Applied Bioelectricity

Applied Bioelectricity
Springer Science+Business Media, LLC

J. Patrick Reilly
Applied
Bioelectricity
From Electrical
Stimulation to
Electropathology
With chapters by
Hermann Antoni
Michael A. Chilbert
James D. Sweeney
With 241 Figures
Springer
J. Patrick Reilly
Applied Physics Laboratory
Johns Hopkins University
and
Metatee Associates
12516 Davan Drive
Silver Spring, MD 20904, USA
Cover art by Anthony E. Randolph, The Johns Hopkins

University Applied Physics Laboratory.
Library of Congress Cataloging-in-Publication Data

Reilly, J. Patrick.
Applied bioelectricity : from electrical stimulation to
electropathology 1 J. Patrick Reilly.
p. cm.
Includes index.
ISBN 978-1-4612-7235-9 ISBN 978-1-4612-1664-3 (eBook)
DOI 10.1007/978-1-4612-1664-3
1. Electrophysiology. 2. Electric shock. I. Title.
QP82.2.E43R437 1998
612'.01427-dc21 97-48860
Printed on acid-free paper.
© 1998 Springer Science+Business Media New York

Originally published by Springer-Verlag New York, Inc. in 1998
Softcover reprint of the hardcover 1st edition 1998
Adapted from Electrical Stimulation and
Electropathology, Cambridge University Press, 1992.
All rights reserved. This work may not be translated or copied in whole or in part without
the written perrnission ofthe publisher (Springer Science+Business Media, LLC), except for
brief excerpts in connection with reviews or scholarly analysis. Use in connection with any
form of information storage and retrieval, electronic adaptation, computer software, or by
similar or dissimilar methodology now known or hereafter developed is forbidden.
The use of general descriptive names, trade names, trademarks, etc., in this publication,
even if the former are not especially identified, is not to be taken as a sign that such names,
as understood by the Trade Marks and Merchandise Marks Act, may accordingly be used
freely by anyone.
While the advice and information in this book are believed to be true and accurate at the date
of going to press, neither the authors nor the editors nor the publisher can accept any legal
responsibility for any errors or omissions that may be made. The publisher makes no warranty,
express or implied, with respect to the material contained herein.
Production coordinated by Chemow Editorial Services, Inc., and managed by Tim Taylor;
manufacturing supervised by Thomas King.
Typeset by Best-set Typesetter Ltd., Hong Kong.
9 8 7 6 5 432 1
ISBN 978-1-4612-7235-9 SPIN 10657964

Preface
The use of electrical devices is pervasive in modern society. The same

electrical forces that run our air conditioners, lighting, communications,
computers, and myriad other devices are also capable of interacting with
biological systems, including the human body. The biological effects of
electrical forces can be beneficial, as with medical diagnostic devices or
biomedical implants, or can be detrimental, as with chance exposures that
we typically call electric shock. Whether our interest is in intended or
accidental exposure, it is important to understand the range of potential
biological reactions to electrical stimulation.
The subject of this book is applied bioelectricity, that is, biological reac-
tions to electrical forces in situations of practical interest. For that purpose,
I have attempted to treat biological reactions with emphasis on human
reactions, from the just discernible to the clearly unacceptable. This book
treats short-term biological reactions, with special attention to reactions of
a detrimental nature. The term "electropathology" is used here to indicate
any undesirable biological reaction to electrical stimulation. Whether a
biological effect is judged beneficial or detrimental often depends on the
context. With controlled medical applications of electrical stimulation, a
reaction may be judged to be beneficial; with uncontrolled chance exposure,
a similar reaction may be judged to be undesirable.
The present work is an adaptation of my previous book, Electrical Stimu-
lation and Electropathology, which was published in 1992 by Cambridge
University Press. Since that time, my continued research in the field of
applied bioelectricity has led to the development of this edition. I have
attempted to make this work useful to the biomedical scientist, as well as to
the individual concerned with electrical safety. The material is directed to a
reader who is familiar with basic physics and basic physiology, but who may
not necessarily be an expert in either discipline. Chapters 2 to 5 provide a
fundamental background by which later discussions on electrical stimula-
tion and electropathology may be understood. Chapters 6 to 10 treat human
reactions to electrical stimulation and are organized according to the nature
of the response.
v
vi Preface
Electrical forces may be introduced into the biological medium through

electrodes that directly contact the subject, or through electric and mag-
netic fields without direct electrode contact. In either case, the biological
system responds to the same bioelectric relationships that are presented
throughout the book. Chapter 9, however, applies these relationships
specifically to electric and magnetic field exposure.
Chapter 11 of this edition treats rationale and standards for exposure to
electromagnetic fields and electric currents in the frequency regime from
zero to several GHz. The chapter emphasizes electromagnetic field expo-
sure, although I have included a section on exposure to electric currents
from consumer products. I have attempted to draw on the material of the
preceding 10 chapters in discussing the rationale that underlies existing and
proposed exposure standards.
As with Electrical Stimulation and Electropathology, this volume focuses
on short-term reactions to electric currents, and electric and magnetic fields.
This focus results in a book that primarily treats biophysical reactions to
relatively high levels of electrical exposure, in contrast to the much lower
chronic exposure issues that have been much discussed in other scientific
and popular forums. The book emphasizes bioelectric mechanisms that are,
for the most part, reasonably well understood, for which theoretical predic-
tive models can explain a wide range of biophysical phenomena, and for
which experimental evidence is available and largely unequivocal. I do not
mean to imply that all the loose ends are tied up with respect to short-term
bioelectric reactions-indeed there remain many issues that are poorly
understood, and much research remain to be done. However, compara-
tively speaking, our grasp of biophysical mechanisms and safety issues
associated with short-term, high-level exposure is much firmer than with
chronic, low-level electrical exposure issues.
As will be evident in Chapter 11, exposure standards are largely devel-
oped with reference to relatively high-level exposures that approach thresh-
olds of demonstrable short-term biological reactions in intact animals and
humans. Nevertheless, public attention is often directed to chronic, low-
level exposure issues. Accordingly, Chapter 11 includes a discussion of
biophysical mechanisms that have been advanced to explain observed bio-
logical reactions to chronic, low-level electrical exposure. Although such
mechanisms are not, for the most part, developed in the preceding chapters,
I can recommend for the interested reader recent books on this subject
(Blank, 1993; Frey, 1994; Gandhi, 1990; Klauenberg et aI., 1995; Lin, 1989;
Norden and Ramel, 1992; Polk and Postow, 1996; Sagan, 1996; Wilson et aI.,
1990).
This book would not have been possible without the contributions of
many people. I am particularly grateful for the help of those at The Johns
Hopkins University Applied Physics Laboratory, and especially for the
support of Stuart S. Janney Fellowships that supported a portion of
my writing and research efforts. The Air Defense Systems Department
Preface vii
(formally the Fleet Systems Department) of the Applied Physics Labora-

tory made available the resources of their art and editorial group. I appre-
ciate the support of the Department Associate Director William Zinger for
making these resources available. I thank the Department staff members
Terry Joslin for editorial and word-processing work, and Anthony
Randolph for new art work for this edition. Art work from the previous
edition, also included here, was done principally by Jacob Elbaz. I am
grateful to the library staff of the Applied Physics Laboratory, who went to
the remote ends of the earth to track down many of my obscure citations.
Many colleagues provided valuable help. John Osepchuk of Full Spec-
trum Consulting reviewed Chapter 11 and made many valuable suggestions.
I particularly wish to acknowledge Professor Willard Larkin of the Univer-
sity of Maryland for the many things I learned from him during the years we
worked together researching electric shock. Professor Larkin is coauthor of
many of the papers that have been adapted into Chapter 7. There were also
many colleagues, too numerous to mention, who sent me papers and
discussed research topics with me. I am grateful to them all.
Special thanks are due to those colleagues who contributed chapters to
this book, namely Hermann Antoni of the University of Freiburg (Chapter
5), James D. Sweeney of the Arizona State University (Chapter 8), and
Michael Chilbert of the General Electric Company (Chapter 10). I also
acknowledge H.A.C. Eaton of The Johns Hopkins University Applied
Physics Laboratory for his collaboration in magnetic brain stimulation, and
for co-authoring Section 9.9.
As anyone who has written a technical book undoubtedly knows, the
process requires a substantial commitment over a long period of time
involving many personal sacrifices that are shared by one's family. I thank
my wife Lynette for her patience and understanding during this project, and
for the many ways that she assisted me.
Laurel, Maryland J. PATRICK REILLY

Contents
Preface .................................................... v
Affiliations ................. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xix
1 Introduction ........................................ 1
1.1 General Perspective ... . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
1.2 Electrical Exposure ............................... 2
Electrical Fatalities ............................. 2
Typical Electrical Exposures ..................... 4
1.3 Scales of Short-Term Reactions
to Contact Current ............................... 5
Sensory Reactions .............................. 7
Muscle Reactions ............................... 7
Cardiac Reactions .............................. 7
Thermal Reactions .............................. 8
Electroporation ................................ 9
1.4 Reactions to Electric and Magnetic Field
Stimulation ....................................... 9
1.5 Variables Affecting Thresholds ..................... 10
2 Impedance and Current Distribution ............... 12

2.1 Dielectric Properties of Biological Materials . . . . . . . . . 12
Conductivity and Permittivity .................... 12
Cellular Membranes ............................ 19
Skin Depth .................................... 20
2.2 Skin Impedance .................................. 20
Detailed Structure of Skin ....................... 21
Equivalent Circuit Models ....................... 23
Area Proportionality of Skin Admittance .......... 26
Skin Capacitance ................................ 28
Time-Variant and Nonlinear Aspects
of Skin Impedance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28
2.3 Total Body Impedance: Low-Frequency and DC ...... 31
ix
x Contents
Distribution of Current and Impedance Within

the Body ...................................... 37
Current Through the Heart ...................... 41
Applicability of Measurements on Corpses ......... 41
Statistical Distribution of Impedance .............. 42
2.4 Impedance at Higher Frequencies .................. 45
2.5 Impedance Through Foot Contact .................. 48
2.6 High-Voltage and Transient Properties .............. 52
Spark and Contact Components .................. 54
Plateau Voltage ................................ 56
Spark Discharges and Corneal Degradation ........ 57
Discharge Impedance ........................... 58
Variation with Stimulus Location ................. 60
Polarity Effects ................................. 61
Stimulus Time Constants ........................ 63
Minimum Impedance with Spark Discharge ........ 64
Internal Body Impedance ........................ 65
2.7 Impedance of Domestic Animals ................... 68
3 Electrical Principles of Nerve and

Muscle Function .................................... 73
3.1 Introduction. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 73
3.2 Cellular Membranes .............................. 75
The Nernst Equation ............................ 76
3.3 The Excitable Nerve Membrane .................... 80
The Hodgkin-Huxley Membrane ................. 81
The Frankenhaeuser-Huxley Membrane ........... 86
Species Dependence of Action Potential
Dynamics ...................................... 87
Propagation of Nerve Impulses ................... 88
3.4 Action Potential Models for Cardiac Tissue .......... 90
3.5 Sensory Transduction ............................. 94
Intensity Coding ................................ 95
3.6 Muscle Function .................................. 97
3.7 Synapses ........................................ 100
General Properties .............................. 100
Synaptic Interactions with In-Situ E-Fields ......... 101
3.8 The Spinal Reflex ................................ 103
4 Excitation Models .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 105

4.1 Introduction. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 105
4.2 Linear Strength-Duration Model ................... 105
Empirical Strength-Duration Relationships ......... 109
4.3 Electrical Cable Representations ................ . .. 111
One-Dimensional Cable Models .................. 112
Contents Xl
Nonlinear Models 116

4.4 Myelinated Nerve Model ......................... . 117
Threshold Criterion ............................ . 121
Position and Number of Nonlinear Nodes ......... . 123
Suprathreshold Response ....................... . 126
4.5 Response to Monophasic Stimulation .............. . 126
Strength-Duration Relations .................... . 126
Polarity Sensitivity ............................. . 128
Current Density and Electric Field
Relationships .................................. . 129
4.6 Response to Biphasic and Repetitive Stimuli ........ . 132
Strength-Duration Relationships ................. . 132
Sinusoidal Stimuli ............................. . 134
Repetitive Stimuli ............................. . 141
Nonlinear Interaction of MUltiple Waveforms ..... . 143
4.7 Parameter Variation Effects ....................... . 145
5 Electrical Properties of the Heart .................. 148

HERMANN ANTONI
5.1 Cardiovascular System: General Anatomical and
Functional Aspects ............................... 148
5.2 Origin and Spread of Excitation .................... 151
Autorhythmicity and Geometry of Propagation ..... 152
Hierarchy of Pacemaker Activity and
Artificial Pacemakers ........................... 153
5.3 Elementary Processes of Excitation
and Contraction ................................. . 154
Ionic Mechanisms of Excitation 154
Mechanism of Contraction ...................... . 156
Excitation-Contraction Coupling ................. . 158
5.4 Stimulation, Propagation, and Refractoriness ........ . 159
Elementary Mechanisms ........................ . 159
Conduction of the Action Potential .............. . 160
Definition and Mechanism of Refractoriness ...... . 161
5.5 Regular and Ectopic Pacemakers .................. . 163
Elementary Events in Impulse Formation ......... . 163
Actual and Potential Pacemakers ................ . 164
Ectopic Pacemakers ............................ . 166
5.6 Effects of Autonomic Nerves and of Changes
in Electrolyte Composition ........................ . 166
Parasympathetic and Sympathetic Innervation ..... . 167
Chronotropy, Inotropy, and Dromotropy ......... . 167
Vagal and Sympathetic Tone .................... . 169
Mechanism of Autonomic Transmitter Actions 169
Effects of the Ionic Environment and of Drugs ..... 170
xii Contents
5.7 Electrocardiogram. . . . . . . . . .. . . .. . . . . . . . . . . . . . . . .. 171

ECG Form and Nomenclature; Relation to
Cardiac Excitation .............................. 171
Origin of ECG ................................. 173
ECG Recording ................................ 174
Use of the ECG in Diagnosis .................... 177
5.8 Abnormalities in Cardiac Rhythm as Reflected
in the ECG ...................................... 178
Rhythms Originating in the A V Junction .......... 178
Rhythms Originating in the Ventricles ............. 180
Extrasystoles ................................... 180
Atrioventricular Disturbances of Conduction ....... 180
Atrial Flutter and Fibrillation .................... 180
Ventricular Flutter and Fibrillation ................ 182
5.9 Mechanism of Flutter and Fibrillation ............... 182
Conditions for Reentry; Anatomical and
Functional Pathways ............................ 183
Length of the Excitation Wave ................... 183
Mechanisms of Abbreviated Refractory Periods .... 183
Mechanisms of Slow Conduction ................. 185
Unidirectional Block and One-Way Conduction .... 186
5.10 Vulnerable Period: Threshold for Fibrillation ......... 187
The Threshold for Fibrillation .................... 188
Relation to the Threshold for Stimulation .......... 189
Effects of Alternating and Pulsed Direct Current ... 190
Significance of Single and Repetitive Extrasystoles ... 192
5.11 Electrical Defibrillation ............................ 192
6 Cardiac Sensitivity to Electrical Stimulation . . . . . .. 194

6.1 Introduction ..................................... 194
6.2 Threshold Sensitivity with Respect to
Cardiac Cycle .................................... 194
6.3 Strength-Duration Relations for
Unidirectional Currents ............................ 198
Electrode Size .................................. 199
Polarity ........................................ 199
Size of Tissue Preparation ........................ 202
Fibrillation by Prolonged Currents ................ 202
Consistency of Data ............................. 203
6.4 Biphasic and Sinusoidal Stimulation ................. 203
6.5 Duration Sensitivity for Oscillatory Stimuli ........... 206
Excitation Sensitivity ............................ 206
Fibrillation Sensitivity ........................... 208
The Z Relationship for Fibrillation by
AC Currents ................................... 212
Contents xiii
6.6 Energy Criteria and Impulse Currents ............... 217

6.7 Body-Size Scaling ................................. 220
Experimental Body-Size Data .................... 220
Scaling Arguments .............................. 224
6.8 Statistical Distribution of Thresholds ................ 225
6.9 Combined AC and DC Stimuli ..................... 227
6.10 Electrodes and Current Density .................... 230
Electrode Area ................................. 230
Current Density ................................ 234
Locus and Direction of Stimulus Current .......... 236
7 Sensory Responses to Electrical Stimulation ....... 240

7.1 Introduction ..................................... 240
7.2 Mechanisms of Electrical Transduction .............. 240
7.3 Perception of Transient Monophasic Currents ........ 244
Strength-Duration Relationships .................. 244
Capacitor Discharges ............................ 248
7.4 Supra threshold Responses ......................... 253
Magnitude Scaling .............................. 253
Categorical Scaling .............................. 254
AC or Repetitive Stimuli ........................ 256
Dynamic Range of Electrical Stimulation .......... 258
7.5 Stimulus Waveform Factors ........................ 260
Biphasic Stimuli ................................ 260
Repetitive Stimuli .............................. 261
Sinusoidal Stimuli ............................... 265
Polarity Effects ................................. 269
7.6 Electrodes and Current Density .................... 270
Cutaneous Electrodes ........................... 270
Current Density Considerations .................. 274
7.7 Body Location Sensitivity .......................... 274
7.8 Skin Temperature ................................ 277
7.9 Tactile Masking .................................. 280
7.10 Individual Differences in Electrical Sensitivity ........ 282
Potential Sources of Variability ................... 282
Distribution of Sensitivity ........................ 283
Correlates of Sensitivity ......................... 287
7.11 Startle Reactions ................................. 290
7.12 Electrical Stimulation of Domestic Animals .......... 291
Reactions to 60-Hz Stimulation ................... 292
Reactions to Transient Stimulation ................ 294
Statistical Distribution of Bovine
Reaction Thresholds ............................. 296
Magnetic Field Stimulation of Cows ............... 296
7.13 Visual and Auditory Effects ........................ 298
xiv Contents
8 Skeletal Muscle Response to

Electrical Stimulation .. . . . . . . . . . . . . . . . . . . . . . . . . . . .. 299
JAMES D. SWEENEY
8.1 Introduction. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 299
8.2 Neuromuscular Structure and Function .............. 299
Skeletal Muscle Innervation ...................... 299
Peripheral Nerve Trunk Structure .... . . . . . . . . . . .. 300
Muscle Action Potentials and Excitation-Contraction
Coupling ...................................... 301
Skeletal Muscle Force Production ................. 302
Differentiation of Skeletal Muscle Fiber Types ..... 303
Recruitment and Firing Patterns .................. 304
Fatigue in Normal Physiological Use .............. 305
Dynamic Nature of Skeletal Muscle Fiber Types .... 306
8.3 Fundamental Principles of Skeletal Muscle Electrical
Stimulation ...................................... 307
Strength-Duration (S-D) Relations for
Neuromuscular Excitation ....................... 307
Regulated-Current Versus Regulated-Voltage
Stimulation ..... . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 309
Biphasic Stimulation ............................ 310
Fatigue and Conduction Failure in Response to
Electrical Stimulation ........................... 310
Sinusoidal Stimulation Strength-Frequency (S-F)
Effects ........................................ 311
8.4 Functional Neuromuscular Stimulation Systems ....... 312
Objectives and Applications of FNS Systems
and Technologies ............................... 312
Electrical Excitation of Skeletal Muscle by Functional
Neuromuscular Stimulation ...................... 314
Electrodes for Functional Neuromuscular
Stimulation .................................... 314
Recruitment and Rate Modulation in Functional
Neuromuscular Stimulation ...................... 319
Electropathology of Functional Neuromuscular
Stimulation Systems ............................. 322
8.5 Skeletal Muscle Stimulation in Electrical Accidents ... 323
History and Overview ........................... 324
Let-Go Thresholds for Hand Contact with
Electrical Stimuli ............................... 324
8.6 Analysis of the Let-Go Phenomenon ................ 330
Functional Anatomy of the Hand and Wrist ........ 330
Excitation Threshold Estimates ................... 334
8.7 Effects of Electrical Stimulation on Respiration ....... 338
Contents xv
Anatomy and Physiology of Respiration ........... 338

Accidental Electric Shocks and Respiration ........ 339
Artificial Respiration by Electrical Stimulation ..... 340
9 Stimulation via Electric and Magnetic Fields ....... 341

9.1 Introduction. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 341
9.2 Electric Field Induction Principles .................. 341
Induction Mechanisms ........................... 342
Body Surface Effects ............................ 345
Induced Electric Shock Principles ................. 347
9.3 Direct Perception of ELF Electric Fields ............. 353
Mechanisms for Human Detection ................ 353
9.4 Human Reactions to AC Electric
Field-Induced Shock .............................. 358
Perception Thresholds ........................... 358
Superthreshold Effects .......................... 360
Multiple Discharges ............................. 360
Extrapolation to Other Frequencies ............... 361
Skin Erosion from Low-Energy Discharges ........ 362
9.5 Time-Varying Magnetic Field Induction ............. 362
9.6 Principles of Excitation by Time-Varying
Magnetic Fields .................................. 368
Pulsed Fields: Large Area Exposure ............... 369
Minimum Excitation Thresholds .................. 370
Strength-Duration Time Constants ................ 371
Anatomical Considerations ...................... 371
Calculated Nerve and Cardiac Excitation
Thresholds ..................................... 373
Sinusoidal Fields: Large Area Exposure ........... 375
Criteria for Nerve and Heart Excitation: Large
Area Exposure ................................. 376
Body Size Scaling ............................... 376
9.7 Experimental Investigations of Magnetic
Excitation: Large Area Exposure ................... 379
Thresholds for Nerve Excitation .................. 379
Thresholds for Cardiac Excitation ................ 382
Stimulus Duration and Frequency Effects .......... 384
Suprathreshold Nerve Excitation Reactions ........ 386
9.8 Visual and Auditory Reactions to
Electromagnetic Exposure ......................... 386
Direct Perception of Magnetic Fields .............. 386
Visual Sensations ............................... 387
Mechanisms for Phosphenes ..................... 390
xvi Contents
Duration and Frequency Relationships for

Visual Effects .................................. 391
Implications of Phosphenes for
CNS Synaptic Interactions ....................... 392
Auditory Sensations ............................ 393
9.9 Local Magnetic Stimulation ........................ 396
Small Coil Stimulators ........................... 397
Induced Electric Field ........................... 400
Local Stimulation of the Brain Cortex ............. 402
9.10 Scales of Reaction: Power Frequency
Magnetic Field Exposure of the Head ............... 406
9.11 Magnetic Forces on Moving Charges ................ 409
Magnetohydrodynamic Effects ................... 409
Ion Resonance ................................. 411
10 High-Voltage and High-Current Injuries ........... 412

MICHAEL A. CHILBERT
10.1 Introduction ..................................... 412
10.2 Modes of Injury .................................. 413
Thermal Injury ................................. 413
Electroporation ................................ 416
Fibrillation ..................................... 417
10.3 Impedance Considerations and Current Distribution
in the Body ...................................... 420
Total Body Impedance .......................... 420
Contact Impedance and Segmental Body
Impedance ..................................... 426
Tissue Current Densities and Current Distribution. .. 430
lOA Thermal Trauma ................................. 435
Heating at the Contact Site ...................... 435
Heating of the Tissues ........................... 438
10.5 Nonthermal Trauma .............................. 442
10.6 Lightning Injuries ................................. 443
10.7 Clinical Observations ............................. 444
Electrical Burns ................................ 445
Tissue Lesions ................................. 446
Neurologic Sequelae ............................ 447
Cardiovascular Effects ........................... 449
10.8 Clinical Treatment ................................ 450
11 Standards and Rationale . . . . . . .. . . . .. . . . . . . . . . .. ... 454

11.1 Introduction...................................... 454
11.2 Electromagnetic Field Exposure Standards ........... 454
Standards of the IEEE and ANSI ................. 454
Contents xvii
Rationale for IEEE/ANSI C95.1 Standards ........ 455

Electric and Magnetic Field Criteria for
IEEE/ANSI C95.1 .............................. 457
Induced Current and Electric Shock
Considerations ................................. 459
Other Standards on Electromagnetic Exposure ..... 461
Electromagnetic Exposure Limits at
Low Frequencies ............................... 463
Static Electric and Magnetic Fields ................ 468
11.3 Pulsed Electromagnetic Fields ...................... 468
Pulsed Field Limits in Magnetic Resonance
Imaging ....................................... 469
Extrapolation of MRI Exposure Limits to
EMF Standards ................................. 472
Limits of Applicability ........................... 475
11.4 Consideration of Spark Discharges in
EMF Standards ................................... 476
11.5 Absorbed Energy and Thermal Considerations in
EMF Standards ................................... 478
Specific Absorption Rate (SAR) Limits ............ 478
Endogenous Field Limits to EMF Interactions ...... 479
11.6 Consideration of EMF Interaction Mechanisms
in Standards Setting ............................... 481
Categories of Mechanisms of Bioelectric
Interaction ..................................... 481
Established Mechanisms for Human Bioelectric
Response ...................................... 484
Proposed Mechanisms for Human
Bioelectric Response ............................ 486
11.7 ELF Magnetic Field Standards Derived from
Established Mechanisms ........................... 489
Threshold Criteria .............................. 489
Limit Criteria .................................. 490
Other Considerations for ELF Standards .......... 493
11.8 Standards in Consumer Products and Installations .... 493
Standards Setting Agencies ...................... 493
Safety Criteria for Consumer Products ............. 494
Hazard Criteria Versus Current Duration and
Frequency ..... ,............................... 496
Limits for Capacitor Discharges .................. 499
Ground Fault Current Interrupter ................ 500
References ................................................. 502

Index ...................................................... 553
Affiliations
HERMANN ANTONI
Physiological Institute
University of Freiburg
D 7800 Freiburg i. Br.
Germany
MICHAEL A. CHILBERT
General Electric Company
Medical Systems Division
Milwaukee WI 53201
USA
J. PATRICK REILLY
Applied Physics Laboratory
The Johns Hopkins University
Laurel, MD 20723
and
Metatec Associates
12516 Davan Drive
Silver Spring, MD 20904
USA
JAMES D. SWEENEY
College of Engineering and Applied Science
Department of Chemical, Biology, and Materials Engineering
Arizona State University
Tempe, Arizona 85287-6006
USA
xix
1
Introduction
1.1 General Perspective

Electrical forces are vital for the functioning of living things, from the
metabolism of individual cells to human consciousness derived from the
activity of the brain. When electric currents are artificially introduced into
a living organism, these natural forces can be activated or modified. The
result is either detrimental or beneficial, depending on the circumstances.
The term electric shock is generally used to describe the response of the
body to inadvertent electrical exposure, and the consequences of electric
shock are usually considered undesirable. However, the biological mecha-
nisms responsible for electric shock may also be activated in a controlled
fashion for beneficial medical purposes.
The fact that electricity can interact with biological processes has been
known for more than 2,000 years (McNeal, 1977). The electrical discharge
of the torpedofish was reported to have been used as early as 46 A.D. to treat
pain. It is tempting to scoff at early claims of beneficial uses of electricity,
but when one considers modern methods of pain management by electrical
means, one is led to suspect that there might have been valid reasons for
some of these early beliefs.
The beginnings of quantitative bioelectric science can arguably be as-
cribed to the investigations of Galvani (around 1790) and later of Volta.
Galvani observed motion in severed frogs' legs when he touched them
with metallic wires. He supposed that he was releasing stored "animal
electricity," and that this was responsible for the observed muscle activity.
It was later demonstrated that the dissimilar metals of the wires used in
his procedures were in fact generating the electrical forces. His efforts
did, however, encourage further investigations of electrical stimulation.
Volta's invention of the "Voltaic pile" gave science a battery that could
be used in systematic and controlled investigations. Later biological
investigations by Faraday (around 1831) demonstrated that interrupted
electric current was an effective means of electrical stimulation of
nerves. The term "Voltaic" stimulation came to be used to indicate direct
J. P. Reilly, Applied Bioelectricity 1

© Springer-Verlag New York, Inc. 1998
2 1. Introduction
current stimulation, and "Faradaic" to indicate pulsed or interrupted

stimulation.
1.2 Electrical Exposure

Electrical Fatalities
It has been long recognized that man-made electricity can be harmful and
can even cause death. This understanding eventually led to the use of
electrical executions of criminals to replace the "less civilized" means that
had previously been devised. The first electrocution in the United States
occurred in 1890 (Leyden, 1990). The number of electrocutions of criminals
since that time is uncertain, but is estimated to number approximately 4,100
in the United States. Today, approximately 900 prisoners in 14 states await
execution in the electric chair.
The National Center for Health Statistics publishes data on the mortality
from electrical accidents in the United States. Figure 1.1 illustrates U.S.
mortality statistics for the years 1975 to 1987 as summarized by Smith
(1990). The figure shows data for all electrical fatalities, excluding lightning
incidents, and for fatalities related to consumer products.
Because the mortality data do not specify the product involved in the
accident, the number of fatalities related to consumer products was inferred
by Smith on the basis of the reported locations of the accidents.
The downward trend of electrical mortalities is striking. And if one takes
into account the increase in the U.S. population during the period reported
in Fig. 1.1, the trend is even more impressive the per-capita rate of fatalities
related to consumer products declined from 3.0 per million in 1975 to 1.3
per million in 1987. According to Smith, this trend could be the result
of a number of factors, including (1) a steady increase in the number of
residences having ground-fault current interrupters (GFCIs); (2) an in-
crease in the number of locations having GFCIs within individual dwellings;
(3) a decrease in citizen-band base-station antennas; and (4) an increase in
the manufacture of double-insulated power tools. Details of some of these
precautionary measures are given in Chapter 11. The downward trend
illustrated in Fig. 1.1 contrasts with a slight increase in electrical mortality
during the period 1960 to 1975 (Daiziel, 1978), a period when electrical
safeguards were being developed but were not yet widely implemented.
Table 1.1 provides a further breakdown of electrical mortalities. The
table also enumerates deaths by lightning, of which there are nearly 100
each year in the United States. Lightning accounts for roughly 10% of the
total number of electrical fatalities. It has been estimated that one third of
all lightning strikes to humans are fatal (Biegelmeier, 1986).
The worldwide experience of industrialized countries also shows a down-
turn in electrical fatalities (Kieback, 1988). The annual number of fatalities
Electrical Exposure 3
1200
(/)
CD
1000
:~
(ij
E
'0 800
....
CD
.0
E Q
::J
Z 600 ~ Accidents involving
~mer products
400
~-o-O'o
200~~~~~~~~~~~~~~~J-~~~~~
ro ~ ~ ~ 00
Year (19--)
FIGURE 1.1. Mortality from electrical accidents in the United States for the years
1975-1987 (per data compiled by Smith, 1990).
per million inhabitants for 1982 (the latest year reported by Kieback)
ranged from a low of 0.42 for the Netherlands to a high of 7.66 for Hungary.
Kieback emphasizes that such comparisons should be treated with caution
because of different reporting standards in various countries.
TABLE 1.1. Mortality because of electrical accidents.

1987 1986 1985 1984
Electric current
Total (all causes) 760 854 802 888
Domestic wiring and appliances 121 150 146 148
Power generating, distr., transmission 177 182 196 215
Industrial wiring, appliances, machinery 64 89 69 68
Other unspecified electric current 398 433 391 457
Lightning 99 78 85
Source: Vital Statistics of the united States, Volume II-Mortality, U.S. Department
of Health and Human Services, National Center for Health Statistics, Washington,
DC.
4 1. Introduction
It is likely that the mortality data referred to above underreport the

actual number of electrical fatalities. The data for the United States have
been developed from death certificates. Not all electrical fatalities may have
been identified in the certificates as such. Furthermore, deaths from injuries
incidental to electric shock, such as falls, are probably not reported in the
electrical category. Delayed fatalities, such as those caused by burns, also
may have been omitted. Even if we allow for underreporting, the number of
electrical fatalities remains small in comparison with other accidental
causes in the United States. Death from choking on food or other ingested
objects, for example, exceeds electrical fatalities by nearly five times.
Typical Electrical Exposures

Notwithstanding the relatively small number of electrical fatalities, the
acceptability of electrical exposure by the public is a matter of growing
concern. One area of concern is related to the increasing number of electri-
cal consumer products. Although potentially fatal electrical exposures are
seldom an issue in most consumer products, in some cases exposures might
be unpleasant, or they might lead to injuries from startle reactions. Both
consumers and manufacturers hope to avoid such exposures.
High-voltage transmission lines provide another source of public expo-
sure. Increases in population and in individual electrical demand press
utilities to increase electricity generation and transmission. The trend in
transmission is to higher voltages because of economies in the cost of
construction and operation of transmission facilities. One environmental
consequence of high-voltage transmission is the potential for electric shock
arising from the electric fields that the lines inevitably produce (see Chapter
9). The electric shock resulting from transmission line fields does not have
the potential for injury, except possibly in certain unusual circumstances.
Nevertheless, perceptible shock from transmission line fields can be un-
pleasant and can provoke fear in the exposed public.
Probably the most rapidly growing area of electrical stimulation is in
biomedical technology. Electrical stimulation is being used increasingly as a
tool for medical diagnosis, therapy, and prosthesis; examples of biomedical
applications are listed in Table 1.2. Even this incomplete list is a testament
to the numerous medical applications of electrical stimulation. Electrical
stimulation for medical purposes may be introduced through electrodes in
contact with the skin, through implanted electrodes, or through magnetic
induction. The latter method precludes the necessity for any electrode
contact whatsoever.
Electrical stimulation that might be considered detrimental in chance
exposures can be beneficial when used in a controlled fashion. For example,
cardiac arrhythmia's caused by chance electric shock are regarded as being
potentially life threatening; these same responses can be life saving when
used in implanted pacemakers. Furthermore, a response that is considered
Scales of Short-Term Reactions to Contact Current 5
TABLE 1.2. Examples of electrical stimulation in biomedical applications.

Restoration of muscle function after nerve injury
Preservation of muscle tone after nerve injury
Treatment of scoliosis
Diaphragm stimulation for respiration control
Electrical stimulation of sphincter for urinary control
Correction of foot-drop
Sensory aids for the blind
Cochlear prosthesis for the deaf
Management of intractable pain
Inhibition of intractable self-injurious behavior
Diagnosis of peripheral nerve function
Diagnosis of muscle function
Functional diagnosis and mapping of the brain cortex
Stimulation of the visual cortex
Electroconvulsive therapy
Automatic cardiac pacing
Automatic sensing and reversal of fibrillation (implants)
Defibrillation in emergency aid (external)
Bone healing
Electrical diathermy
Imaging of soft tissue
undesirable in one medical application may be considered beneficial in

another. Electrically induced pain, for example, is considered a difficulty to
be avoided in procedures involving the introduction of currents through
electrodes on the skin. But painful electrocutaneous stimulation has also
been used in a beneficial way to inhibit otherwise intractable self-injurious
behavior (Newman, 1984).
In this work, the term "electropathology" is used to indicate any
undesirable biological reaction to electrical stimulation. The study of
electropathology can help to reduce the possibility of unacceptable expo-
sures to electrical equipment and to understand better how to use electrical
stimulation for beneficial reasons. Regardless of our particular orientation,
it is useful to understand the range of potential human reactions and their
underlying mechanisms.
1.3 Scales of Short-Term Reactions to Contact Current

Several scales of reactions are of interest in the study of electropathology
and electrical stimulation. These scales can be best appreciated by way of
example. Figure 1.2 illustrates possible reactions to 60-Hz alternating cur-
rent flowing between the hand and feet. It is assumed that an adult is
gripping a large electrode in one hand, while a return electrode is contacting
a foot or both feet. The exposure duration is assumed to be approximately
5 s. Categories of reactions are identified in the five columns in the figure,
6 1. Introduction
Muscle Cardiac Thermal Electroporation

Sensation
reaction reaction effect (EP)
Perception
Twttch
Discomfort
Pain
Intolerable pain
10
<n Grip tetanus
E Resp. interfer.
Resp. tetanus
«
S
Qi
> 6T=I°C
.!!!
C
~
::> Reversible EP
()
100 Excitation T = 45° C
T = 70° C
Fibrillation
Irreversible EP
1000 Defibrillation
t
FIGURE 1.2. Potential short-term reactions to 60-Hz current. Assumed conditions:
5-s exposure; hand grip of large electrode; return electrode at feet; large adult
subject; median response.
each involving a distinct continuum of reactions. Various thresholds of

reaction are identified by descriptors whose vertical placement indicates an
approximate median value of the electrical threshold, which is read on the
left-hand scale. These thresholds are only approximate, and are subject to
many variables. The intent of Fig. 1.2 is to provide a rough comparison of
thresholds, rather than precise numerical values. As the indicated values are
approximate medians of a statistical distribution, they would not normally
be suitable as acceptability criteria unless some safety factor were applied.
The following is a commentary on each category. The commentaries
call attention to particular chapters (6-10) where a detailed treatment may
Scales of Short-Term Reactions to Contact Current 7
be found. Additional chapters (2-5) provide a foundation of bioelectric

principles and models that explain the experimental data and related
mechanisms described in later chapters.
Sensory Reactions (Chapter 7)

Electrical sensation involves the same pathways that are used to sense our
external and internal environment. The dynamic range of electrical thresh-
olds from perception to pain is, however, a small fraction of that for natural
stimuli. The dynamic range for 60-Hz current is particularly small. Al-
though pain thresholds can be measured with a reasonable degree of re-
peatability, the measurement of thresholds for "intolerable pain" may vary
considerably with the experimental context. The tolerance threshold indi-
cated in Fig. 1.2 indicates an experimental tolerance limit, described in
Chapter 7, but does not necessarily represent the most extreme limit of
human endurance. Indeed, subjects participating in hand-grip experiments
(the "let-go" level in the next column) willingly endured considerably
higher levels.
Muscle Reactions (Chapter 8)

The quantum of muscle reaction is called a "twitch." The threshold for the
smallest measurable twitch for a hand-grip electrode is expected to be
nearly equal to that for sensation. For the conditions represented in Fig. 1.2,
the threshold twitch (around LOrnA) would occur in the small musculature
of the hand. At much higher levels (around 15mA), one encounters grip
tetanus, where the hand "freezes" to the conductor (the so-called let-go
threshold). The muscles responsible for grip tetanus lie much higher up in
the forearm. It is probable that the small muscles of the hand would be
tetanized at much lower current levels, but their effects on grip would be
easily overcome by the finger extensors in the forearm. However, at current
levels that cause grip tetanus, electrical stimulation of more powerful
flexors located in the forearm would dominate.
Respiratory interference can also result from electrical exposure.
Quantitative data supporting particular thresholds are rather sparse. Anec-
dotal reports gathered during grip tetanus tests indicate that respiratory
interference may occur at levels somewhat above the grip tetanus thresh-
old, and respiratory tetanus somewhat above that. The current levels
indicated for respiratory effects in Fig. 1.2 are rough estimates based on
these reports.
Cardiac Reactions (Chapters 5 and 6)

The cardiac excitation threshold indicated at about lOOmA in Fig. 1.2
applies to an extra beat elicited during the normally relaxed state of the
8 1. Introduction
heart. Such excitation is not necessarily life threatening, but is nevertheless

treated as a potentially serious effect. At a higher current (around 240mA),
it is possible to induce ventricular fibrillation, in which uncoordinated
contractions of heart muscle preclude pumping. Since the human heart
rarely recovers spontaneously from electrically induced fibrillation,
death will soon result unless defibrillation equipment can be applied to the
victim.
Defibrillation occurs at still higher current levels (several amperes).
Here, the fibrillating heart is forced into a uniform state of excitation, after
which normal rhythmic activity becomes possible. Consequently, fibrilla-
tion is most probable within a band of current. Below the lower limit of the
band, the current is too feeble to excite the heart; above the upper limit,
the heart is defibrillated. This would explain why electrical fatalities due
to cardiac failure occur most often at low voltages, whereas high-voltage
injuries are principally from burns (see Chapter 10).
Thermal Reactions (Chapter 10)

The greatest temperature rise due to the current from a gripped conductor
would probably occur in the high-current-density region of the wrist, or
possibly in the hand at the edges of the conductor. With a 60-Hz stimulus,
a 1 DC temperature rise would not be sensed, because the current necessary
to produce it would also produce severe pain that would mask any thermal
perception.
Tissue heating depends largely on the root-me an-square value of the
current and relatively little on its waveform. Excitation of sensory nerves,
on the other hand, is very sensitive to the stimulus waveform. If the wave-
form were inefficient for electrical stimulation, it would be possible for
electrical perception thresholds to exceed thermal perception thresholds.
Such a condition occurs with continuous sinusoidal current if the stimulus
frequency exceeds 105 Hz. In that case, the current would be sensed as
warmth, without electrical stimulation.
When skin or muscle tissue is heated to about 45 DC for prolonged
periods, thermal damage can result. At that temperature, cutaneous
nociceptors would probably be stimulated, resulting in pain (the body's
natural defense against heat injury). Thermal damage, however, would not
likely occur at 45 DC for a duration as short as 5 s. A temperature rise of
somewhere around 70°C would be needed to sustain permanent heat
damage at that duration. Thermal perception for current frequency
above 105 Hz is indicated at 35mA for a touched contact (see Fig. 7.12). A
large area grip contact will require an increased current to produce the
same local temperature rise. Assuming that perception occurs with a tem-
perature rise of 1 DC, the current for higher temperature rises can be
estimated by using the relationship that the current is proportional to ~ I1T
[see Eq. (10.3)].
Reactions to Electric and Magnetic Field Stimulation 9
Electroporation (Chapter 10)

Current flowing within a biological medium will create potential differences
across a cellular membrane. This effect is strongest in elongated cells (e.g.,
nerve and muscle cells) that are oriented in a direction parallel to the
current flow. At the lowest levels on the scale in Fig. 1.2, the alteration
of membrane potential will excite nerves and result in both sensory and
muscular reactions. At much higher levels, the intense electric field that
develops across the cellular membrane will promote the formation of
pores. This process of electroporation (EP) is reversible at formative levels,
but becomes irreversible at higher levels, leading to cellular death.
All cells normally maintain a natural "resting" potential, which for
muscle cells is around 90mV, with the inside negative relative to the outside
(see Chapter 3). In response to an external current flowing parallel to the
long axis of the cell, the membrane will be hyperpolarized at the cathode-
facing end and depolarized at its anode-facing end. The thresholds indi-
cated in Fig. 1.2 are predicated on a membrane voltage of 200mV for
reversible EP and 800mV for irreversible EP (see Chapter 10). These
thresholds correspond to hyperpolarization of HOmV for reversible EP
and 710mV for irreversible EP in the hypothetical muscle cell used in this
example. The corresponding current levels indicated in Fig. 1.2 are with
reference to the maximum current density (and cellular polarization)
occurring in the wrist.
1.4 Reactions to Electric and Magnetic

Field Stimulation
The exposure conditions assumed in Fig. 1.2 apply to current conducted
through direct contact with an energized electrode. It is also possible to
produce short-term bioelectric reactions without direct electrode contact,
as in the case of exposure to electromagnetic fields (EMF). With EMF
exposure, an electric field and associated current density is induced within
the biological medium by the external electric or magnetic field. Stimula-
tion of excitable tissue is typically achieved most easily by magnetic fields at
frequencies below 1 MHz, and thermal effects are most readily produced at
higher frequencies. The principles governing EMF stimulation and thresh-
olds of reaction are treated in Chapter 9.
Whereas the bioelectric mechanisms applying to contact current also
apply to EMF stimulation, the measures of subject exposure may be
quite different. To illustrate this point, Fig. 9.28 provides an example of
scales of human reactions to short-term exposure ofthe head to a magnetic
field at power frequencies (50, 60Hz). The reactions listed include
phosphenes, visual evoked potential effects, excitation of brain neurons,
and seizures. Measures of exposure shown in Fig. 9.28 are the applied
10 1. Introduction
flux density in units of Tesla (T), the time rate of change of the flux density
in Tesla per second (Tis), the induced electric field within the brain in
volts per meter (Vim), and the associated current density in amperes
per square meter (Nm 2 ). These four metrics of exposure can be tied
together only for specified conditions which include the frequency of the
incident field, the anatomical region of exposure, and the spatial distribu-
tion of the applied field.
The lowest threshold shown in Fig. 9.28 is associated with phosphenes,
which are visual sensations produced by nonphotic stimuli. Experimental
evidence on electrical phosphenes, discussed in Sect. 9.8, indicate that
the site of electrical interaction is at synaptic processes within the retina.
One significance of electrical phosphenes is that they are informative
of synaptic interactions with in situ electric fields that may take place
within the brain. Additional discussion of synaptic interactions is provided
in Sect. 3.7.
At levels of exposure above the thresholds for phosphenes, Fig. 9.28
indicates excitation thresholds for brain neurons. The responsible bioelec-
tric mechanisms are the same as those responsible for nerve and muscle
excitation effects, which are represented in the first three reaction columns
of Fig. 1.2. These mechanisms are discussed in Chapters 3 and 4.
Other short-term reactions to EMF exposure include auditory effects
(Sect. 9.8) and heating effects (Sect. 11.5), both of which invoke thermal
mechanisms related to the induced current density within the biological
tissue. Thermal effects may become significant at much higher frequencies
of EMF exposure than applicable to the example of Fig. 9.28. At the
opposite end of the frequency spectrum, with static or extremely low fre-
quency magnetic fields, short term reactions include magneto hydrody-
namic effects, as discussed in Sect. 9.11.
1.5 Variables Affecting Thresholds

The thresholds discussed in Sects. 1.3 and 1.4 are illustrative values for a
particular set of conditions. Many variables strongly affect these thresholds.
These variables can be grouped into categories associated with the stimulus
waveform, the spatial distribution of the stimulus, and the subject.
A significant part of the parametric sensitivity to electrical stimulation
can be understood in the light of basic principles of bioelectricity. These
basic laws are treated in the initial chapters of this book. Chapter 2 covers
impedance and internal current distribution. In Chapter 2 the biological
subject is considered for its passive electrical properties. Chapter 3 develops
the principles that govern the electrical response of nerve and muscle. In
Chapter 4 these principles are extended to computational models that allow
one to study the excitatory effects of electrical stimulation. An "electrical
cable" model is developed to help explain how electrical excitation is
Variables Affecting Thresholds 11
related to the temporal and spatial aspects of the stimulus. The electrical
properties of the heart are developed in Chapter 5.
The excitation model of Chapter 4 has been extensively referred to in this
book when discussing electrical excitation of the heart (Chapter 6), of
sensory processes (Chapter 7), and of muscle (Chapter 8). The model is also
used to derive excitation thresholds pertaining to electric and magnetic field
exposures (Chapter 9).
Considering the extensive number of parameters that affect electrical
sensitivity, one might wonder whether it is feasible to define "safe" or
"acceptable" exposure levels. The subject of safety criteria is treated in
Chapter 11, where performance criteria and electrical safeguards are dis-
cussed. The approach for setting electrical safety standards described in
Chapter 11 is to make conservative assumptions regarding parameters, such
as body size, impedance, stimulus waveform, and statistical threshold varia-
tions. While this approach does not necessarily protect against the most
extreme combination of sensitivity factors, these standards are intended to
provide an acceptable margin of safety.
In selecting criteria for the acceptability of electrical exposure, we are
confronted with questions that cannot always be answered with scientific
objectivity. Frequently there are policy or judgment issues that are settled
on the basis of historical precedent, or on some other basis. One example of
a judgment issue is the selection of the population percentile that should be
assumed for a particular threshold reaction. Should a median sensitivity for
a large segment of the population be used because it is "representative" in
some sense? Should a lower percentile be selected to minimize the prob-
ability of an unacceptable exposure? If a safety factor is to be applied, on
what basis should it be selected? The object of this book is not to define the
"correct" criteria for the acceptability of electrical exposure. Rather, its
object is to present the scientific data that will aid in the process of criteria
selection.
The material of this book is the short-term human reactions to electrical
stimulation, with particular attention to those of a detrimental nature. In
this context, the term electropathology is meant to include those human
reactions that might be considered undesirable in some context. It is
necessary to understand the full range of reactions, from just-noticeable
reactions to clearly undesirable or life-threatening ones. The intent of this
book is to cover fundamental principles by which such responses may be
understood.
2
Impedance and Current Distribution
This chapter examines tissue and body impedance as it affects the evalua-
tion of electrical stimulation. It is concerned generally with current levels
exceeding lOOIlA, and voltage levels exceeding 1 V. As a result, many ofthe
special problems associated with characterizing the body's response to
microampere currents or millivolt potentials will not be relevant. For
a thorough discussion of impedance in biological measurements at very
low voltages and currents, the reader is directed to the work of Geddes
(1972).
2.1 Dielectric Properties of Biological Materials

The bulk impedance properties of biological materials are important in
many applied problems of electrical stimulation. They dictate the current
densities and pathways that result from an applied stimulus. In order to
appreciate these bulk properties, consider dielectric properties in a more
general context.
Conductivity and Permittivity

Figure 2.1a illustrates a simple measurement of the electrical resistivity (e)
or conductivity (a = lie) of a substance. Two electrodes of area A contact
a cylindrical block of biological material having length d. The resistance
between the electrodes, given by the ratio VII, will be directly proportional
to the length of the material and inversely proportional to the area of the
electrodes and the material's resistivity in accordance with
ed
R=-=-
d (2.1)
A aA
Inversion of Eq. (2.1) yields a simple definition of the material's resistivity

or conductivity:
12
J. P. Reilly, Applied Bioelectricity
Dielectric Properties of Biological Materials 13
~------~d------~·~I
Area = A
R =pd/A
Q=ohm' cm
~----------~ V ~----------~
~
I
(a) Resistivity for volume material
rm,cm
R = rm/A
rm = ohm·cm 2
Cm = ~F/cm2
(b) Area proportional resistance of membrane

FIGURE 2.1. Electrical resistivity in volume and area materials: (a) resistivity for
volume material; (b) area-proportional resistance of membrane.
1 RA
(}=-=- (2.2)
a d
While the illustration in Fig. 2.1a indicates a simple conceptual method
for measuring resistivity, a more accurate and practical method would use
four electrodes-two to supply a current within the medium, and two
intermediate electrodes to sample the voltage drop (Ruch et aI., 1963).
Resistivity is ordinarily cited in units of Qm or Qcm; the relationship
14 2. Impedance and Current Distribution
between the two units of measurements is e(Qcm) = lOOe(Qm). The in-

verse of resistivity in Qm is conductivity, expressed in units of siemens
per meter (S/m).!
Although the simple concept of conductivity expressed in Eq. (2.2) is
adequate for many calculations, a more complete description of the dielec-
tric properties of a material is often needed. The dielectric characteristics of
a material are described in complex notation and include both conductive
and capacitive properties. The concept of "polarizability" will help to ex-
plain the relationship among the variables and the remarkable electrical
properties of biological materials.
Figure 2.2 illustrates a nonconductive material held between a pair of
parallel-plate electrodes on which a potential of V volts is applied. The
amount of charge accumulated on the electrodes is directly proportional to
the product of applied voltage and capacitance in accordance with
Q=CV (2.3)
where Q is charge in coulombs, C is capacitance in farads, and V is the
potential in volts. If the plate separation, d, is much smaller than its linear
dimensions, the capacitance is
C = cocrA (2.4)
d
where Co is the dielectric constant of free space (8.85 X lO- 12 F/m), and Cr is
the permittivity of the material relative to that of free space (expressed in
dimensionless units, with Cr ~ 1). The relative permittivity of air is very
nearly equal to unity.
The relative permittivity constant, c" is a measure of a material's ability
to become polarized in response to an applied electric field. In the parallel-
plate example, the electric field is simply given by E = Vld. The hypotheti-
cal material indicated in Fig. 2.2 is assumed to be nonconductive; that is, it
lacks free electrons or ions. Consequently, there is no net current flow in
response to the applied field. Nevertheless, the material is assumed to
contain units of separated charge that are bound together into electrically
neutral entities called dipoles. The displaced charge centers represent at-
tractive forces within the dielectric medium, enhancing the internal electric
field as indicated by the arrows in Fig. 2.2. Analogously, the dipoles act as
if they were reducing the plate separation to an effective value of dlc" and
thereby increasing the capacitance.
The capacitive current is given by the change of charge versus time,
dQldt. For a time-varying applied voltage, it follows from Eq. (2.3) that
1= C dV (2.5)
dt
iFormerly called the "mho," the siemen is now the accepted international unit of
conductance.
~--------,d--------~
Area =A
FIGURE 2.2. Dielectric property. Dipoles within material enhance internal field as
indicated by flux lines (arrows).
For a sinusoidal voltage, dVldt = jwV, where w = 2nf,fis the frequency of

oscillation, and j is the phaser operator indicating 90° phase shift. Along
with Eq. (2.4), the capacitive current can be expressed as
I' AV
= JWEOErd (2.6)
When the material contains both dipoles and free charges, its dielectric
description requires complex notation. Figure 2.3 illustrates an equivalent
circuit of a partially conductive material. The total current is the sum of the
resistive and capacitive components 1= IR + 10 which is given by
1= V + C dV (2.7)
R dt
For a sinusoidal voltage, Eq. (2.7) can be expressed as
1= V~A + V(Eo~A}w
or, equivalently
(2.8)
The macroscopic electric field is E = Vld, and the current density is J =

IIA. Consequently, Eq. (2.8) can be written as
J= a*E
v v
FIGURE 2.3. Complex permittivity. A partially conductive dielectric material is sub-
jected to an alternating voltage V, resulting in resistive and capacitive current.
where a* is the complex conductivity given by

a* = a + jweoe r (2.9)
Alternatively, Eq. (2.8) can be written as
1= jweoA (e r - ja)v (2.10)

d weo
By analogy with Eq. (2.6), the term in parentheses is the complex permit-
tivity constant, given by
e* = er _ ja (2.11)
weo
or, in conventional notation,
e* = e' - jeff (2.12)
in which e' = en and e" = a/(weo). Equations (2.9 and 2.11) are equivalent
descriptions of the material. The form indicated by Eq. (2.11) is more often
used; its symbolic representation is given as in Eq. (2.12) and is called the
complex permittivity. Its imaginary part includes the conventional conduc-
tivity, and the real part is the relative dielectric constant.
In the above development, it is assumed that the dipoles orient them-
selves to the internal alternating field, requiring that they reverse their
orientation everyone half cycle of the applied voltage. The dipoles have a
certain degree of inertia and cannot follow the field oscillation if it is too
rapid. Therefore er is at maximum at low frequencies. It drops when the
frequency is raised above a critical value. With very high frequencies,
the dipoles retain random orientation, and the relative permittivity of
the material approaches unity.
The ability of a dielectric material to respond to an applied field can be
expressed in terms of its relaxation time constant Tn or, equivalently, in
terms of the relaxation frequency Wr = 2nt, = lIir• In a typical biological

medium several mechanisms may exist for producing dipoles, each with a
different relaxation time constant. In addition, the presence of boundaries
between regions of differing permittivity will result in an equivalent relax-
ation time constant because of the buildup of charges at those boundaries
(referred to as "interfacial" effects).
Figure 2.4 illustrates an example of the complex permittivity of a typical
biological material (Pethig, 1979; Foster and Schwan, 1996). The dips in the
curve arise from different mechanisms of polarization and are termed di-
electric dispersion. At the lowest frequency, the so-called a dispersion has
been attributed to electronic bilayers in organic molecules, ionic dispersion
processes in micrometer-sized particles, active membrane conductance
phenomena, and other membrane effects. The f3 dispersion is attributed to
capacitive charging of cellular membranes (interfacial effects), and dipolar
relaxation of proteins. The y dispersion represents dielectric relaxation of
water molecules.
Table 2.1 lists experimental values of conductivity (Part A) and relative
permittivity (Part B) for various biological material. The tabulated values
are geometric means of data published by Foster and Schwann (1996).
Useful summaries have also been published by others (Geddes and Baker,
1967; Schwann, 1968; Stuchly and Stuchly, 1980; Stoy et al., 1982).
The relative permittivity of biomaterials at low frequencies can be on
the order of 106• These values are remarkably large in comparison with
10 7 r--.---.--.---~--r-~---'--'---'--'---'10
6
.... 10
-
w
~
.s; 10 5
+=
+=
'E 10 4
CD
a.
CD
> 10 3
+=
«l
Q)
a: 10 2
10
1
10
Frequency (Hz)
FIGURE 2.4. Frequency variation of complex permittivity typical of soft tissue.
.......
00
TABLE 2.1. Dielectric properties of biological materials.
Brain Brain
Skeletal muscle tv
Frequency white gray Whole ......
(Hz) Parallel Perpend. Liver Lung Spleen Kidney matter matter Bone Blood Fat S
"0
(1)
Part A: Conductivity (values given in S/m) 0..
po
::s
(')
10 0.52 0.076 0.12 0.089 (1)
HY 0.52 0.076 0.13 0.092 0.013 0.60 po
10' 0.52 0.08 0.13 0.096 0.013 0.68 0.04 ::s
0..
104 0.55 0.085 0.15 0.11 0.15 0.07 0.013 0.68 (J
10' 0.65 0.47 0.16 0.62 0.24 0.13 0.17 0.014 0.61 ..,Ei
(1)
106 0.71 0.20 0.63 0.38 0.16 0.21 0.017 0.71 ::s
107 0.52 0.024 1.11
....
0.87 0.46 0.63 0.59 0.28 tJ
lOS 0.85 0.65 0.53 0.83 0.80 0.48 0.68 0.057 0.82 0.04 Ij;'
....
10" 1.41 1.03 0.73 1.31 0.97 0.85 1.05 0.05 1.43 0.05 g;
10lD 8.23 7.3 8.1 5.8 8.0 10.0 0.92 9.8 0.35 ....
=
o·
::s
Part B: Relative Permittivity
10 107 106 5 X 107 2.5 X 107
102 1.1 X 106 3.2 X 105 8.5 X 105 4.5 X 105 3,800 1.5 X 10'
10' 2.2 X 105 1.2 X 105 1.3 X 105 8.5 X 104 1,000 2,900 5 X 104
104 8 X 104 7 X 104 5.5 X 104 2.5 X 104 2.2 X 104 4.8 X 104 640 2,810 2 X 104
105 1.5 X 104 2.1 X 104 1.2 X 104 3,260 1.2 X 104 2,500 3,800 280 3,300
106 2,200 1,970 1,450 2,540 670 1,250 87 2,040
107 184 232 357 294 190 309 37 200
10· 68 72 35 78 76 62 81 23 71
10" 55 49 35 51 44 39 46 8 63 5
lO lD 36 37 38 33 25 40 45
Source: Table adapted from Foster and Schwan (1996). Reprinted with permission from CRC Handbook of Biological Effects of Electromagnetic
Fields, Copyright CRC Press, Inc., Boca Raton, FL.
synthetic dielectric materials, for which relative permittivities of 5 to 10 are

typical. Despite the appearance of such large permittivity, the biological
material remains overwhelmingly resistive, as can be seen from the
following example. Assume the following values: a = 0.1 S/m, c, = 106 ,
and f = 100Hz. In accordance with Eq. (2.9), the complex conductivity
is a* = 0.1 + j5.6 x 10-3• In this example, the capacitive component
of current is only 5.6% of the resistive component. In general, the bulk
properties of biological tissue are dominantly resistive.
The conductivity of some biological preparations is markedly anisotro-
pic. Skeletal muscle, for example, is shown in Table 2.1 to be approximately
6 to 7 times more conductive when low-frequency current is orientated
parallel to the muscle fibers, as compared with a perpendicular orientation.
The degree of anisotropicity of animal muscle tissue appears to vary greatly
with the tested species (Chilbert et aI., 1983). Anisotropic conductivity
ratios ranging from about 5: 1 to 10: 1 apply to cardiac tissue, depending on
the method of measurement (Plonsey and Barr, 1986). An anisotropic ratio
of about 10:1 has been attributed to nerve bundles (Nicholson, 1965;
Ranck, 1963). For a critique of anisotropic impedance measurements in
cardiac and other muscle tissue, the reviews of Plonsey and Barr (1986) and
Roth (1989) are recommended.
The anisotropic behavior of muscle tissue can be understood by envision-
ing a collection of individual muscle fibers, as in Fig. 3.21. Current in a
direction perpendicular to the fibers must travel in a circuitous
path, whereas current parallel to the fibers travels in a shorter direct
path-the path length difference accounting for the directional difference
in resistivity. If the frequency of the current were high enough, the
insulation afforded by the cellular membranes would become bypassed
through capacitive coupling, and the anisotropic properties would
disappear.
Cellular Membranes
The previous discussion has dealt with the bulk dielectric properties of
composite biological materials. Characteristics of microscopic biological
components, such as the cellular membrane, may also be important in
applied studies of electrical stimulation. The cellular membrane consists of
a bimolecular lipid structure whose impedance properties are usually ex-
pressed as area-proportional quantities-unit area resistance in Qcm2, and
capacitance in {lF/cm 2 as in Fig. 2.1b. These units stand in contrast to the
bulk resistivity units (Qcm) indicated in Fig. 2.1a for composite materials.
The reason is that the thickness of the biological membrane cannot be
subdivided without altering its basic structure. The capacity of a cellular
membrane generally is in the region 0.5-1{lF/cm2 , resistivity in the range 102
to 104 Qcm2, and relative permittivity around 2.5 (Pethig, 1979). The ionic
permeability is specific to particular ionic species. For excitable membranes
(nerve and muscle tissue), ionic permeability is highly dependent on the

transmembrane voltage, as discussed in Chapter 3.
Skin Depth
The penetration depth of incident electromagnetic energy is often de-
scribed in terms of the material's "skin depth." As described in Chapter 9,
an incident magnetic field will set up eddy currents in a conducting material.
The eddy currents, in turn, create their own magnetic field, which tends to
oppose the incident field and resist its penetration into the material. Conse-
quently, the current induced within the material will drop off in an expo-
nential fashion from the surface. The distance at which the current density
falls to e- I of its surface value is known as the skin depth.
The skin depth of a material of arbitrary conductivity was described in
1888 by Oliver Heaviside (Nahin, 1987):
Jr'
0= ------------.......,..- (2.13)
2,,+#£/2)[11 + (O/2"Je)' - 1
where f1 = f1of1r is the magnetic permeability of the material, C = COCr is its
dielectric permittivity, and f is the frequency of the induced current. The
magnetic permeability of free space or air is f10 = 4n X 1O- 7 H/m; for all
practical purposes, the permeability of biological materials is that of free
space, that is, f1r = 1. For a good conductor, a/(2nfc) » 1, and Eq. (2.13)
reduces to
-1/2
0= ( nff1a ) (2.14)
which is the expression for skin depth found in most engineering texts.
The skin depth of biological materials for frequencies below about
10 MHz is generally much greater than any practically attainable material
thickness. This can be seen by way of example. Assume thatf1r = 1, Cr = 200,
a = 0.1 Slm, f = 10 MHz. Then, from Eq. (2.13),0 = 0.8m. It is only when
the frequency is well above 10MHz that skin depth becomes a significant
consideration in most cases. One interpretation of this result is that for
frequencies below 10MHz, a magnetic field will pass readily through bio-
logical material, and the internal magnetic field differs negligibly from the
external field.
2.2 Skin Impedance

In most situations involving electrical stimulation, current is introduced
through metallic electrode contact with the skin. The total circuit imped-
ance will include contributions from the source, the electrode interface, the
Skin Impedance 21
skin, and the internal tissues of the body. Of these, skin impedance is the
most difficult to characterize. It is nonlinear, time-variable, and depends on
environmental and physiological factors that are usually difficult to control
in an experimental setting. Yet skin impedance is often the primary factor
that limits current flow in the body, particularly where the applied voltage
is moderately low «200V) and where the skin is undamaged.
Detailed Structure of Skin2

The skin consists of various layers, as indicated in Figs. 2.5 and 3.16. An
outer layer (the epidermis) overlays the inner dermis. The epidermis con-
sists principally of keratin, derived from dead cells of the lower layers,
arranged in a flattened and irregular fashion. The bonding strength of these
cells is low. They are constantly flaking away naturally and can be easily
broken or removed. The germinating layer at the boundary of the epidermis
contains a mixture of living and dead cells. The dermis contains living cells
and a great density of blood vessels that are related to both nutrition of the
skin and its thermoregulation. The dermis consists of bundles of collagen
fibrils oriented in all directions, giving it strength and elasticity. The distri-
bution of skin thickness varies greatly for different body areas. The epider-
mis itself ranges from about 10 to more than 100,um. It is typically 10 times
Non dividing
living cells Hair
shafts
} - - Epidermis
Dividing cells----Ioi·
~D"m;'
Nerve endings -_!l~
Blood capillary ---l"'li::,·.Il\
Sweat gland - - = > I !

} - Subcutaneous
tissue
~~<)ii;r
FIGURE 2.5. Structure of the skin.
2See also Edelberg (1971), Harkness (1971), and Tregear (1966).

as thick on the palms of the hands as compared with other body areas and
is endowed with a much greater density of sweat glands.
The corneum (the outermost layer of dead skin cells) is a relatively poor
conductor when dry. But when wet or sweaty, or when bypassed (as with an
injury), the conductivity of the skin can rise dramatically. The contribution
of the corneum to the total impedance can be studied by stripping the skin
with cellophane tape (Harkness, 1971; Lykken, 1971; Tregear, 1966; Reilly
et aI., 1982; Clar et aI., 1975). Figure 2.6 illustrates the drop of skin resistivity
when the corneum is successively stripped away, showing an ultimate drop
by a factor over 300. Impedance drop with corneal stripping has also been
noted with high-voltage spark discharge stimuli (refer to Sect. 2.6). When a
microelectrode penetrates the corneal layer, the resistance drops suddenly,
as noted in Table 2.2 (Suchi, 1954).
107~----T-----~-----r----~----~
10 6
-
C\I
E
(.)
9-
.i!'
·5 10 5
~
·in
e
.2
:u
~
rJ)
103~----~----~----~----~----~
o 0.5 1.0 1.5 2.0 2.5
Mass removed (rngIcm 2 )
FIGURE 2.6. Impedance of skin as related to the mass of corneum removed from dry
skin; sinusoidal current at 1.5 Hz. (Adapted from Tregear, 1966.)
Skin Impedance 23
TABLE 2.2. Impedance layers in the epidermis.

Ball of thumb Forearm
Dry. outer layer thickness (urn) 200 45
Penetration distance for
impedance breakdown (,urn) 350 50
Source: Suchi (1954).
Sweat is chiefly a 0.1 to 0.4% saline solution of sodium chloride, with a

resistivity of 140Qcm at 37°C (for a 0.3% solution). The density of sweat
ducts (per cmZ) is approximately 370 on the palmar and plantar surfaces of
the hands and feet, 160 on the forearm, 750 at the bend of the elbow, 150 to
250 on the breast, and 60 on the buttocks.
Current appears to be conducted in dry skin through discrete channels
beneath a contact electrode. With a multipoint electrode, current is found
to be preferentially conducted at one point (Mueller et aI., 1953). In experi-
ments with electroplating on the skin surface (Saunders, 1974), the pattern
of silver deposition on the arm shows that current is conducted in discrete
channels at a density of about I channel/mm2• Other electroplating experi-
ments reveal that the channel density on the palm is about three times that
on the forearm (Panescu et aI., 1993). These observations approximately
correspond to the density and distribution of sweat ducts on the hand and
forearm. The punctate nature of skin conductance was also explored with
high-voltage discharges (Sect. 2.6).
The sweat ducts form electrical weak points in the epidermis, acting as
conductive tubes into the well-conducting dermis and tissues below. Suchi
(1954) explored the impedance role of sweat ducts using a silver microelec-
trode to scan various parts of the skin. When the electrode touched a duct
filled with sweat, the impedance dropped by a factor of 10 as compared with
adjacent areas of the skin. If the duct was dry, the drop was approximately
a factor of two.
Equivalent Circuit Models

It would be desirable to represent the impedance of the skin and body as an
equivalent circuit model that allows one to determine internal currents for
a wide range of exposure conditions. Unfortunately, the skin is not readily
expressible as a simple passive circuit-it is a distributed electrical system
having markedly nonlinear and time-variant properties complicated by
electrolytic interactions at the electrode interface. Nevertheless, impedance
models are often valid under a sufficient range of circumstances to be of
practical use.
A simple model sometimes used to represent skin impedance is a parallel
network consisting of a capacitor and resistor, followed by a series resistor
(cf. Yamamoto and Yamamoto, 1977; Burton et aI., 1974). The parallel
resistor and capacitor represent the resistivity and capacity of the skin, and
the series resistance the well-conducting subepidermal medium. Evidence
for this simple model can be seen when measuring the current response to
a constant-voltage stimulus pulse applied to the skin as illustrated in Fig. 2.7
(from Lykken, 1971). The response, illustrated in Fig. 2.7b, shows an
initial current spike that is limited by the series resistance, Rs. Afterward,
the current decays to the value limited by Rp + Rs. After the corneum
has been removed by abrading, the current, illustrated in Fig. 2.7c, is limited
by Rs.
A more complete model considers the skin as composed of numerous
layers of cells, each having capacitance and conductance, as illustrated in
Fig. 2.8a (Edelberg, 1971; Lykken, 1971). The individual strings of elements
are meant to represent the parallel paths beneath an electrode. In addition
to the resistance and capacitance elements, the electrical model contains
DC potential sources to account for the observed bulk potential of the skin
of about 15 to 60mV, with the surface negative relative to the underlying
layers. These potentials are extremely small in comparison with the stimu-
lus potentials typically needed for cutaneous electrical stimulation. The
element RB is treated as the body impedance, exclusive of the skin.
Figure 2.8b illustrates a model of intermediate complexity. The element
Ze represents the impedance at the electrode interface; it is seldom deter-
mined explicitly, but rather is lumped together into total impedance. The
parallel RC circuit represents the epidermis, with the element Rn added
in the capacitive branch. It is shown in parentheses to indicate that it is
(a) (b) (c)
FIGURE 2.7. Current response of skin to square-wave constant-voltage pulses: (a)

equivalent circuit; (b) response of intact skin; (c) response of skin with corneum
removed. (Copyright 1971, The Society for Psychophysiological Research. Re-
printed with permission of the publisher and the author from Lykken, 1971.)
Skin Impedance 25
Rp Skin
Spreading
resistance
Body
resistance
Rp
(a) (b)
FIGURE 2.8. More complex impedance models: (a) multilayer model for skin imped-
ance; (b) simplified body impedance model.
frequently ignored in impedance representations. The subepidermal layer

is shown as consisting of a spreading resistance element, R" which is in-
versely related to electrode size, and a body resistance element, R B , which
increases with electrode separation. In most studies of body impedance, Rs
and RB are lumped together into a single resistive term. The element RB is
generally treated as a pure resistance, although in fact it contains a small
reactive component, indicated by the components in parentheses. The reac-
tive component can be ignored for most practical applications. The circuit
representation includes additional terms to account for the second elec-
trode. The total circuit impedance is designated ZT' The circuit models
discussed here are merely approximations that are sometimes suitable; the
individual terms may behave in a more complex fashion than would a
laboratory component, as discussed below.
Area Proportionality of Skin Admittance

From geometric considerations, it might seem reasonable to suppose that
skin admittance is directly proportional to contact area, as long as the
electrode diameter is significantly larger than the skin thickness. Many
investigators report equivalent circuit parameters as being area propor-
tional. Lykken (1971), for example, notes an area-proportional admittance
for DC currents with electrodes of 0.72cm2 and larger and for pulse stimuli
with electrodes 2.4cm2 and larger. Tregear (1966) reports that the product
of impedance and area is constant for wet skin only for electrodes
having areas larger than 2cm2 • For smaller areas, the impedance departs
considerably from inverse area proportionality, with significant differences
between wet and dry skin. Biegelmeier and Rotter (1971) evaluate equiva-
lent circuit parameters for electrodes of 1.5 and 100cm2 (an area difference
of 67: 1); they note a 4: 1 change in Rn + Rs as defined in Fig. 2.8b, and a
25: 1 change in R p , showing that admittance rises more slowly than does
contact area.
The area-dependent portion of the resistive path (Rs in Fig. 2.8b) com-
prises the region where the current still undergoes spherical spreading from
the stimulating electrode. At the point where current no longer spreads out
with distance, the resistance is no longer dependent on electrode area, but
is more a function of electrode separation and body geometry. Thus, R8 in
Fig. 2.8b is taken to be independent of electrode contact area.
Area-proportional parameters are usually determined by dividing admit-
tance by electrode contact area. This simple calculation is confounded by
the fact that the distribution of current beneath a contact electrode may be
very nonuniform, as demonstrated in theoretical (Caruso et aI., 1979;
Rattay, 1988) and experimental (Lane and Zebo, 1967) studies. Figure 2.9
illustrates the theoretical current density for a three-layer conductivity
model representing a layer of skin, fat, and muscle tissue; a two-layer model
consisting of fat and muscle; and a single-layer model consisting of muscle
tissue (Caruso et aI., 1979). For the three-layer model, the calculated
current density near the electrode edge is nearly a factor of 10 greater
than that at the center. For the single-layer model, the current density
from center to edge differs by a factor of about 2.2. Area conductivity is
further complicated by the fact that current travels laterally beyond the
confines of the electrode, such that the effective contact area may be
larger than the physical contact area. Rather than rely on area-proportional
values, contact impedance at high-currents can be expressed more accu-
rately in terms of both the area and perimeter of the contact, as given by
Equation 10.5.
With these caveats in mind, Table 2.3 lists some of the area-proportional
values reported for low-frequency (0- to 65-Hz) skin impedance (Tregear,
1966). Because of the low frequency, the list applies primarily to the resis-
tive component Rp + Rs defined in Figs. 2.7a and 2.8b. The values are a
Skin Impedance 27
1.0r----------r----------r---------~--------~
(a) 3-layer model
(b) 2-layer model

,,
:: ,,
:.
"
. , ,,
f:
\
(c) Single-layer model i: \
," , \
... ~ \
..................................-:;;/1 \ \
,.,," \ ,
-------~~' \
.... ,' ....
~. .......... ..........
......... -..
.... _----
O~--------~----------~---------L----------J
o 0.5 1.0 1.5 2.0
Distance from center, r/r 0
FIGURE 2.9. Current density beneath contact electrode; ro = electrode radius; verti-
cal axis on dimensionless scale. (a) three layer model with skin, fat, and muscle; (b)
two layer with fat and muscle; (c) single layer with muscle. (From Caruso et aI.,
1979.)
TABLE 2.3. Low-frequency resistivity of human and animal

skin.
Body Frequency Impedance
Species Hydration Location (Hz) (kQcm2)
Human Dry Arm 0 100-1,000
Human Dry Arm 1.5-10 600-1,200
Human Dry Fingertip 0-1 120-130
Human Dry Palm 30-65 60-80
Human Wet Forearm 1.5 880
Pig Wet Flank 1.5 12
Rabbit Wet Flank 1.5 18
Source: Adapted from Tregear (1966); data compiled from a variety

of sources.
compilation from a variety of sources and do not necessarily represent data

measured in a consistent manner.
Skin Capacitance
A variety of testing methods indicate that the skin's capacity lies in the
range 0.02 to 0.06,uF/cm2 (Edelberg, 1971); by conventional calculations,
this is considered very high. Consider, for example, a corneum thickness of
10,um and a dielectric constant of 2.5 for biological membranes. With these
values, the capacity is calculated to be about 2 x 1O- 4 ,uF/cm2-a small
fraction of the experimental values.
Skin capacitance will be affected by polarization capacitance-the phe-
nomenon of stored charges that appear around an electrode in an electro-
lytic medium, forming an effective ionic capacitor that is dependent on
excitation frequency (Schwan, 1966). Lykken (1971) argued that the appar-
ent frequency-dependent property of skin capacitance is simply a conse-
quence of the choice of equivalent circuit and that a representation with a
resistively coupled capacitor (C p and Rn in Figure 2.8b) will demonstrate
a fixed value versus frequency.
Much of the skin's capacity lies in the corneum. If the corneum is stripped
away, the skin capacitance is reduced with each successive stripping opera-
tion (Edelberg, 1971). When the corneum is removed entirely, the capacity
drops to a small fraction of its intact value (Lykken, 1971; vanBoxtel, 1977).
These observations are counter to a model in which the corneum is simply
a dielectric separating the electrode and the underlying conductive dermis.
If that model were correct, we would expect to see an increase of capaci-
tance as the thickness of the corneum is reduced. Biegelmeier and Miksch
(1980) postulate that the skin's capacity is derived from membrane capaci-
tance of the sweat gland duct. However, this explanation does not account
for the leading-edge current spikes that are observed when constant-
voltage pulses are applied to the sweat-gland-free skin of Rhesus monkeys
(Bridges, 1985).
An alternative explanation for the skin's high capacity was provided by
Tregear (1966), who treated the skin capacitance as being due to individual
cell membranes as in Fig. 2.8a. If we assume that each cell's membrane may
have a capacity as large as 5,uF/cm2, and acknowledging that each cell
accounts for two membrane layers, then 200 cell layers could account for a
capacitance of 0.05,uFlcm2. A related mechanism that might account for
skin capacitance has been described as an ionic bilayer surrounding indi-
vidual corneal cells (Clar et aI., 1975).
Time-Variant and Nonlinear Aspects of Skin Impedance

When an electrode is placed on dry skin, impedance gradually falls with
time, as noted in Fig. 2.10 (Mason and Mackay, 1976). During the first 2 min,
Skin Impedance 29
150 1/8" dia*

150
130
130
C- 1/4" I.D.,
oX
110 9/16" O.D.
Q) 110
0
c
C1S
U)
·iii
90
Q) " 90 0 0.5 1.0 1.5 2.0
c:
' .....
'/-------~---
70
Wet skin
50
o 10 20 30
Time (min.)
FIGURE 2.10. Variation of skin resistance with time. Concentric electrode as illus-
trated (inner diameter not clearly described in source). Wet skin treated with tap
water. Inset shows expanded scale for first 2min. Stimulus current: 1 rnA. (Adapted
from Mason and Mackay, 1976, ©1976 IEEE.)
the impedance undergoes rapid fluctuations superimposed on a relatively

sharp overall drop (see insert). Thereafter, the impedance continues to
drop, reaching an apparent asymptotic value after some 20 or 30min. This
effect may be explained by gradual hydration of the corneum from sweat
buildup beneath the electrode. In the same experiments, skin pretreated
with conductive electrode paste started with an initially lower impedance,
rose somewhat, and attained a final value only slightly below its initial value
after about 1 min.
Past studies have shown that the electrical properties of the skin are
nonlinear. Edelberg (1971) suggests that the linear region for current den-
sity lies in the range below 2mNcm2• Lykken (1971) concludes that it is
applied voltage, and not current density, that determines an upper limit of
linearity around 2V. Biegelmeier and Rotter (1971) note variations in
equivalent circuit values of over 2: 1 for voltage changes from 4 to 45 V.
Stevens (1963) defines the DC current conducted in dry skin in terms of
1= aV+ bV2, in which the V2 term dominates above 3V.
Nonlinear phenomena in living skin was studied by Grimnes (1983a). He
hypothesized that the time scale of impedance changes could be explained
by either a nondielectric mechanism or a dielectric breakdown mechanism,
depending on the applied voltage in the range of 600 to 1,OOOV. In this

work, the current was limited to a few microamperes, and the results may
not necessarily apply directly to higher current densities.
A dramatic display of skin nonlinearity is seen in the sudden break-down
of impedance above a critical voltage. Dielectric breakdown of dry, excised
corneum has been observed by Mason and Mackay (1976) at 600V for a
15-,um-thick sample and at 450V by Yamamoto et al. (1986). Dielectric
breakdown of intact skin from high-voltage spark discharges appears at
similar voltages, as described in Sect. 2.6.
With intact, living skin, a sudden impedance breakdown has been ob-
served at significantly lower voltages than that needed to break down
excised corneum. A discrete thermal model has been developed to account
for nonlinear skin impedance, including breakdown (Panescu et aI.,
1994a,b). Pricking pain has been associated with the impedance breakdown
(Nute, 1985; Gibson, 1968; Mason and Mackay, 1976; Mueller et aI., 1953).
During breakdown, current becomes concentrated in discrete channels.
According to Mason and Mackay, microscopic examination of the electrode
site reveals small blackened punctures under a dry-skin electrode, but no
such evidence when the skin has been pretreated with water or electrode
paste.
The major source of skin nonlinearity lies in the corneum. Equivalent
circuit values Rp and Rs were evaluated by van Hoxtel (1977) from the
transient current response to constant-voltage pulses. With intact skin, Rp
varied significantly with the applied voltage, whereas Rs remained indepen-
dent of the stimulus level (up to a voltage of 40V, and 40mA final current).
With the corneum removed, both Rp and Rs remained independent of the
stimulus level.
Figure 2.11 illustrates nonlinear impedance response for pulses of 5-,us
duration (Saunders, 1974). At these short durations, the corneal layer is
probably bypassed through capacitive coupling to the dermis. However, the
skin exhibits marked nonlinear properties, as noted in the knee of the
curves at the higher current and voltage values-nonlinear breakdown
occurs with voltages from about 150 to 250V.
Freiberger (1934) describes breakdown as occurring at small contact
areas within a fraction of a second for voltages above 100V. Hiegelmeier
and Miksch (1980) noted breakdown above 200V for large hand-held
electrodes (refer to Sect. 2.3). With small electrodes and voltages above
100V, breakdown may occur within fractions of a second with dry-skin
contact.
Nonlinear response to sinusoidal voltage shows up as distortions from
a sinusoidal current waveform. Short-term distortions appear as an
instantaneous nonlinear response within individual cycles; longer term
nonlinearities appear as a gradual reduction of impedance. Figure 2.12
illustrates impedance breakdown with 50-HzAC stimulation on a relatively
long time scale (Freiberger, 1934). An initial precipitous drop occurs within
Total Body Impedance: Low-Frequency and DC 31
10 50
Current (rnA)
FIGURE 2.11. Nonlinear voltage/current relationship for 5-#s cathodic pulses applied
to abdomen; resistance (R) applies to linear regions of curves; d = electrode diam-
eter. (Adapted from F.A. Saunders, 1974 in Conference on Cutaneous Communica-
tion Systems and Devices, pp. 20-26, reprinted by permission of Psychonomic
Society, Inc.)
a small fraction of a minute, a second drop during the first minute, and a
final gradual drop during the next 7 or 8 min.
2.3 Total Body Impedance: Low-Frequency and DC

The previous section showed that body impedance depends on a variety of
factors, including size and location of contact electrodes, skin hydration,
and applied voltage. These factors were studied by Biegelmeier (1985a).
Measurements consisted of the peak and root-mean-square (rms) voltage
and current, dissipated power, and the phase angle between the voltage and
current applied to the subject. The subject grasped cylindrical electrodes of
8cm diameter, providing a hand-to-hand current path. The contact area on
each hand was estimated to be 82cm2 • After covering the cylindrical elec-
trodes with insulating material, smaller hand contact areas consisted of 12.5,
~ 15
I
25
20 10
<-E
I B (left scale) a
::.
-C 15 <D
0
c:
...<D co
-g
::J
<.) a.
E
10 ." 5
...... ... ,
" ZT (right scale)
'-
5 ~----------------
o~~~--~~~~--~~~--~~~--~~~o
o 5 10 15
Duration of current (min.)
FIGURE 2.12. Breakdown of the skin at 50V. Forearm-to-forearm current path;

circular electrodes, 12cm2; dry skin. (From Freiberger, 1934.)
1.0,0.1, and 0.01 cm2 on each hand. Foot electrodes consisted of copper foil
inserted into the shoes.
Figure 2.13 illustrates example current and voltage waveforms at the
onset of a stimulus having a steady-state voltage and currene of 200V and
118mA, respectively (Biegelmeier, 1985a). Corresponding peak values
were 280V and 370mA. Total body impedance (ZT) and internal body
resistance (R B ) were determined by
ZT = VT (2.15)
IB
(2.16)
3When referring to sinusoidal voltage and current, cited magnitudes indicate rms
values, unless specifically stated otherwise. This convention applies throughout this
book.
where V T and IB are steady-state (rms) values of applied voltage and body
current; V TP and V BP are the peak values of voltage and current. In the
example of Fig. 2.13, ZT = 1.7kQ, and RB = O.7SkQ.
Figures 2.14 to 2.16 illustrate the relationship between impedance and the
applied voltage for large-area contacts and different electrode placements
(from Biegelmeier, 1985b). Figure 2.14 applies to hand-to-hand contacts
and indicates ZT and RB separately; ZT is further subdivided into wet and
dry contacts. Each measurement point is the average of six procedures with
a single subject. The dry-contact data are indicated by filled circles; the
error bars show the mean and extreme measurements. In Fig. 2.14, the
wet-contact conditions were obtained by soaking the bands in either tap
water or a saline solution prior to the measurement. The tap water
treatment lowers the impedance somewhat relative to the dry condition;
.".----~ ........ "

,,
,
'" , I rms = 118 rnA
o 5 10 15 20
Elapsed time (ms)
FIGURE 2.13. Oscillograms of applied voltage (upper trace) and body current (lower
trace), with contact at the approximate peak of the voltage waveform; applied
voltage = 200Vrms. (From Biegelmeir, 1985a.)
5r---~---r--~r---~---r--~----.----r--~
4
I Average and range
Dry
)t" ....
Tap;-....
water
0.3% Saline
cr---o--~--~ __ ~
~-----------o
}
Zr
o 25 50 75 100 125 150 175 200 225

Voltage (V)
FIGURE 2.14. Impedance versus voltage for large-area hand-to-hand contacts.

(From Biegelmeler, 1985b.)
I Average and range
a
~
2
B
c:
C'II
~
E 1
A---~---n---n---b---A-_~---6RB
o 25 50 75 100 125 150 175 200

Voltage (V)
FIGURE 2.15. Impedance versus voltage for large-area hand-to-feet contacts. (From
Biegelmeier,1985b.)
3.---.----.---.,---.----.---.----.---.----,
I Average and range
o 25 50 75 100 125 150 175 200

VoHage (V)
FIGURE 2.16. Impedance versus voltage for large-area foot-to-foot contacts. (From
Biegelmeier,1985b.)
treatment with saline lowers the impedance much more. At 200V, there is
little difference between wet and dry electrodes.
Figures 2.14 and 2.15 also show internal body resistance, R B , for the
indicated paths. The value of RB drops only slightly as the applied voltage is
raised from 25 to 200 V, dropping to 650Q for the hand-to-hand path and to
550 Q for the hand-to-foot path. These values are reasonably close to the
value of approximately 500 Q determined for RB from high-frequency
measurements (refer to Sect. 2.4), and also with high-voltage capacitive
discharges (refer to Sect. 2.5).
The cited values of RB are similar to the resistance of a solution of NaCI
of physiological concentration and of geometric dimensions similar to that
of the body (Biegelmeier, 1986). If the resistivity of the NaCI solution is
taken to be 80Q cm, Eq. (2.1) indicates that the resistance of a cylindrical
cross-sectional area of 25cm2 and of length 150cm (for two arms in series)
is approximately 480 Q. This simple calculation correlates quite well with
measured values of R B •
Figure 2.17 illustrates further results of Biegelmeier's (1985a) experi-
ments showing the relationship between total impedance and applied
voltage for various contact areas. At 25V, impedance is nearly inversely
proportional to contact area. As the voltage is increased, impedance drops
rapidly, and area dependence is markedly reduced. At 225 V, total imped-
ance varies by only about 4: 1 for a change in electrode contact area of
8200:1.
Impedance-versus-voltage characteristics depend very much on the body
locus of skin contact. With tests on cadavers, Freiberger found impedance
a
~
Q)
0
c:::
ctI
-0
Q)
0.
.§
>.
-0
10
0
co
1 --------------
0.1~--~--~----~--~--~----L----L--~~--~--~
o 50 100 150 200 250
Voltage (V)
FIGURE 2.17. Total body impedance as a function of applied voltage for various
contact areas; dry hand-to-hand electrodes. The various curves are identified by the
contact area on each hand. (From Biegelmeier, 1985b.)
of dry contacts to drop precipitously for voltages above 200 V on the palmar
and plantar surfaces of the hands and feet. On the forearm, the point of
impedance breakdown was about 50V.
Figure 2.18 illustrates measurements of Freiberger (1934) showing im-
pedance values for large-area hand-to-hand or hand-to-foot contacts at
50Hz. The illustrated values are reasonably consistent with those of
Biegelmeier (Fig. 2.17) within the voltage range 25 to 200V. In Fig. 2.18,
impedance declines with voltage, reaching a minimum plateau somewhat
beyond 500V. Freiberger attributed the data in Fig. 2.18 to living persons,
although the measurements above 50 V were actually conducted on corpses,
and those below 50 V on living persons. He adjusted the measurements on
corpses to living persons using a correction procedure that will be described
presently.
6~--~~--~----~-----r-----r----'-----.-----'
Lower limit
100 200 300 400 500 600 700 800

Voltage (V)
FIGURE 2.18. Total body impedance attributed to living persons; large-area hand-to-
hand or hand-to-foot contacts. Measurements above SOY conducted on cadavers,
and corrected for living persons. (From Freiberger, 1934.)
Distribution of Current and Impedance Within the Body

The pathway and density of current within the body, as well as the total
impedance presented to the stimulating device, depend critically on the size
and placement of the electrodes. Figure 2.19 (from Roy et aI., 1986) illus-
trates the current distribution measured for electrodes of various sizes
within a saline-filled tank. A large (400-cm2) "indifferent" electrode and a
smaller "active" electrode were separated by 16cm. The current density
was measured as a function of distance from the active electrode; the
applied current was 3 rnA. The current density near the active electrode
depends substantially on electrode size. But when measured farther into the
medium (beyond 6 cm in this example) the current density depends but
little on electrode size. This is an illustration of "spreading" resistance,
mentioned in connection with Fig. 2.8. When the electrodes are widely
spaced on the body, the total impedance will depend on the separation of
body contact points, plus the skin and spreading resistance. For high volt-
ages, large contact areas, and hydrated skin, the total impedance will be
dominated by internal body impedance.
A substantial body of knowledge concerning body impedance was pro-
vided in early experiments by Freiberger (1934). Even today, after the
passage of more than 60 years, we can still learn much from his remarkable
work. Much of Freiberger's research was conducted on cadavers. He deter-
mined the contribution of skin impedance by removing the cornium from
the electrode site using a process of inducing heat blisters on the skin. The
impedance after this treatment was called "body" impedance. Figure 2.20
illustrates the distribution of body impedance for various current paths. The
numbers indicate internal impedance for various electrode placements as a
percentage of the total hand-to-foot impedance. These data apply specifi-
cally to the internal component, analogous to RB in Fig. 2.8. Roughly 50%
of the internal impedance for hand-to-hand or hand-to-foot contacts resides
in the wrists or ankle; these high-impedance regions are dominated by
relatively poorly conducting bone and ligament. The impedance distribu-
tion of Fig. 2.20 is consistent with the investigations of Taylor (1985), who
used high-voltage capacitive discharges on living people (see Sect. 2.6). The
internal body impedance for hand-to-hand or hand-to-foot paths as mea-
sured by Freiberger averaged 1,000 Q for 60 corpses-male and female, of
10
100 em 2
5 10 15
Electrode separation (em)
FIGURE 2.19. Normalized current density in medium with resistivity of 1,860 Q cm;
l,rlle is current density in medium divided by average current density of electrode.
Indifferent electrode area is 400cm2• (From Roy et aI., 1986.)
\
\
\
\
I
I
I
FIGURE 2.20. Distribution of internal impedance of the human body. Numbers

indicate percentages of total internal impedance for hand-to-foot contacts.
(Adapted from Freiberger, 1934.)
various ages and body structure. This value is somewhat larger than the
previously discussed measurements (500--750Q) in living persons.
Figure 2.21 shows a simplified representation of body impedance for
evaluation of electrical accidents in which the current path is from one
extremity to another. If the contact area is large and if the skin is sweat-or
saline-soaked, the impedance to low-frequency currents is largely resis-
tive-a typical example value would use ZT = 1,000 Q. As a further approxi-
mation, we can consider the impedance in each extremity (ZTE) as being
equal, and the impedance of the body trunk (R BT) to be negligible in
comparison. The internal body impedance across both extremities lies in
the range 500 to 750 Q. If ZT = 1,000 Q, then the skin impedance at each
electrode must be 125 to 250 Q. Using the model of Fig. 2.21, the example
indicates that approximately 1,000 Q would be measured across any combi-
nation of two extremities. If a person were to grasp equipment with both
hands while sitting on a conductive floor, the impedance would be reduced
to 250 Q. For other current paths, calculation of total impedance can be
Rs
FIGURE 2.21. Simplified impedance model for current paths across extremities-wet
skin, large-area contacts. Example values: ZT = 1,OOOQ; ZTE = SOOQ; RBE = 2SO-
37SQ; Rs = 12S-2S0Q; RBT = SOQ.
carried out by reducing RB in accordance with Fig. 2.20. If the skin is

damaged, the value of Rs would also be reduced.
Current Through the Heart

Freiberger applied 50-Hz currents to corpses to determine the percent of
current conducted through the heart. After opening the chest of a corpse,
he encircled the heart with a ring-shaped current transformer, and mea-
sured the current flowing through its cross-sectional area (51 cm2) as a
function of electrode placement on the body. Measurements made in this
manner are sensitive only to current in a direction along the long axis of the
heart. The percentage of current flowing through the ring transformer,
listed in Table 2.4, is quite small for current paths through the limbs. The
greatest percentage was 8.5% for a right-hand-to-feet current path. For a
foot-to-foot path, the current through the heart does not surpass 0.4% of
the total body current. It is likely that a more favorable orientation of the
ring transformer was responsible for the larger currents measured from the
right-hand in comparison with the left-hand contact. Tests similar to those
of Freiberger have been carried out on dogs (Kouwenhoven et aI., 1932),
with similar results.
Applicability of Measurements on Corpses

Tests on biological tissue reveal impedance changes after death. The dielec-
tric permittivity of frog muscle, for example, is reported to decrease by a
factor of 2 at 10Hz within 2.5h after death, but changes above 100Hz are
much less pronounced (Schwan, 1954). Permittivity changes would have
little effect on tissue conductivity, since most tissue is dominantly resistive,
even at low frequencies.
Freiberger measured systematic increases in body resistance with time
after death. The internal body impedance increase after 5 h, for example,
was about 20%. More gradual increases continued up to 30h, at which time
the increase from time of death averaged about 45 %. These increases were
TABLE 2.4. Percentage of longitudinal

heart current for various current paths.
Heart current (%)
Current Path Min Av. Max
Hand to hand 1.9 3.3 4.4
Left hand to feet 1.2 3.3 5.1
Right hand to feet 4.8 6.7 8.5
Foot to foot <0.1 0.4
Head to feet 4.8 5.5 5.9
Source: Data from Freiberger (1934).

attributed almost entirely to the cooling of body temperature to room

temperature after death. Freiberger determined that the resistivity of physi-
ological saline varies with temperatures in a manner that would account for
the observed change in body impedance. Similar changes after death in
the resistivity of muscle tissue of animals have been observed; resistivity
changes in fatty tissue, however, are found to be minimal (Chilbert et al.,
1983).
Changes in the processes of blood circulation and sweat activity after
death should have the greatest influence on skin impedance. Considering
that skin impedance is most significant at low voltages, we expect the
impedance difference between living persons and corpses to be small when
the voltage is raised above 200V. Skin hydration should playa prominent
role in the impedance changes after death. It is very difficult to measure
skin hydration in a meaningful way. Freiberger instead studied the correla-
tion of impedance with body temperature after death, and used a tempera-
ture correction factor to relate the impedance of corpses to living persons.
These correction factors were applied in Fig. 2.18 to the measurements on
corpses for voltages above SOY; below SOY, the data were derived from
measurements on living persons.
Statistical Distribution of Impedance

Table 2.5 provides statistical data from Swiss and Austrian measurements
at 50Hz (Biegelmeier, 1985c). Data are shown for total impedance (ZT)
with large-area contacts on dry skin, and also for the body impedance
component (RB)' Freiberger also determined the statistical distribution of
50-Hz impedance on corpses above SOV and on living persons below SOY.
The statistical data of Freiberger have been adapted by the International
Electrotechnical Commission (1EC) to produce an impedance model appli-
cable to living persons. (Biegelmeier, 1986). Table 2.6 presents the IEC
model at the 5, 50, and 95 percentile ranks, and for voltages ranging from 25
to 1,OOOV. The asymptotic values are the presumed values of internal body
impedance (RB)'
Freiberger tested 25 subjects using direct or 50-Hz alternating voltages
ranging from 5 to SOY. The current path was hand to hand, with large
TABLE 2.5. Statistical impedance measurements.
Voltage (V) Number Avg. (kO) S.D. (kO) Avg. (kO) S.D. (kO)
25 100 3.52 1.40 0.78 0.11
15 50 3.72 1.12 0.64 0.10
Large area contacts, 50Hz, dry skin, hand-to-hand.

Data from Swiss and Austrian measurements, as summarized by Biegelmeier
(1985c).
TABLE 2.6. Statistical data for total body

impedance (Zr) adopted by the IEC for
50-Hz currents.
Total body
impedance (Q)
at the indicated
percentile rank
Touch voltage (V) 5% 50% 95%
25 1,750 3,250 6,100
50 1,450 2,625 4,375
75 1,250 2,200 3,500
100 1,200 1,875 3,200
125 1,125 1,625 2,875
220 1,000 1,350 2,125
700 750 1,100 1,550
1,000 700 1,050 1,500
Asymptotic value 650 750 850
Large area contacts, hand-to-hand, dry skin.

Source: Data from lEe (1984).
cylindrical electrodes on dry skin. At 5V, the average impedance was about
4.8kQ for DC voltages and 3.8kQ for AC voltages. At 50V, the impedance
dropped to about 70% of the 5-V values. Maximum and minimum imped-
ances at DC were typically a factor of 2 and 0.5 times the averages. At AC,
the extremes were 1.5 and 0.53 times the averages. These extremes prob-
ably represent approximately the 5 and 95 percentile ranks of a cumulative
distribution. The mean DC impedance exceeded the AC value by a typical
factor of 1.25, although there was considerable overlap in the distributions.
Lower AC values are attributed to 50-Hz capacitive coupling through the
high-impedance corneal layer.
Statistical impedance data were developed by Underwriters Laboratories
(UL) for DC currents and low voltages (approximately 12V) (Whitaker,
1939). Figure 2.22 illustrates statistical distributions for 40 adults; Fig. 2.23
applies to 46 children. Table 2.7 summarizes the data of adults and children.
Adult weight ranged from 45.4 to 94.4kg (median = 62.2 kg), and age
ranged from 18 to 58 years (median = 30 years). For children, weight
ranged from 14.1 to 58.1kg (median = 31.8kg), and age ranged from 3 to 15
years (median = 10 years). Hand electrodes consisted of two No. 10 AWG
bare copper wires twisted together; the foot electrodes were large copper
plates. Under "wet" conditions, the hands and feet of subjects were initially
soaked in a 20% NaCI solution. Preliminary tests showed that resistance
was independent of measuring current in the range of 1 to 15 rnA, provided
a constant contact area and pressure were maintained and the subject's
hands were wetted in the solution before each measurement. The test
voltage was adjusted up to a maximum of 12 V to maintain a current of
99
2 hands! Hand!
98 ,2 feet
,i/',
2feet~ Hand! Hand!
Hand!
95 I ' Vhand 2 feet hand
I
I I ,
: I
, ,
90 I
,,,# ,,, ,,,

I I ,
,,, I, I
80
CD
~ I
'E
CD
60 , : I
,t
~
CD
a.
4I II'
,
CD
> I
~ 40
,,,
I
,I
I
"S I
E I
,,, ,I• ,,,

::::I I
() I
I
,,, ,4 ;,
20 I I
10 Dry skin
, I
I
----- Saline treated
5
,,
• J ./
I
f /
/
N =40
Weight = 64.5:!::10.9 kg
2 (mean:!::cr)
1
0.5 10 50
Resistance (k Q)
FIGURE 2.22. Cumulative distribution of DC resistance of adults, with various elec-

trode paths. Hand electrodes = 2 No. 10 Awg twisted wires, feet on plate, I = SmA.
(Data from Whitaker, 1939.)
1 rnA with children and 5 rnA with adults, except when body resistance was
too great to allow these values.
The curves plotted in Figs. 2.22 and 2.23 are approximately straight lines,
which indicate the log-normal distribution on this plotting format. The
slopes of the distribution curves are substantially less for wet than for dry
conditions, and dry-skin slopes are greater for children than for adults. In
general, children's resistance is greater than that of adults. Apparently, the
shorter current paths of children is more than offset by their reduced
volume. This can be appreciated by a simple calculation treating the limbs
as conducting cylinders of length L and cross-sectional radius r. According
Impedance at Higher Frequencies 45
99
98 2 hands/-__-.~l ,
2 feet I I
I I
95 Handl ______ f " I
2 feet ~ I
90 I j II I
Handl~'
hand .,
/
II
Cll
CI
80
4
III (
,
E
ctI
I
I 'I l
I
Cll
~
60 I I I
Cll
Q. I I 4
Cll
> ,f #' II
~ 40
"'S
, f
,
I
I
E 4 , #
~ I I I
u I I I
20
I
f ~ I
'I
, I
''II
10
I
I
4
I,I Dry skin
----- Saline treated
5 , I I
N =47
# 4 I
I I I
4 J 4 Weight=32.3±11 kg
2 (mean±cr)
1
0.5 1 10 50
Resistance (k Q)
FIGURE 2.23. Cumulative distribution of DC resistance of children, age 3-15yrs.

Electrodes as in Fig. 2.22, I = 1mA. (Data from Whitaker, 1939.)
to Eq. (2.1), the limb resistance would be calculated by R = L(}/(nr). If we

assume that the ratio Ur is independent of body size, then it is concluded
that resistance would vary inversely with r, or equivalently, inversely with
body weight to the 113 power.
2.4 Impedance at Higher Frequencies

Skin impedance decreases as the frequency of current is increased, as noted
in Table 2.8 (Schwan, 1968). The capacitive component of skin impedance
is primarily responsible for its frequency dependence. Figure 2.24 illustrates
TABLE 2.7. Statistical summary of DC Body impedance measurements.

(Data listed in kQ).
Adults Children
5% 50% 95% 5% 50% 95%
A. Dry conditions
Hand/hand 6.96 11.45 15.69 4.04 14.35 51.10
Hand/two feet 2.62 4.00 6.51 2.62 5.70 18.72
Hand/two feet 2.00 4.25 12.68
B. Wet conditions
Handlhand 1.28 1.86 2.45 1.70 2.55 4.47
Hand/two feet 0.93 1.20 1.67 1.43 1.80 3.02
Two Hands/two feet 0.63 0.84 1.16 0.90 1.30 2.04
(a) Hand electrodes: two No. 10 Awg twisted copper wires.

(b) Voltage: 12V DC; current -lmA (children), -5mA (adults).
(c) Wet conditions apply to treatment with 20% NaCI solution.
(d) Children's ages: 3-15 years; adults ages: 18-58 years.
Source: Data from H. B. Whitaker (1939).
the frequency dependence of impedance over the range 0 to 2,000 Hz,

using large hand-to-hand electrodes on dry skin (from Biegelmeier, 1986).
At the highest frequency, impedance is reduced to about 750Q, a value
approximately that for the internal body impedance with hand-to-hand
contacts.
Figure 2.25 illustrates the measurements of Osypka (1963) in the
frequency range 0.3 to 100kHz, using low voltages (approximately 10V)
and large copper-cylinder hand-to-hand contacts. The test results with dry
and wet hands show that above 5 kHz, the effects of skin hydration become
negligible. Table 2.9 shows impedance measurements at 0.375 and 1 MHz
(Schwan, 1968). The hand-to-hand data are similar to the values for internal
body impedance as discussed in Sect. 2.3.
TABLE 2.8. Dry-skin impedance versus

frequency.
Frequency Magnitude Phase
(kHz) (Qcm2) angle (deg)
1 14,000
5 3,000 -70
10 1,800 -65
20 1,000 -55
50 500 -30
100 300 -20
200 250 -10
Phase angle given by tan- 1 X/R, where X is

capacitive reactance, and R is resistance.
Source: Schwan (1968).
Impedance at Higher Frequencies 47
a
::.l- 3
N
§
co 2
IE
o~r~~~~~~~----~~~~~~~--~--~
o 20 40 100 200 400 1000 2000 4000

Frequency f (Hz)
FIGURE 2.24. Frequency dependence of impedance with large electrodes, dry hand-
to-hand current path. Applied voltage = 25Vrms. (From Beigelmier, 1986.)
5
- - Dry skin
~
---
Q)
()
- - - . Wet skin
c Hand/hand
CO
"0
Q)
a.
.§
>- 0.5
"0
0
.0
(ij
;§ /
2 hands/2 feet
Hand/2 feet
0.1
0.3 0.5 1 5 10 50 100
Frequency (kHz)
FIGURE 2.25. Total body impedance for large-area contacts in the frequency range
0.3-100kHz. (From Osypka, 1963.)
TABLE 2.9. Body impedance at 0.375 and IMHz.

Impedance
Area f= 0.375MHz f= IMHz
Electrode location (cm2) (Q) (Q)
Hand to hand 90 475 460
Finger to other arm 10/65 500 470
Across left arm 32 34 21
Across elbow joint 32 37 21
Across shoulder joint 32 47 31
Across neck 32 36 18
Forehead to neck 32 82 57
Chest to back 150 31 20
Across thorax 150 29 19
Right wrist to left leg 75 248 234
Left wrist to right leg 75 274 266
Area applies to each electrode, except for finger/arm, where

two different electrode sizes were used.
Source: From Schwan (1968).
Chatterjee and colleagues (1986) measured impedance in the range

of 10kHz to 10 MHz. Adult subjects numbered 367 (170M, 197F). The
hand electrode was a brass rod of diameter 1.5 cm; subjects stood barefoot
on a copper plate. Prior to measurements, the subject's hand was moistened
with 0.9% physiological saline. The subject's feet were not treated with
saline. Figure 2.26 shows results from these experiments. The authors as-
sumed that impedance was related to body size, rather than sex per se, and
that impedance was inversely proportional to subject's height. The phase-
angle data show that the impedance is largely resistive in the indicated
frequency range. The magnitude of impedance drops steadily with increas-
ing frequency, to a value nearly equal to the internal body impedance
reported in Sect. 2.3 for hand-to-feet contacts.
2.5 Impedance Through Foot Contact

The current in an electrical accident may travel to ground through a person's
feet, and the path resistance will include the contact resistance of the feet
against the ground, and the resistance of the footwear. A conservative
assumption treats the feet as bare or wearing conductive shoes-
accordingly, the feet are treated as metallic discs (IEEE, 1986), for which the
contact resistance can be calculated from the equations of Sunde (1968) as
(2.17)
(2.18)
Impedance Through Foot Contact 49
where Rf is the contact resistance of a single foot, (} is the resistivity of the

surface material, r is the equivalent radius of the foot print, RM is the mutual
resistance of the two feet, and df is the separation distance of the two feet.
We are often interested in the resistance of the two feet in human contact
accidents. A pathway from a hand to two feet would involve the two feet in
parallel. Alternatively, one may be interested in a current path where the
Magnitude
§: 600
CD
'0
.a
'E
f6>
E 500
CD
g
CIS
-g
c.
£ 400
300
-14
-12 Phase angle

(i)
CD
....
CD
Cl
CD -10
~
CD
Cl
r::::
CIS -8
CD
III
CIS
.s::.
a.. -6
-4
10 10 2
Frequency (kHz)
FIGURE 2.26. Magnitude and phase of impedance measured on 367 subjects, hand-
to-feet contact. Hand electrode: 3.5-cm rod; wet hand contacts; bare feet on plate.
Vertical bars show standard deviations: (From Chatterjee et aI., 1986 © 1986 IEEE.)
TABLE 2.10. Resistivity of several materials-

uniform composition.
Resistivity
Surface (Qm)
Generic soil types
Wet, loamy soil 10
Moist soil 100
Dry soil 1,000
Bed rock 10,000
Specific surface types
Concrete, moist 30-60
air dry 910
Gravel, dry 106
wetted w/ground water 8,500
wetted w/salt water 24
Source: IEEE (1986); Hammond & Robson (1955).
two feet are in series with a potential difference on the ground, as with so-
called "step potentials" that arise from current flow within the earth. The
contact resistance of two feet can be calculated by
(2.19)
(2.20)
where R Zfs and R zfp are the resistance of the two feet in series and parallel
respectively.
Table 2.10 lists resistivity of several surface materials (IEEE, 1986).
Contrary to common expectations, concrete can actually be a relatively
good conductor-being a hygroscopic material, it absorbs moisture if it is
in exposed to water, or is in contact with moist soil. Tables 2.11 and 2.12
TABLE 2.11. Resistivity of soil by

water content.
Resistivity (Qm)
Water content Top Sandy
(% by weight) soil loam
0 >10' >10'
2.5 2,500 1,500
5 1,650 430
10 530 185
15 190 105
20 120 63
30 64 42
Source: IEEE (1982).

Impedance Through Foot Contact 51
TABLE 2.12. Variation of soil resis-

tivity with temperature (Sandy
loam, 15.2% moisture content).
Temperature CC) Resistivity (Qm)
20 72
10 99
o (liquid water) 138
o (ice) 300
-5 790
-15 3,300
Source: IEEE (1982).
show that temperature and especially water content can greatly affect
resistivity.
As an example, consider a foot print with area 200cm2 (equivalent radius
= 0.08m), and a separation distance between the two feet of O.Sm. Table
2.13 lists the series and parallel contact resistance of two feet for soil
resistivity (} = 10, 100, and 1,000Qm-attributed to wet-loamy soil, moist
soil, and dry soil respectively (IEEE, 1986).
Footwear can add significantly to the total path resistance, as illustrated
in Fig. 2.27 (Reilly, 1979b). This figure plots the distribution of DC resis-
tance of individuals standing on various surfaces, with a current pathway
from a large electrode held in the hand, to a nearby driven ground rod in the
soil. The measurement voltage was SOOV. In all cases, footwear was dry,
except for surface moisture on which the person stood. In general, leather
soles are much more conductive than rubber soles. If leather soles become
wet, their resistance can fall greatly. Grass blades that touch the sides of the
shoes can also significantly lower their resistance. The curves labeled
"damp grass" apply to individuals standing on short grass; in the "wet grass"
condition, subjects first stood briefly in 1 cm of water before stepping on the
grass. In tall grass (e.g., 8cm), we would expect to see much larger percent-
ages of low resistances.
TABLE 2.13. Examples of contact resistance of

two feet.
Contact resistance (Q)
Wet soil Moist soil Dry soil
56.2 562 5,620
17.2 172 1,720
Q = 10, 100, I,OOOQm for wet, moist, and dry soil,

respectively.
100
80
--
>R..
0
ctS
(J)
(J)
'0 60
(J)
.0
ctS
-
V
(J)
cQ):
E 40
~
:::I
Dry grass (n =42)
(J)
ctS
Q)
:::?:
20
o
0.04 0.1 1 10 100 1000 2000
DC leakage resistance (Mil)
FIGURE 2.27. DC resistance of people through shoes, standing on various surfaces.

(From Reilly, 1979b.)
2.6 High-Voltage and Transient Properties

We studied current and voltage relationships under high-voltage conditions
with living subjects (Reilly et aI., 1983, 1984, 1985). These experiments
could be safely conducted on living subjects because the duration and
energy of the stimuli were limited by the use of capacitive discharges. These
experiments provided the first detailed account of high-voltage current
response of humans in the microsecond time scale.
The experimental apparatus is illustrated in Fig. 2.28. A high-voltage
transformer T was followed by a resistor Rc (variable from 1 to 128MQ)
and a capacitor Co) (variable from 100 to 6,400pF in steps of 100pF). An
internal transformer resistance of 1 MQ, and permanent series resistance of
2MQ, limited the current for human safety. The voltage at the energized
electrode could be varied from 0 to ± 15 k V in the DC mode, or 15 kV peak
in the AC mode.
The charge stored on Co could be discharged to a subject by one of
several methods: by actively touching the energized electrode; by bringing
High-Voltage and Transient Properties 53
to some location on the body a probe energized by Co; or by closing a switch

between Co and a passive electrode already in contact with the body. Stimu-
lus voltage and current could be simultaneously sampled at a maximum rate
of 50ns, and stored in a dual-channel digital oscilloscope which was con-
nected to a digital computer for manipulation and plotting of the data. The
return electrode was silverplated, with a contact area of 15 cm2 • It was
covered with conductive electrode paste and generally worn near the site of
stimulation, such as the forearm for stimulation on the fingertip. Because of
the size and treatment of the return electrode, its impedance was negligible
compared with that at the stimulus site.
What may be thought to be a simple stimulus-a single capacitive
discharge-actually has complexities that distinguish it from other
electrical stimuli discussed previously. These complexities include a separa-
tion of the discharge into spark and contact components, a plateau voltage,
marked impedance nonlinearities, differences in response to positive and
negative stimuli, and significant waveform differences at different body
locations.
Charge from a capacitor can be transferred to the skin by two modes:
the stimulus can be presented through a metallic contact on the body
Cu rrent and {
voltage waveforms Energized
H.V. probe electrode
AC/DC select
Current
transformer
FIGURE 2.28. High-voltage stimulator schematic. (From Reilly and Larkin, 1983.)
or through a spark contact. When a subject touches an energized

electrode, there can result both a spark (if the initial voltage is sufficient)
and a subsequent discharge when mechanical contact is made. To
understand the characteristics of the spark and contact components,
we studied discharge waveforms under a variety of stimulating
conditions.
Spark and Contact Components

When a voltage is applied across the air gap between two electrodes, the
free electrons and negative ions that normally exist in air will move toward
the anode (positive electrode). The positive ions move toward the cathode
(negative electrode). If the electric field produced by the voltage is suffi-
cient, the free electrons will be accelerated to velocities such that their
collisions with neutral atoms or molecules create more free electrons. This
process is known as electron avalanche.
When one of the electrodes is the human body, the nature of the electri-
cal breakdown will be greatly affected by the nonlinear impedance proper-
ties of the skin. The nature of the breakdown phenomenon is evident in
Fig. 2.29, where a subject lightly tapped with his finger an electrode having
a 1-mm-diameter tip that protruded by 0.5 mm above an insulated housing.
For each set of conditions depicted in the figure, four successive stimuli
were applied; the four corresponding voltage and current waveforms were
then averaged.
In this sequence of four recordings, the discharge capacitance is 400pF;
the initial voltage is varied. In Fig. 2.29a, the initial voltage of 405 V is too
small to produce a spark because of the dielectric protection of the skin. As
a result, current is conducted entirely by direct contact when the finger
touches the electrode. In Fig. 2.29b, the initial voltage of 515 V is just barely
at the point of a spark discharge at the initiation of the current trace
as revealed by the small initial current transient at about t = 75#s. In
Fig. 2.29c, the initial voltage has been raised to 586V (a 14% increase over
Fig. 2.29b), and the magnitude of the spark discharge grows by a factor of
10. In Fig. 2.29d, with the initial voltage at 989V, the spark discharge
component of the stimulus has increased by another factor of 10, and we
can see the convergence of voltage to a plateau around 450V, followed by
the contact at about t = 130#s.
In this sequence of waveforms, an initial voltage of about 500V delin-
eates a level above which spark discharges occur and below which they do
not. This is also approximately the plateau voltage level that is discussed in
the succeeding paragraphs. The experimental data suggest that the break-
down voltage for spark discharges to the skin is at about the level of 500V.
This is consistent with the dielectric strength of the excised human corneum
reported in Sect. 2.2.
40CHr--" 1.12
\ 500r~\
\
a) Vo=405V 400 \ b) Va" 515V 1.12
300 \ 0.84
.1\\
A \\ loot I \\
' \ " I '. ' 0.84 «
" 41
«E ->
\ -'..... 0.66 ,: .§.
200 i ~ .'
\"
• 'I "~-' ...
->1
--._,. :.::0 200+ \ .....,-
/ , ..
) ~ 0.66 E
~ --.. .--------------..1 > ..."
" .-'-.-~~---,-----,.----
..,
/
\',-- ~:::I
100 0.28 a "
\."'\~ 100 I 0.28 ()
)' }l ~~"".~
-\.......A""'......"',,"'_'<A.,,\~./,__~........\ .....
1--+-01 ""v¢ , " I ,~O r*l''''' I -t- " 1~...l!l!IIO
o 100 200 300 400 500 80 160 240 320 400 480
Time (J.lS) Time (ps)
1000 ::t:
6OOt....J 1.68 t
900
1.40 800 ~
' ) Vo" 686V _ 700 d) Vo=989V ~
'" c.. 1.12;;: ;:: 600 §
. 0.
- 400J I \ , 6.84
~
-ie ~II 400 :;3
> I \ • > 6.6 E ~
o , 6.56d iii'
" f t:l
-j~ V~ ~~ 4.2 5
()
...
> - "- 0.28 2.8
(1)
1001 I .... , •• -.,0

~~ 1.4
1
::4-
I I
./-............- iii'
CIl
I---i-U'" I I I
200 300 400 500 200 300 400 stJf
100 Time (ps) Time (ps) VI
VI
FIGURE 2.29, Stimulus waveforms for light tap force, capacitive discharges with C = 400pF. (From Reilly et aI., 1982.)
Plateau Voltage
In the example illustrated in Fig. 2.29, separate voltage plateaus can be seen
for the spark and the contact phases of the discharge. The contact plateau
is approximately 100V, and the spark plateau is approximately 450V.
These plateaus suggest that as the voltage declines, the impedance becomes
very large in both phases of the discharge. The presence of a voltage plateau
reveals that a charged capacitor cannot be totally discharged through a
spark to the skin. It also delineates a voltage level above which spark
discharges can occur.
The plateau voltage for a spark stimulus can be studied by bringing an
energized electrode slowly to the body to produce a spark discharge that is
not followed by a contact. Figure 2.30 illustrates four successive discharge
waveforms for a pencil-shaped probe (0.8-mm-diameter tip) that is brought
slowly to the same point on the finger pad. Discharge to the same point was
ensured by the use of a dielectric mask with a O.5-mm hole. The initial
voltage has been varied from 750 to 1,900V. In each case, a plateau voltage
in the range of 450 to 500 V is evident. The presence of this plateau indicates
that the spark component does not completely discharge the capacitor and
that the discharge impedance converges to a very large value in the region
where the discharge current approaches zero. When tested over various
subjects and body locations, the plateau voltage was found to range from
about 450 to 650V, with an average of 525V. Stripping away the corneal
700
CO" 800 pF
+ polarity 600
500
~
....'" «
'0 400 E
...o
> ...c:
;!:: OJ
(.) 300 "-
"-
'"
0-
to
::l
U
U
200
100
0
5 10 15 20 25
Time (I-ls)
FIGURE 2.30. Four successive discharges to the same point (second finger, point A),
with variable initial voltage. (From Reilly and Larkin, 1983.)
layer of skin on the forearm lowers the plateau voltage to about 330 V as
described below.
Spark Discharges and Corneal Degradation

The contribution of the corneum can be studied by stripping the skin with
cellulose tape. With each stripping operation, a portion of the corneum
adheres to the tape. This procedures was applied to a portion of the volar
forearm to study the effects of spark discharges (Reilly et aI., 1982). A
charged capacitor (400pF@ 1,000V) was discharged through a small probe
to a precise point on the skin defined by a small hole in a dielectric mask.
Altogether 11 stimuli were applied to each of four target points on both
stripped and control skin, and a total of 60 strippings were performed.
At the end, the skin appeared glossy, and was sensitive to the stripping
procedure.
Table 2.14 shows experimental data; each entry is the average of mea-
surements at four distinct points on both stripped and control skin. Listed
are the plateau voltage Vp , and the initial impedance R i , defined by the ratio
of the initial voltage and peak current. Vp in the stripped area dropped from
480V to a low of 330V. At the same time, Ri dropped by a factor of over
5: 1. At the control skin, we see a slight increase in Vp and a 16% decrease
in R i. The slight drop in Ri in the control area probably results from burning
away a small portion of the corneum; the slight increase in Vp remains a
mystery. The data also indicate that subcorneallayers also exhibit a signifi-
cant plateau voltage. Vp mirrors the minimum voltage from a static or low-
frequency source needed to support a spark discharge to the skin.
Tests were conducted with repeated spark discharges on the dry skin
of the leg (Reilly et aI., 1982) from a 200pF capacitor charged to 2,000V
(stored energy = O.4mJ). With 100 repeated discharges to the same
point on the skin, a small but noticeable change in skin impedance was
produced, along with a small reddened spot about O.5mm in diameter.
Similar discharges applied to the fingertip did not result in observable skin
changes due to the greater thickness and conductivity of the palmar
corneum.
The energy of a spark discharge may be dissipated in a very small volume
in the skin. The effective area of a spark contact is less than 1 mm2 as
evidenced by the fact that a spark to the finger varies substantially over
distances of about 1 mm-a distance that is similar to the density of sweat
ducts in the palm of the hand. The previously presented data on initial
impedance suggest that the effective diameter of contact is roughly a few
tenths of a millimeter, and possibly as small as 0.1 mm.
We can compare these discharges with ordinary carpet sparks. Taking the
capacitance of the human body as 100pF and the voltage in the range of 2
to 3kV (a common range for carpet sparks), the energy in a typical carpet
spark would be in the range of 0.2 to 0.45mJ. We see that the minimum
TABLE 2.14. Plateau voltage and minimum resistance

for epidermal stripping.
Epidermal Stripped skin Control skin
Stimulus stripping Vp R, Vp R,
number number (V) (kQ) (V) (kQ)
1 0 * * * *
2 1 480 22.5 470 30.5
3 2 435 15.5 450 35.9
4 3 450 24.1 440 32.5
5 5 410 10.9 500 36.4
6 8 415 8.8 475 26.3
7 12 * * 510 23.9
8 20 385 8.1 500 24.4
9 30 400 4.9 525 25.0
10 46 390 4.8 * *
11 60 370 4.3 * *
1** 60 330 3.7 * *
* Data unavailable.
** Last data set appliesto simulus in stripped area, but at
four points not previously stimulated. Each measurement
represents an average over four spatial points. (From Reilly
et aI., 1982).
energy per spark cited above for skin erosion is similar to the energy
commonly encountered in carpet sparks.
In an alternating field, each half cycle of oscillation presents a new
opportunity for a spark. The maximum rate of sparking will increase as we
raise the frequency of the energizing field. At power frequencies, the spark
discharge phenomenon is very much like that for static discharges (see
Chapter 9). For frequencies substantially above 60 Hz, the breakdown volt-
age of air becomes lower as a result of the transit time of positive ions and
electrons with respect to the gap length and oscillation frequency (Craggs,
1978). At frequencies of several MHz, and small metallic electrode gaps
«1 mm), the reduction in breakdown voltage at atmospheric pressure is
approximately 15 to 20% compared with that at 60Hz. At much higher
frequencies in the GHz regime, breakdown voltages can be more signifi-
cantly reduced. Section 11.4 discusses standards for human exposure to
spark discharges induced by radio-frequency electromagnetic fields.
Discharge Impedance
The dynamic properties of discharge impedance are pertinent to transient
high-voltage shock exposure. With capacitive discharges, for example, the
impedance affects the time constant of the discharge. The time constant, in
turn, significantly affects the neural excitation potency of the discharge, as
discussed in Chapters 3 and 4. In other applications, knowledge about body
impedance is needed to define the limiting value of current in a high-voltage

exposure-such as with fault conditions in high-voltage installations.
Figure 2.31 expresses impedance data for Fig. 2.30 as the ratio voltage/
current. These dynamic impedance functions reveal that the skin imped-
ance to a spark discharge exhibits a nonlinear relationship with voltage in
which peak current is not directly proportional to the initial voltage. The
variation in initial impedance with initial voltage is evident in Fig. 2.3l.
The impedance attains a minimum value in a fraction of a microsecond
after the stimulus onset and increases during the time course of the stimu-
lus, during which time the stimulus voltage is decreasing. The large excur-
sions beyond 16.us are the result of digital quantization effects.
The impedance to a capacitor discharge consists of two contributions: the
voltage drop across the arc itself and the voltage drop across the subject.
We measured the arc voltage/current relationship with the human body in
direct contact with one of two metallic electrodes forming the arc gap, and
in series with the electrical path. The voltage drop across the arc for this
case varied from nearly zero at the point of arc initiation to approximately
50 to 75 V at the point of arc extinction. As a result, when two metallic
electrodes form the discharge path, the arc does not contribute significantly
to the overall measured impedance. We cannot be certain that the same arc
voltage drop will apply when the spark terminates at the human skin rather
106
S
Q)
u
c
ctI 105
-0
Q)
a.
E
104
10 20
Time (tis)
FIGURE 2.31. Impedance versus time for four successive discharges to the same
point (second finger, point A) with variable initial voltage. Voltage and current
waveforms as shown in Fig. 2.30. (From Reilly et aI., 1982.)
than a metallic electrode. It does not seem possible to make a more direct
measurement of the arc impedance when the discharge terminates at the
human skin without significantly altering the skin's electrical response. If,
however, the arc voltage drop is similar to that when two metallic electrodes
are used, then the arc voltage drop must be small compared with that across
a human subject.
Variation with Stimulus Location

Waveforms for discharges to different discrete body locations can differ
significantly even though the stimulated points may be separated by only
2mm. Figure 2.32 illustrates voltage and current waveforms for anodic
(positive polarity) spark discharges to four points on the fingertip separated
by 2mm. To localize the spark stimulation, a dielectric mask with a 0.5-
mm-diameter hole was placed on the finger. But repeated positive
polarity stimuli at the same point produce nearly identical waveforms. This
closely spaced pattern of skin impedance may reflect the mosaic of sweat
ducts.
The fact that identical nonlinear current response is seen with repeated
discharges to the same point on the finger suggests that, with the tested
discharge parameters, the impedance breakdown is not permanent.
Perhaps ionic disassociation rather than membrane rupture is involved.
1200
Co = 800 pF
1000 + polarity
Point A
s;-
a>
01 800
.-
to
0
>
"-
0
600
.~
U
to
Q.
to 400
U
200 50
a
0 5 10 15 20
Time (j.Ls)
FIGURE 2.32. Waveforms for stimuli to four different points on the same fingertip
(second finger, points A-D); each point separated by 2mm. (From Reilly et aI.,
1982.)
Repeated discharges at similar energy levels on the dry skin of the calf on
the leg, however, eventually resulted in visible erosion of the corneum.
Polarity Effects
Positive-polarity current transients appear relatively smooth and repeat-
able, but negative-polarity waveforms often are bistable and have lower
initial impedance than the positive waveform, as illustrated in Fig. 2.33. In
this example, four cathodic spark discharges are applied to point A (also
used for the anodic discharges shown in Figs. 2.31-3.33).
The ionized discharge process may be responsible for these polarity
differences. Electrons move so rapidly that by the time the avalanche has
reached the anode, the newly created positive ions are virtually still in their
original positions, forming a positive space charge conically concentrated at
the anode, and tapering or decreasing toward the cathode as shown in Fig.
2.34a (Howatson, 1965). This distribution of positive ions modifies the field
near the anode as shown in Fig. 2.34b. The conical spreading of the plasma
discharge from the cathode to the anode results in a larger area of a spark
contact at the anode than at the cathode. The lower initial minimum imped-
ance for negative polarity may result from the greater ion contact area with
the body as an anode than as a cathode. The bistable character of negative
polarity discharges may result from an unstable space charge distribution
and its effective area of anodic contact.
1200
300
Co = 800 pF
1000 - polarity
250
800
200 «
E
...
600
150 e
...
:::!
U
400
100
200
50
o~~~~~~~~~~o
a 5 10 15 20 25
Time (ps)
FIGURE 2.33. Waveforms for successive discharges to the same point (second finger,
point A); negative polarity. (From Reilly et ai., 1982.)
Anode +
Field enhanced by
space charge
Enhanced field produces~

auxil iary avalanches
Positive space charge

left by main avalanche
electrons
( Cathode -
a) Formation of a steamer
Electric field
Anode
T~~
Q,)
u
...
c::
ro
.!!!
Field enhanced by space charge
1
Cl
f---Uniform field before discharge
Cathode
b) Electric field
FIGURE 2.34. Spark discharge formation: (a) formation of a steamer; (b) electric
field. (Reprinted with permission from A. M. Howatson, An Introduction to Gas
Discharges, © 1965, Pergamon Press PLC.)
Stimulus Time Constants

A capacitance C discharged to an ideal linear resistance R will have an
exponentially decaying current and voltage with a time constant r = RC,
which defines the time at which the voltage or current decays to e- 1 of its
maximum value. As evident in the previous examples, many of the
discharge patterns can be approximated by exponential functions with
current decaying to zero and voltage decaying to the plateau value.
Although an ideal RC circuit results in an exponential discharge pattern,
the converse statement should not be assumed: an exponential current
decay does not necessarily imply constant resistance. This can be appreci-
ated with reference to the discharge patterns of Fig. 2.30: both current
and voltage waveforms are approximately exponential functions, but,
because of the nonzero voltage plateau, the impedance varies from a
minimum value at the onset of the discharge to a very large value at its
extinction.
The instantaneous impedance to a capacitor discharge departs signifi-
cantly from what would be expected in a linear model. Linear circuit models
for skin impedance, commonly invoked in studies of low-voltage cutaneous
currents, do not adequately represent the measured impedance response.
However, the measured transient response might be accounted for if the
model parameters were voltage- or time-dependent.
We define time constants for capacitor discharge stimuli as the time
required for the voltage to decay to e- 1 dV, where dV is the difference
between the initial and plateau voltages. The decay time is measured
from the onset of the current waveform. The measured plateau for the
touch component was generally in the vicinity of 100V (as in Fig. 2.29),
in contrast to the plateau of about 500 V for the spark component.
The value of d V defines the amount of charge passed, in accordance with
Q = CdV.
Figure 2.35 depicts time constants for spark and contact components as a
function of voltage for a single subject who discharged a capacitor by
actively touching an energized electrode. The discontinuity around 500 to
600 V occurs because spark discharges could be produced only above this
range. The time constants decrease as the initial voltage on the capacitor is
increased and as the capacitance is decreased. In the test procedure a
subject discharged a capacitor by actively touching an electrode. The result-
ing effective discharge contact area is very small for both spark and contact
phases. We hypothesize that at the instant of a tapped electrode contact, an
initially small point of conduction establishes a preferred current channel.
As a result, both spark and contact phases may utilize a small contact area,
regardless of the electrode size. In contrast, discharges to an electrode held
in contact with the skin appear to utilize multiple current channels. The
result is a much lower impedance at the electrode/skin interface, and signifi-
cantly smaller discharge time constants. For example, discharges from a
+ polarity
1:.,
c:
8.
E
8
t;
!l
c:
100 8
...c:
ra
t:
c:
.,8
'E..
.2
E 10
~
c.,
c:
8.
E
ou
-t:
II
Vl
0.5~~~ ____~__~~~~~~~~~____~____L -__~~__- J

0.15 0.2 3.0 4.0
Initial stimulus voltage (kV)
FIGURE 2.35. Time constants for monophasic capacitor discharge stimulation of the
fingertip. Parameter shown is the capacitance in picofarads. (From W. D. Larkin
and J. P. Reilly, Perception and Psychophysics 36 (1): 68-78, 1984, reprinted by
permission of Psychonomic Society, Inc.)
6,400-pF capacitor to a 1.3-cm-diameter contact electrode result in dis-

charge time constants under 3fls, even when the initial voltage is as low as
SOY.
Minimum Impedance with Spark Discharge4

The minimum impedance to a spark discharge occurs at the onset of the
stimulus, when the applied voltage is greatest. Figure 2.36 shows the mini-
mum impedance when the spark discharge and return electrode are on the
calf of the leg. The curves are second-order least-squares fits to the experi-
mental data. Two striking features are the voltage dependence and the
4Adapted from the work of R.J. Taylor, as reported in Reilly et al. (1982,1983) and
Taylor (1985).
o Positive polarity
o Negative polarity
102L--L~~-----L----~--L-~
0.4 1 4 10
Initial applied voltage (kv)
FIGURE 2.36. Minimum impedance of leg (volar calf) versus voltage, for spark
discharges to skin; leg grounds. (From Reilly et aI., 1982)
polarity sensitivity. The voltage dependence reflects the nonlinear skin

impedance as discussed in Sect. 2.2. Additionally, the diameter of spark
contact area may increase with voltage-at higher voltages, the distance
between the skin and the discharge electrode is greater, allowing the base
area of the space charge cone (Fig. 2.34) to increase. When the spark
discharge terminates at an electrode already in contact with the skin,
the impedance is much lower than when the spark terminates directly on
the skin.
Internal Body Impedances

The internal body impedance can be estimated from transient measure-
ments of capacitor discharges to large contact electrodes, where the large
contact area and high voltage minimize skin impedance. Figure 2.37 shows
an example of the ratio of voltage and current during a capacitor discharge
to a 3-cm-diameter cylinder electrode grasped by a subject with the left
hand while standing with the bare left foot on a copper plate; the discharge
capacitance was 800pF, and initial voltage was SOOV. The discharge process
5 Adapted from the work of R.I. Taylor, as reported in Reilly et aI. (1982, 1983) and
Taylor (1985).
800r-~~-.-------.-------'------~------~------~
...
CD
lij 600
X
E
Co = 800 pF capacitor
Vo = 500 V initial voltage
400~----~~----~~----~~----~------~------~
0.0 0.2 0.4 0.6
Elapsed time (~s)
FIGURE 2.37. An example of body impedance measurements versus time; 3-

cm-diameter electrode in left hand, left bare foot on plate. Spark discharge was
initiated between 0.0 and 0.05,us. (From Taylor, 1985.)
was essentially complete during the interval illustrated in the figure. Data
point 1 is not considered meaningful because of the limited rise time of the
instrumentation (O.l.us). At data point 2, the impedance is 434Q (434 V and
1A current). Data point 2 is strongly influenced by the distributed capaci-
tance of the body to ground as shown in Figure 2.38; an initial current surge
is required to supply the body's surface charge in addition to the current
that passes internally. The impedance measurement at data point 2 changed
as much as 50% as the proximity between the body and grounded objects
was varied. The measurements at data point 3 changed only a little (about
10%) with variations in distributed capacitance. The measurements at data
point 3, termed "initial body impedance," were considered by Taylor to
represent the core body impedance.
Table 2.15 shows initial body impedance from various electrode
configurations (from Taylor, 1985). The treated skin had conductive elec-
trode paste applied to it. The impedance for dry and treated skin differs but
a little, thus supporting the hypothesis that skin impedance is negligible
with this measurement technique. The data in Table 2.15 agree well with
internal body impedance data determined by other methods (refer to
Sect. 2.3).
Voltage waveform
Current transformer
High
vottage
probe
/ I
-
...... ......
"- '\
/ I \
/ I \
/ I \
I \
:::;::; I I
Transformer I I I
coupling from
110 Vac I
I 1
'T'
I
I
~II I
I
I
!
I
I
-'-
I
I
I
'T'
. Coupling
I
capacitance I
/ I I
I I I
I I I
I I I
I I I
I I I
.-L --L -L
FIGURE 2.38. Body impedance measurement with capacitive coupling of body to

ground. (From Taylor, 1985.)
TABLE 2.15. Initial total body impedance

(Q).
Standard
Path Mean deviation
Dry skin
Left hand/left foot 533 52
Right handlleft foot 521 37
Right hand/left hand 508 29
Treated skin
Left handlleft foot 516 55
Right hand/left foot 507 36
Right hand/left hand 490 36
Source: Data based on measurements of seven

individuals (from Taylor, 1985).
Discharge
• Discharge measurements
o Freiberger data
o 100
Percentage of total internal body impedance
FIGURE 2.39. Internal body impedance between sole of foot and various locations of
the active electrode along the body. Open points from Fig. 2.20. (From Taylor,
1985).
Figure 2.39 and 2.40 illustrate the internal body impedance for different
locations of the active electrode, measured as a percentage of the total
hand-to-hand and hand-to-foot internal impedance. The filled circles repre-
sent the transient discharge measurements, averaged over seven adult
individuals. The open circles are derived from the model of Freiberger
(Fig. 2.20). The agreement between the two experiments is apparent.
2.7 Impedance of Domestic Animals

Farmers are sometimes concerned about electrical exposure of farm ani-
mals. The term stray voltage has been coined to indicate unwanted electrical
potentials to which farm animals may be exposed. Stray voltage is usually
related to the grounding of equipment on the farmstead, on- and off-site
transformers, and associated ground currents. Potential differences applied
Impedance of Domestic Animals 69
Discharge
• Discharge measurements
o Freiberger data
a 100
Percentage of total internal body impedance
FIGURE 2.40. Internal body impedance from one hand to various locations of the
active electrode along the arms. Open points from Fig. 2.20. (Adapted from Taylor,
1985.)
to a farm animal can result in unpleasant shock, which can lead to animal
handling problems and aversion to feed or water (Lefcourt, 1991). Con-
cern about stray voltage is particularly strong in the dairy industry, where
many researchers, consultants, and legal practitioners have focused on this
problem.
Potential differences of only a few volts can be disturbing to cows (see
Sect. 7.12) while being unnoticed by the farmer. The animal's sensitivity to
such low voltages does not mean that the cow is acutely sensitive to electri-
cal stimulation, but rather that her electrical impedance is typically much
lower than that of the farmer. This low impedance is a consequence of the
cow's size and weight, and the fact that she contacts surfaces that are wet
and contaminated with urine and manure. And like the human skin, cows'
hooves are good conductors when impregnated with such effluents.
The animal may access potential differences across various points on its
body-for instance, from a metal water or feed bowl to the feet, from foot
to foot, or from a shoulder contacting a stanchion to the feet. One can
determine impedance for various points of contact using the impedance
FIGURE 2.41. Cow impedance model for stray voltage application.
model of Fig. 2.41 (Reilly, 1994). The figure shows impedance components,
front and hind foot contact resistance (RCF and RcH ) and muzzle contact
resistance (RcM). Table 2.16 lists median values of cow impedance compo-
nents from the data of Norell et al. (1983). Foot contact impedance is listed
in Table 2.17 using Eqs. (2.19 and 2.20). The contact area of each hoof is
120cm2 (equiv. radius = 6.2cm); distance between right and left feet is
O.4m; distance between front and rear feet is 1.0m.
To illustrate cow impedance from the model of Fig. 2.41, consider moist
soil conditions (e = 100 Qm). With a muzzle to all feet pathway, the model
TABLE 2.16. Cow impedance model values.

Resistance
Symbol Impedance Component (Q)
RH Head & neck 183
RB Body 30
RFL Front leg 882
RHL Hind leg 525
R CH, RCF Hind, front foot contact see text
Rs Source impedance variable
RCM Muzzle contact variable
Cow impedance componets derived form median data of

Norell et al. (1983).
Impedance of Domestic Animals 71
TABLE 2.17. Cow foot contact impedance.

Soil R contact
Soil resistivity One foot Two feet parallel
condition (gm) (g) (g)
Wet 10 40 22
Moist 100 400 220
Dry 1,000 4,000 2,200
Assumptions: Contact area each foot = 120cm2 ; distance

between right and left feet = O.4m; distance between from
and hind feet = 1.0m.
indicates an impedance of 472 Q, excluding the muzzle and source imped-

ance (ReM and Rs). If the cow lifts one front foot, the impedance increases
to 549Q. The impedance between the two front and two rear feet would be
1,173 Q; with one front foot raised, the impedance would be 1,794Q. If the
animal stands on a perfectly conductive surface, the impedance from
muzzle to four feet would be 358 Q. These data are attributed to an adult
cow, for which the average weight at 2 years is approximately 500kg
(ASAE, 1993).
Much smaller cow impedances were reported by Lelcourt and Akers
(1982). However, in Lefcourt's procedures, the leg electrodes were placed
on shaved areas above the hock or knee, and these areas were treated with
electrode paste. By placing the electrode above the knee rather than on the
hoof, one eliminates more than 50% of the leg impedance. This occurs
because a cross-section of the leg above the knee contains a large bulk of
high-conductivity muscle, but the cross-section below the knee is much
smaller, and consists largely of low-conductivity bone and ligament, as
suggested by Fig. 2.42.
Impedance of growing or finishing pigs of average body weight (89.2kg)
were measured when standing on a metal grid (Gustafson et aI., 1986). The
mean impedance from mouth to four feet was 789 Q, with a standard
deviation of 262 Q. For constant body proportions, impedance varies in-
versely with body weight to the one third power (see Sect. 2.4). When
compared with the corresponding impedance of cows (358Q), one finds
that the cow and pig resistances compare roughly as the inverse cube-root
of their respective body weights, even though their body configurations are
different.
1st Phalanx
2nd Phalanx
3rd Phalanx
FIGURE 2.42. Distribution of current with cow foot contact.

3
Electrical Principles of Nerve
and Muscle Functionl
3.1 Introduction
This chapter begins with some basic electrical properties of biological cells
and then examines the function of a special class of cells that have electri-
cally excitable membranes. It shows how these properties are organized for
sensory and muscle function. This information will help explain how exter-
nally applied currents can modify or interfere with normal function.
Consider a familiar example that points out several pertinent electro-
physiological functions. You touch a hot object with your finger-after a
brief delay, you feel pain, and jerk your hand away. This sequence of events
involves a number of electrical functions of nerve and muscle. First, the
skin's temperature rise is converted to an electrical potential by a special-
ized transducer or receptor in the skin-an example of a pressure-sensitive
receptor is shown in Fig. 3.1. The receptor's voltage response, called a
generator potential, initiates a nerve impulse, called an action potential
(AP). The AP travels along an electrical cable known as a nerve axon,
which runs from the receptor to the spinal column, where connections,
called synapses, are made with additional nerve cells; some of these ulti-
mately carry information to the brain. The receptor, axon, and synaptic
terminus comprise a single nerve cell (also called a neuron). Bundles of
neurons are commonly called nerves. In our example, the length of
the nerve cell, from the fingertip's receptor to the spinal synapse, will be
approximately 1 m.
Thus far, we have discussed an afferent nerve cell, that is, one that carries
information from the body's sensory system to the spinal cord and then to
the brain. Our example sensory nerve cell is also a slowly conducting type
in which the conduction velocity may be only a few meters per second.
Because of the delay in the conduction and synaptic processes, a substantial
fraction of a second may pass before the brain is notified of the finger's
1For background material, see Kandel et al. (1991); Ruch et al. (1968); Stein (1980);
and Plonsey (1969).

74 3. Electrical Principles of Nerve and Muscle Function
Synapse
Pressure
stimulus
Receptor
Nucleus
Axon
. . . ..-_--Myelin--_
. _ _-.._"'-
)~ Node of - - - - . u l
Ranvier
III •
Muscle contraction
(a) (b)
FIGURE 3.1. Functional components of (a) motor and (b) sensory neurons. Arrows
indicate the direction of information flow. Signals are propagated across synapses
through chemical neurotransmitters and elsewhere by membrane depolarization.
Synapses are inside the spinal column. The sizes of the components are drawn on a
distorted scale to emphasize various features.
temperature rise and pain is consciously registered. Even before that,

however, a reflex action may be set in motion within the spinal column. In
the reflex action, sensory inputs, and possibly other inputs from the brain,
are processed within the spinal column, and, if the inputs satisfy threshold
criteria, AP signals are sent to the muscles of the hand and arm-traveling
along efferent (motor) neurons, that is, ones that carry signals from the
central nervous system (eNS) to the muscles (Fig. 3.1a). When the efferent
APs reach the neuron terminus at the muscles, they initiate a series of
Cellular Membranes 75
Myelin
Membrane
FIGURE 3.2. Detail of myelinated fiber at a node.
events leading to electrical excitation and contraction of muscle fibers. The

result is that the hand pulls away in a reflex action.
The mechanisms responsible for normal function also respond to exter-
nally applied currents. With appropriate control, electrical stimulation can
provide medical assistance and diagnostic benefits, such as muscle control,
pain relief, sensory prosthesis, and diagnosis of nerve and muscle pathol-
ogy. However, in chance encounters with electric currents (electric shock),
the result can be pain, hazardous muscle reactions, heart disturbances, and
ultimately death.
Figure 3.1 illustrates several functional components of a sensory and a
motor (muscle) neuron. In this example, the neuron is myelinated, that is,
covered with a fatty layer of insulation called myelin, and has exposed nodes
of Ranvier. Other neurons are unmyelinated. The conducting portion of the
neuron is a long, hollow structure known as the axon (illustrated in Fig. 3.2).
The axon plus myelin wrapping is frequently referred to as a nerve fiber.
The arrows in Fig. 3.1 indicate the direction of information flow. For the
motor neuron in Fig. 3.1a, APs propagate from synapses in the spinal
column to the terminus at the muscle. For the sensory neuron in Fig. 3.1b,
APs originate from one of a variety of specialized receptors (a pacinian
corpuscle is illustrated) and proceed to a synapse in the spinal column.
Communication across the synapses is accomplished through chemical
substances known as neurotransmitters.
3.2 Cellular Membranes

Cells are the basic building blocks of both plant and animal life. The
functional boundary ofthe cell is a thin (about lO-nm) bimolecular lipid and
protein structure. One role of the membrane is to regulate chemical ex-
change from within the cell (the plasm) to its surroundings (the interstitial
fluid). The electrochemical forces at the membrane are intimately involved
in the regulation.
Inside of cell Membrane Outside of cell
~ Na+
[Na+); = 12 [Na+]o = 145
[K+]i = 155 [K+]o = 4
[CI-]i = 4 [CI-]o = 120
V = 90 mV V=OV
- Na+
FIGURE 3.3. Schematic of a typical cell membrane. The channels allow the passage
of ions. Numerical values indicate approximate steady-state concentrations
(umoUcm 3 ) for typical mammalian muscle cells. An active metabolic pump drives
Na+ out of the cell and K+ into the cell. The interior potential is about -90mV
relative to the outside.
The plasm and interstitial fluids are composed largely of water containing
ions of different species. The concentration of ions inside and outside the
cell differs leading to the electrochemical forces across the cell membrane.
The membrane is said to be semipermeable; that is, it is basically a dielectric
insulator that allows some ionic interchange. Figure 3.3 represents a
membrane as a barrier, with channels that permit the passage of ions; the
individual channels may be very selective with respect to the ionic species
that are allowed to pass. Typical concentrations inside and outside a cell are
shown in Table 3.1 for Na +, K+, and Cl- ions. The concentrations indicated
in Table 3.1 are markedly different inside and outside the cell. The differ-
ences lead to two forces that tend to drive ions across the membrane: a
concentration gradient and a voltage gradient. In order to understand these
forces, first consider an environment where only one ionic species, sub-
stance S, is present.
The Nernst Equation

In a solution where the concentration varies from one region to another,
there will be a net flux from the region of higher concentration to the lower.
TABLE 3.1. Example cellular ionic concentrations.

Concentration Nernst
CuMlcm3) potential
Species Inside Outside (mV)
A. Mammalian muscle cells
Na+ 12 145 66
K+ 155 4 -97
cr 4 120 -90
Resting potential -90
B. Squid axon
Na+ 50 460 59
K+ 400 10 -98
cr 40--100 540 -45 to -69
Resting potential -60
Source: Data from Ruch et al. (1968) and Katz (1966).
The concentration potential energy difference, W Co is work required to

move a mole of S against the gradient. This quantity is proportional to the
logarithm of the concentration difference in accordance with
We = RT(ln[Sl- In[SL) (3.1a)
=RTln[sl (3.1b)
[SL
where [Sl and [S]o represent the concentrations of S inside and outside the
cell, R is the gas constant, and T is absolute temperature. The product RT
has units of energy per mole.
If S is ionized, an electrical potential difference will occur between
the two regions of differing concentration. The electrical potential energy,
We> is
(3.2)
where Z is the valence of S, F is the Faraday constant (the number of
coulumbs per mole of charge), and Vm is the potential difference across the
membrane.
The total electrochemical potential difference is the sum of the concen-
tration and electrical potentials:
I:!W = We + We (3.3)
Substituting the quantities from Eqs. (3.1b) and (3.2) results in
I:!W = RTln [Sl + ZFV (3.4)

[SL m
When ~ W = 0, S is at equilibrium across the membrane, that is, there is no

net force in either direction, and the net flux across the membrane is zero.
Under conditions of equilibrium, the membrane will attain the potential:
v = RT In[sL (3.5)
m FZ [Sl
Equation (3.5) is known as the Nernst equation. It is a statement of the
membrane potential for an ionic substance in electrochemical equilibrium.
Using the values R = 8.3IJ/moIK, T = 310K (37°C), F = 96,500Clmol, and
Z = + 1 (for a monovalent cation), converting to the base 10 logarithm, and
expressing Vm in millivolts, we obtain:
v = 61l0g [SL (mV) (3.6)

m [Sl
For a system with more than one permeable ionic species, the equilibrium
voltage will depend on the concentration and relative permeability of the
individual ions. For a system consisting of K+ and Na+, for example, the
expression is
(3.7)
where P K and P Na are the permeabilities (expressed in units of centimeters

per second) for K+ and Na +, respectively. An alternate expression for Eq.
(3.7) uses the ratio q = PN/PK to obtain
(3.8)
One can get a feel for the changes in Vm during excitation by considering
the simplified circuit diagram in Fig. 3.4. The membrane permeability
is represented by conductances gNa and gK, and the electrochemical
gradients as potential sources ENa and E K • For an excitable membrane
in the resting state, gNa < < gK and the membrane potential approaches
the Nernst potential for K+, as indicated by Eq. (3.8). In the excited
state, gNa > gK' and the switches in Fig. 3.4 would be connected in the
alternate positions, forcing the membrane to move toward the Nernst
potential for Na+.
Consider the individual Nernst potentials for the ionic species listed in
the right-hand column of Table 3.1. The equilibrium potential for Na+
(66mV) is far removed from the membrane potential (-90mV), K+ is
slightly out of equilibrium, and Cl- is essentially in equilibrium. The magni-
Outside of cell
Reference voltage (0 V)
Membrane
ENa Ek
(66 mV) (98 mV)
Resting potential: -90 mV

Inside of cell Excited potential: + 20 mV
FIGURE 3.4. Circuit diagram representing membrane conductance for Na+ and K+
ions. In the resting condition, the inside of the cell is at a potential of -90mV. In the
excited state, the inside of the cell is at a potential of 20mV.
tudes and signs of the potentials show that a strong electrochemical force
tends to drive Na+ into the cell and a relatively weaker force tends to drive
K+ out of the cell. Given that the membrane is at least somewhat permeable
to the ions discussed here, these forces ought eventually to bring the species
into equilibrium. Clearly, another force is working to maintain dis-
equilibrium. The responsible force, is the so-called sodium pump, an active
system that pumps Na+ out of the cell and K+ into the cell. The energy
for the pump is derived from the cell's metabolism. A dead cell would
eventually reach equilibrium potential.
The electrical forces on the membrane are quite large. Considering the
membrane potential (=10- 1 V), and thickness (=10- S m), the electric field
developed across the membrane is about 107 Y/m. Conductivity properties
of the excitable membrane are intimately tied to the membrane electric
field; disturbances from the resting condition can lead to profound changes
in the membrane's electrical properties. These changes ultimately initiate
and sustain the functional responses of nerve and muscle.
3.3 The Excitable Nerve Membrane

Nerve and muscle cells possess membranes that are excitable, such that an
adequate disturbance of the cell's resting potential can trigger a sudden
change in the membrane conductance. The resulting membrane voltage
change will affect adjacent portions of the membrane, and in a nerve, will
propagate as a nerve impulse. The response of the excited membrane is an
action potential (AP).
To illustrate the properties of excitability and propagation, consider the
experiment illustrated in Fig. 3.5, in which a small stimulating electrode
(SE) is near a nerve fiber, and two small recording microelectrodes (RE)
pierce the membrane; the return electrodes are assumed to be immersed in
the conducting medium some distance away. The stimulating electrode is
connected to a current source. In Fig. 3.5, the small arrows represent the
distribution of conventional current flow. The voltage disturbance due to
the stimulating electrode reduces the membrane potential (depolarization)
near the cathode, and increases the potential (hyperpolarization) elsewhere
along the axon.
Figure 3.6 illustrates the response of the membrane to the rectangular
current pulses shown in the upper part of the figure. Six possible current
FIGURE 3.5. Nerve excitation and measurement arrangement. Excitation is initiated

near the cathode of the stimulating electrode. Arrows indicate direction of conven-
tional current flow.
The Excitable Nerve Membrane 81
120
>E
"iii
''::;
...c:
Ol
0
c-
Ol
c:
...co
.0
E 0
Ol
~
-40
0 0.1 0.2 0.3 0.4 0.5
Membrane response time (ms)
FIGURE 3.6. Excitation of a nerve fiber by an applied current, as in Fig. 3.5.
magnitudes labeled a to f are shown. Pulses a to c apply if SE is an anode;

pulses d to f apply if SE is a cathode. The membrane response is shown as
measured by REI and RE2, where OV represents the resting potential.
Responses a to c are in the direction of hyperpolarization, and responses d
to f are in the direction of depolarization. Responses a to d exhibit the
characteristic of a linear RC network (see Chapter 4). Response e is slightly
below the excitation threshold, and response f illustrates a fully developed
action potential. The signal at RE2 is a delayed version of the AP, demon-
strating conduction along the axon. The membrane response is often re-
ferred to as "all-or-nothing," because the peak AP response is not normally
graded-the membrane is either excited or it is not.
To develop a quantitative model of excitation of the neuron by externally
applied currents, nonlinear membrane models and electrical cable theory
must be used. The membrane itself acts as the dielectric, and the ions on
either side of the membrane act as the conductive plates of the capacitor.
The nature of the leakage channels in the dielectric distinguishes the excit-
able membrane from the ordinary cellular membrane.
The Hodgkin-Huxley Membrane

With a series of ingenious experiments that eventually lead to a Nobel Prize
in Physiology, Hodgkin and Huxley (1952) provided the first detailed de-
scription of the electrical properties of the excitable membrane of unmyeli-
nated nerve cells. This work was later extended by Frankenhaeuser and
Huxley (1964) to describe the myelinated nerve membrane. For brevity, we
shall refer to the Hodgkin-Huxley and Frankenhaeuser-Huxley work as
HH and PH equations, respectively. The system of HH and FH equations

is largely empirical.
Figure 3.7 illustrates the HH membrane schematically. The electrical
model consists of membrane capacitance, nonlinear conductances for Na+
and K+, and a linear leakage element. For a parallel combination of capaci-
tance and conductance, the current through the membrane is related to the
capacitive and leakage currents by
I m = C m ~~ + (INa + I K + I L) (3.9)
where I m is the membrane current density, Cm is the membrane capacity, V

is the membrane voltage, and INa, I K, and h are the ionic current densities.
The ionic terms are expressed by
INa = gNa(V - VNa ) (3.10)
IK = gK(V - VK) (3.11)
h =gL(V - VL) (3.12)
where gNa' gK' and gL are the ionic conductances, and V Na , V K, and V L are the
ionic Nernst potentials. The gL conductance is a linear conductance; the
other two conductances are more complex nonlinear functions of the form
gNa = gNa m3h (3.13)
- 4
gK =gKn (3.14)
where gNa and gK represent the maximum conductance values; and m, n, and
h are so-called activation and deactivation variables that modulate the
maximum conductances. The m, n, and h variables are governed by the first-
order differential equations:
Outside
Membrane
FIGURE 3.7. Hodgkin-Huxley membrane model.

~~ = a n(1 - n) - f3nn (3.15)
~7 = am (1 - m) - f3 m m (3.16)
~~ = a h (1 - h) - f3 h h (3.17)
The a and f3 terms in Eqs. (3.15) to (3.17) are functions of the membrane
voltage. At the experimental HH temperature of 6 ec, the a and f3 constants
are
0.01(10 - ~V)
(3.18)
an = -ex~p[(-,--1O..:....~---:V)---'-/1--=0]:-'---1
f3n = 0.125 exp( - ~~) (3.19)
0.1(25 - ~V)
(3.20)
(3.21)
ah = 0.07 exp (20

-~V) (3.22)
p, ~ [ex~30 ~OAV) + { (3.23)
In Eqs. (3.18) to (3.23), ~ V is the change in membrane potential, expressed

in millivolts, relative to the resting potential. Solutions to Eqs. (3.15) to
(3.17) for a step change in ~ V can be obtained in the form:
(3.24)
(3.25)
(3.26)
10 .........,r-------,.------,.-----, 1.0
o L..-""-O=::::::;"_ _ _...L..._ _ _ _....I-_ _ _ _--! 0

1 1.0
en
.sE 0.5
~
o b""j1..::::::._-=:::::::::.L_ _ _ _1-_ _ _..J 0

10 1.0
o LJ_ _ _ _....L_ _...::::::::=......_ _ _-.I 0

Vr - 50 Vr Vr + 50 Vr + 100
Membrane potential (mV)

FIGURE 3.8. Relationship of m, n, and h constants to the membrane voltage. V,
represents the resting potential. (Adapted from Stein, 1980.)
where n(O) = no, and n(t ~ 00) = n",. Expressions similar to Eqs. (3.24) to
(3.26) are obtained for met) and h(t). The m, n, and h variables are con-
strained between 0 and 1, and can be regarded as the fraction of ion gates
open at anyone time. As indicated by Eqs. (3.13) and (3.14), these gates
modulate the maximum conductance of Na and K. The am, an> and 0h'
variables determine the rate at which the gates can open and close. The
asymptotic values of the m, n, and h variables, and the associated time
constants, are all functions of membrane depolarization voltage as illus-
trated in Fig. 3.8 (after Stein, 1980).
Depolarization of the membrane is necessary for excitation. Figure 3.9
illustrates the membrane events accompanying excitation of the HH
40
35 90
'" 30
.E 80 >E
~
0 70 -;
..c
E 25 60 ~
§.
20 50 g
B
c: 40 B
III 15 c:
0 e!
30 .J:J
:l
"0 10
c:
0
20 ~
()
5 10 ::::i!
0
0.5 1.5_ 2.0 2.5 3
Elasped time (ms)
1000
'"E 800
!
600
400
C 200
~
5(.) 0
~ -200
til
15 -400
E
~ -600
-800
0 0.5 1.5 2.0 2.5 3
Elasped time (ms)
FIGURE 3.9. Membrane events during propagating action potential. The top figure
shows the membrane voltage change (Ll V) and the conductances gNa and gk. The
bottom figure shows the sodium current (INa) potassium current (IK) and total ionic
current (IJ The positive axis indicates the ionic current influx. (Adapted from
Hodgkin and Huxley, 1952.)
membrane. The upper part shows the AP voltage waveform along with the
membrane conductance. The lower part shows membrane current-the
positive axis refers to ionic current influx. An initial surge of Na+ influx
serves to further depolarize the membrane; this influx is followed by K+
efflux, which repolarizes the membrane.
The Frankenhaeuser-Huxley Membrane

The FH equations for the myelinated membrane use four ionic terms, in
contrast with the three for the HH unmyelinated membrane. The FH ionic
current densities are
(3.27)
The terms INa, I K , and IL have the same interpretation as they do with the
HH membrane. The term Ip was described as a "nonspecific" ionic compo-
nent that responds to the concentration gradient of Na+. Whether Ip repre-
sents a second type of Na+ channel, or another ionic species, was left
unresolved by the FH researchers. In the FH membranes, the individual
ionic current densities have the expressions:
I =
_ hm 2 (EF2
P __
) [Na]0 - [Na]., exp(EF/RT) (3.28)
Na Na RT 1 - exp(EF/RT)
I = P. n2(EF2)[KL - [Klexp(EF/RT) (3.29)

K K RT 1 - exp(EF/RT)
I = P 2(EF2) [NaL - [Nalexp(EF/RT) (3.30)

p pP RT 1- exp(EF/RT)
IL = gL(V - VL) (3.31)
where E = V - V" V is the membrane potential, and Vr is the

resting potential. The variables m, n, h, and p are defined by
differential equations of a form identical to Eqs. (3.15) to (3.17). The a and
fJ variables for the m, n, h, and p rate terms of the FH equations have the
form
where the a, b, c, and d parameters are provided by Frankenhaeuser

and Huxley (1963). Specific constants for the FH equations are given in
Table 3.2.
TABLE 3.2. Constants for FH equations.

Constant Value Description
PNa 8 X 1O- 3 cm/s Sodium permeability constant
PK 1.2 X 1O-3 cm/s Potassium permeability constant
Pp 0.54 X 1O-3 cm/s Nonspecific permeability constant
gL 30.3mS/cm2 Leakage conductance
VL 0.026mV Leakage equilibrium potential
[Na]o 114.5mM External sodium concentration
[Na]; 13.7mM Internal sodium concentration
[K]o 2.5mM External potassium concentration
[K], 12mM Internal potassium concentration
F %,514C/gmol Faraday constant
R 8.3144 JIK mol Gas constant
T 295.18K Absolute temperature
V, -70mV Resting potential
Cm 2p,F/cm2 Membrane capacitance per unit area
Initial conditions: m(O) = 0.0005, h(O) = 0.8249, nCO) = 0.0268, p(O) =

0.0049.
Source: Frankenhaeuser and Huxley, 1963.
Differences between the FH and HH equations reflect the specific prop-

erties of the myelinated and unmyelinated membranes. The activation vari-
ables have different powers, and the FH equations have an additional ionic
term. The ionic terms appear more complex in the FH membrane as com-
pared to the HH membrane. Despite the seemingly greater complexity of
the PH membrane, the electrical properties associated with the myelinated
membrane's AP development are very similar to the HH membrane. One
observed difference includes a somewhat longer AP duration for the HH
membrane. In addition, the HH equations respond to a prolonged current
stimulus by producing multiple APs, whereas the FH membrane produces
a single AP.
Other differences are seen between myelinated (A fiber) and unmyeli-
nated (C fiber) response when the overall excitatory behavior of the neuron
is considered. Some of these differences include faster conduction rates and
lower thresholds to external currents for A fibers (Ruch et aI., 1968). The A
fiber is an ideal one to model for electrical stimulation studies. Because of
its lower excitation threshold for external current stimulation, the A fiber
class will generally determine the limiting value for threshold currents.
Nevertheless, it is important to consider the potential role of both fiber
classes in electrical stimulation, and their functional properties in sensation
and muscle response.
Species Dependence of Action Potential Dynamics

The HH and FH equations were developed from experiments On non-
mammalian species-the HH experimenters used the giant axon of a squid,
and the FH experimenters used the myelinated fiber of a toad. These

species were chosen primarily for the large size of their axons, and relative
ease of experimentation. The principles of membrane dynamics developed
in that work have been found to be applicable to other species, including
mammals. A quantitative description of membrane currents in rabbit myeli-
nated nerve was developed by Chiu et al. (1979) using the HH experimental
approach and mathematical framework. They showed that, as in squid and
frog nerve, a transient inward sodium current was responsible for the ini-
tially rapid membrane depolarization. However, in contrast to squid and
frog nerve, repolarization of the rabbit nerve was the result of outward flow
of only a single passive ionic component, and potassium current outflow was
virtually absent. Despite the simpler description of the rabbit nerve, its
calculated AP was very close to that of the frog nerve in both amplitude and
time course.
Propagation of Nerve Impulses

Action potential propagation can be understood by referring to Fig. 3.10.
Consider that point A on the axon is depolarized. The local point of depo-
larization causes ionic movement between adjacent points on the axon, thus
propagating the region of depolarization. If depolarization were initiated
from an external source on a resting membrane at point A, an AP would
propagate in both directions away from the site of stimulation. Normally,
however, an AP is initiated at the terminus of the axon and propagates in
only one direction.
After the membrane has been excited, it cannot be reexcited until a
recovery period has passed. This period is termed the refractory state of the
membrane. Before full recovery, the membrane becomes partially refrac-
tory; that is, it requires a stronger depolarizing force to become excited. The
refractory property is principally the result of the prolonged decrease of
the sodium deactivation variable, h (Fig. 3.8), which effectively turns off
the membrane's sodium conductance until the passage of a recovery period
that extends beyond the repolarization process.
Refractory behavior in frog nerve is illustrated in Fig. 3.11 (adapted from
Katz, 1966), which shows the membrane response to an external current
stimulus. An initial stimulus, applied at t = 0, results in response a. Succes-
sive stimuli lead to responses b to g. The absolute refractory period exists
from the AP spike to somewhat within the negative after potential. During
this period, a second stimulus, no matter how strong, fails to produce a
response. In a mammalian nerve at body temperature, the neuron is abso-
lutely refractory for typically about 0.5 ms. Afterwards, for a period of
several milliseconds, the nerve is relatively refractory. During this period,
an increased stimulus is needed to produce a response, and that response
will initially be feeble, as noted by responses b to e. After several millisec-
onds, the neuron fully recovers from its less excitable state.
FIGURE 3.10. Spread of the depolarization wavefront. Depolarization occurring in

region A results in charge transfer from the adjacent regions.
The refractory recovery period sets an upper limit on the number of APs
per second that the membrane can support. With externally applied electri-
cal stimuli, an AP rate cannot be produced beyond about 2,000 per second.
In natural conditions in the body, the repetition rate rarely exceeds 500
per second and is more typically in the range from 10 to 100 per second
(Brazier, 1977, Hoffmeister et at, 1991).
The AP velocity depends on the rate at which electrical charge is trans-
ferred from the locus of excitation to the region of membrane ahead of the
AP. The charge-transfer rate, in turn, depends on the membrane capacity
and the longitudinal resistance of the axon. Depending on the assumptions
o 3 6 9
Elapsed time t (ms)
FIGURE 3.11. Illustration of the refractory period in frog nerve. The initial stimulus
was applied at t = 0, resulting in response a. Subsequent stimuli were applied at
various time delays, resulting in responses b through g. (Adapted from B. Katz,
Nerve, Muscle, and Synapse, 1966, reproduced with permission of The McGraw-Hill
Companies. )
TABLE 3.3. Characteristics of A and C fibers.

Fiber class
A C
Fiber diameter Cum) 1-22 0.3-1.3
Conduction velocity (m1s) 5-120 0.6-2.3
AP duration (ms) 0.4-0.5 2.0
Abs. refractory period (ms) 0.4-1.0 2.0
Velocity/diameter ratio (m1sflm) 6 -1.7
Myelinated Yes No
Source: Adapted from Ruch et al. (1968).
made, theoretical arguments suggest that conduction velocity ought to vary

either as the square root of, or linearly with, fiber diameter (Paintal, 1967).
Experimental evidence demonstrates that conduction velocity indeed in-
creases with fiber diameter, and experimental data are frequently repre-
sented in terms of the ratio of conduction velocity to fiber diameter
(Paintal, 1973).
The myelinated fiber is nature's means of obtaining fast conduction ve-
locity without requiring unduly large fibers. The effective area requiring
capacitive charging is limited mainly to the myelin-free internodes. Further-
more, the depolarization process is saltatory; that is, it jumps from node to
node. As a result, propagation can proceed at a much faster rate than would
be the case with an unmyelinated fiber of the same diameter.
Table 3.3 lists some general characteristics of A and C fibers. The myeli-
nated A fibers are sometimes subdivided into diameter classes designated
A o' A p, and Aa. The Ao fibers are typically related to cutaneous pain and
temperature sensation, the Ap fibers are related to mechanoreception, and
the Aa fibers are related to proprioception and contraction of striated
muscle (Li and Bak, 1976). Figure 3.12 illustrates the distribution of afferent
A fibers (Kandel et aI., 1991). Unmyelinated fibers have a range more
typically from 0.3 to 1.3.um. The distribution of efferent neurons shows
prominent clusters from 12 to 20.um, and from 2 to 8.um, with a pronounced
nadir in the range from 8 to 12.um (Ruch et aI., 1968).
The distribution of myelinated fibers will differ substantially in the
Central Nervous System (CNS) as compared with peripheral nerves. For
instance, in the human pyramidal tract, 89.6% nerves were found in the
diameter range from 1 to 4.um, 8.7% from 5 to lO.um, and 1.7% from 11 to
20.um (Lassek, 1942).
3.4 Action Potential Models for Cardiac Tissue

The electrical properties of excitable cardiac tissue have been defined using
the experimental techniques and mathematical formalism that lead to the
HH model. The membrane electrodynamics are specialized for different
Action Potential Models for Cardiac Tissue 91
cardiac tissue. One common feature is a prolonged excited state, as com-

pared with a much shorter period in nerve tissue.
The HH experimental techniques and mathematical formalism were ap-
plied to cardiac Purkinje fibers by Noble (1962). The Noble model used
three ionic components, including a nonlinear sodium term similar to that in
the HH equations, a nonlinear potassium term that exhibited a rectification
process, and a nonspecific linear "anion" term. This model was later modi-
fied and extended by McAllister, Noble, and Tsien (1975) (referred to as
the MNT model). The MNT Purkinje model is much more complex than
the Noble Purkinje model-it includes nine separately described ionic com-
ponents, along with five activation and deactivation variables. Some of the
ionic terms in the MNT model are analogous to the HH nonlinear compo-
nents; others are analogous to the HH linear leakage terms. The MNT
model was modified and adapted for more efficient computer simulation by
Drouhard and Roberge (1982b).
Further modifications and elaboration's of the MNT model were made by
DiFrancesco and Noble (1985) (DN model). This model included no less
Muscle nerve
...
III C
Q)
~
AS
...
'0
Q)
.D
E
:J
Z
C Cutaneous nerve
o 2 4 6 8 10 12 14 16 18
Fiber diameter (11m)
FIGURE 3.12. Distribution of peripheral afferent, nerve fibers: lower, cutaneous;

upper, muscle. (Adapted from Boyd and Davey, 1968, with additional data by
Martin, 1991.)
50r-----~------~----~----~----~
:;-
.S-
iii
~ 0
~c.
Q)
c
...
III
.0
-50
E
Q)
E
-100
0 2.5
Elapsed time (5)
FIGURE 3.13. Computed action potential of cardiac Purkinje cell based on

DiFrancesco-Noble model (From DiFrancesco and Noble, 1985.)
than 11 separate ionic components. A further feature of the DN model

consisted of nonlinear differential equations governing the intracellular and
extracellular ionic concentrations. None of the other models, including the
various nerve membrane models, includes this feature. Figure 3.13 illus-
trates the simulated AP of the DN model, showing pacemaker activity.
A review of the ionic processes in the Purkinje fiber and their role in the
development of computational models was presented in a tutorial paper by
Noble (1984). He draws attention to the growth over time of the complexity
of Purkinje computational models, and remarks: "It is sincerely to be hoped
that this is not a case of indefinite exponential growth, or, by the year 2000,
we shall all have great difficulty in explaining cardiac excitation to our-
selves, let alone to students with formidable memories" (p. 42).
The various Purkinje models may be regarded as representing differing
degrees of precision in the treatment of ionic processes. From the point of
view of electrical stimulation studies, it is not clear how much of this
precision is required. At a minimum, we would like to reproduce the pace-
maker AP process and simulate evoked extrasystolic excitation. Even the
simplest model (Noble, 1962) gives a good reproduction of the Purkinje AP
and its pacemaker action. Evoked extrasystolic excitation has been demon-
strated in both the Noble (1962) and MNT (McAllister et aI., 1975) papers.
Figure 3.14 illustrates the response of the Noble model to repeated stimu-
lation at a rate of 3/s. The alteration in the duration of the evoked APs seen
in Fig. 3.14 is similar to experimental observations (Trautwein and Dudel,
1954; Geddes et aI., 1972).
The DN equations described above were modified by Noble and Noble
(1984) to describe electrical activity of the sinoatrial node. Beeler and
Reuter (1977) used the HH formalism to develop a mathematical model for
the action potential of ventricular myocardial fibers (referred to as the BR
Action Potential Models for Cardiac Tissue 93
> 50
.s
(ij
'';:::;
...c:
0>
0
0
c.
~ -50
...
CtI
.0
E
~ -100 ~--~--~~~ __~~~__~__~__~L-__L-~
Elapsed time (s)
FIGURE 3.14. Effect of repetitive stimulation on the computed AP in the Noble

Purkinje model. Arrows indicate application of stimulus at the rate 3/s. (From
Noble, 1962.)
model), and later improved by Mogul and colleagues (1984). An interesting

property of the BR model is its ability to produce oscillatory potentials
when depolarized by steady outward current, as seen in Fig. 3.15. The
authors point out that although the electrical properties of the myocardium
seldom produce spontaneous activity, numerous experimental studies with
steady depolarizing current have demonstrated oscillatory AP behavior,
such as seen in Fig. 3.15.
To adequately model cardiac excitation, one needs to spatially couple the
temporal electrodynamics of individual cells to account for the spatial and
40 . .-----.----.-----r----,,-----.
:;-
S
]i o
C
0>
'0
c.
0>
c: -40
~
.0
E
0>
E
-80
o 2 3 4 5
Elapsed time (s)
FIGURE 3.15. Oscillatory AP response of Beeler-Reuter model for ventricular myo-

cardial fiber in response to steady outward current. (From Beeler and Reuter, 1977.)
temporal distribution of the stimulus. In the case of nerve cells, one can use
cable models, such as the SENN model described in Chapter 4. For cardiac
tissue, the bidomain model treats cardiac tissue as two coupled, continuous
domains, one for the intracellular space, and one for the interstitial space
(Henriquez, 1993).
3.5 Sensory Transduction2

The body is equipped with a vast array of sensors (receptors) for monitor-
ing its internal and external environment. The receptor converts a stimulus
to an electrical potential that initiates a propagating AP. Receptors are
specialized to respond most efficiently to a specific type of stimulus, al-
though they may also respond to a variety of stimuli if the intensity is great
enough. We shall be concerned here with the somatosensory system, i.e., the
system of receptors found in the skin and internal organs. There are other
specialized receptors in the visual and auditory systems, chemical receptors
for taste and smell, and other special chemical receptors by which neurons
communicate with one another.
The somatosensory receptors can be classified as: mechanoreceptors,
thermo receptors, chemoreceptors, and nociceptors. Numerous specializa-
tions of mechanoreceptors respond to specific attributes of mechanical
stimulation. Thermoreceptors are specialized to respond to either heat or
cold stimuli. Nociceptors are unresponsive until the stimulus reaches the
point where tissue damage is imminent, and are usually associated with
pain. Many nociceptors are responsive to a broad spectrum of noxious
levels of mechanical, heat, and chemical stimuli (Ruch and Patton, 1979).
Both myelinated and unmyelinated nociceptors have been identified.
Figure 3.16 illustrates several cutaneous mechanoreceptors. The pacinian
corpuscle is a so-called rapidly adapting receptor because it responds to the
onset or termination of a pressure stimulus. As such, it transduces accelera-
tion of skin displacement (Schmidt, 1978). The Meisner corpuscle is a less
rapidly adapting receptor, and is most responsive to the velocity of skin
displacement. The Rufini ending and tactile disks are slowly adapting, and
respond most efficiently to steady pressure. Figure 3.16 also shows hair
receptors that respond to displacement of hair follicles. Other mechanore-
ceptors, such as stretch receptors and muscle spindles, respond to move-
ment and position of the muscles, and serve to monitor and regulate body
posture and movement. An additional class of receptors, known as free
nerve endings, appear as branching structures on the end of the neuron.
Free nerve endings are involved in a variety of specialized transduction
processes in both nociceptive and non-nociceptive classes. It is not always
2The following are useful references for this section: Kandel and Schwartz (1981);
Lamb et al. (1984); Ruth et al. (1968), Ruch and Patton (1979).
Sensory Transduction 95
Free
nerve
Dermis ending
Fatty
tissue
Hair Pacinian Iggodome

receptor corpuscle
FIGURE 3.16. Morphology of several mechanoreceptors in the hairy skin.
clear from morphological evidence what distinguishes the transduction

specializations of free nerve endings.
Intensity Coding
When a receptor is stimulated, it produces a voltage change called a genera-
tor potential at the terminal ending of its axonal connection. Unlike the all-
or-nothing response of the axon, the generator potential is graded: if you
squeeze a pacinian corpuscle, it produces a voltage at the axon terminus; if
you squeeze it harder, it produces a greater voltage.
The generator potential initiates one or more APs that propagate along
the axon. The AP repetition rate will depend, in part, on the intensity of the
stimulus. Figure 3.17 illustrates the response rate of a slowly adapting (Fig.
3.17a) and an intermediately adapting (Fig. 3.17b) receptor (from Schmidt,
1978). The slowly adapting receptor responds to a constant pressure stimu-
lus, although the AP rate is greatest at the onset of the stimulus. The
intermediately adapting receptor AP rate is more or less constant during a
constant-velocity pressure stimulus. For a rapidly adapting pacinian cor-
puscle, a single AP is produced at the onset or termination of a pressure
stimulus. However, for a vibratory stimulus, a train of APs is produced at
the rate of the vibration frequency. The intensity of the threshold amplitude
~!F a.
100
~ 20
--
"iii"
en
en
10
OS~s
Q)
J!l
~
c. .5
a.. 5
:§. 10 <
'0
Q)
...
iii ...
Q)
2
a.. .c
<
z
E
~ •
0.01 0.1 1.0 10.0 100 0.2 0.5 1 2 5 10
Stimulus force (N) Indentation rate (~m I ms)
(a) (b)
FIGURE 3.17. Response of (a) slowly adapting receptor and (b) intermediate adapt-
ing receptor to constant force stimulus. Part (a) shows AP rate of various time
delays (0) after onset of pressure stimulus. Part (b) indicates total number of APs
during O.5-s constant velocity stimulus. (Adapted from Schmidt, 1978.)
of the sinusoidal mechanical stimulus on the skin is shown in Fig. 3.18.

Sensitivity to vibratory mechanical stimuli on the skin is greatest in the 200
to 300-Hz range. Anesthetization of the skin elevates thresholds only for
frequencies below about 50 Hz. This can be explained by two modes of
transduction: low-frequency sensations are the result of more superficial
104~--~----~----~-----r------~----~----'
Cutaneous response Meissner corpuscle
3 10 30 100 300 1000 3000

Vibration frequency (Hz)
FIGURE 3.18. Threshold response for vibratory mechanical stimulus applied to the
skin. (Reprinted by permission of the publisher from Martin, 1991. © 1991 by
Elsevier Publishing Co. Inc.)
Muscle Function 97
60 60
en
('t)
Cl
c: 40 40
/
.;::
::J
'0
en
Q)
en
:; 20 20
c.
E
0 0
45 47 49 51 53 0 40 80 120
Stimulus temperature (Oc) Stimulus pressure (gm)
FIGURE 3.19. Response of myelinated nociceptive afferents innervating monkey

hand. Stimulus duration is 3 s for both types of stimuli. (From Campbell et ai., 1979.)
hair follicle receptors or Meissner corpuscles; high-frequency vibration is

detected by pacinian corpuscles located in deeper strata (Martin and
Jessell, 1991).
Figure 3.19 illustrates AP rate versus stimulus intensity for unmyelinated
nociceptors (Campbell et aI., 1979), showing a monotonic relationship be-
tween the average AP rate and intensity of both heat and pressure stimuli.
Other studies (LaMotte et aI., 1984) demonstrate a monotonic relationship
between heat intensity and both AP rate and pain rating in nociceptive C
fibers. Although AP rate is important in sensory encoding of stimulus
intensity and painfulness, there are a number of other important factors as
well. These include the spatial and temporal patterns of stimulation, and the
number of neurons brought to excitation (recruitment). Chapter 7 provides
additional discussion of the role of these factors in electrical stimulation.
3.6 Muscle Function

The previous section discussed the role of afferent neurons in conveying
information from the body's sensory system to the central nervous system
(CNS). Efferent neurons carry information from the CNS to the muscles to
effect contraction. Figure 3.20 illustrates structural features of the neuro-
muscular junction at the end plate of the muscle (after Birks et aI., 1960;
Woodbury et aI., 1966). When the AP reaches its terminus at the end plate,
a chemical neurotransmitter is released across the nerve/muscle gap, which
causes depolarization of the muscle cells. The result is that the muscle
membrane is excited, and a depolarization wave is propagated in the muscle
away from the end-plate region.
Figure 3.21 illustrates the anatomical structure of muscle, showing
progressively smaller systems of structural elements. When the muscle is
Terminal axon
branch
FIGURE 3.20. End-plate region of frog muscle fiber. (After Birks et aI., 1960.)
excited, the individual fibrils (shown in the lower part of the figure) slide
together resulting in muscle contraction. The muscle illustrated in Fig. 3.21
is called striated because of the microscopically visible striations that arise
from the arrangement of contractile elements. Striated muscle is found in
the skeletal system. Skeletal muscle is under voluntary control, and is
innervated by the somatic nervous system. Smooth muscle differs from
striated muscle in that the characteristic cross-striation pattern is absent.
Smooth muscle is involuntary, and is found in vessel artery walls, air pas-
sages of the lungs, and in various tissues of the intestines and reproductive
system.
An action potential launched along a motor neuron results in a single
contractile quantum called a twitch. When a succession of APs is produced
on the motor neuron, the individual twitch quanta fuse as illustrated in Fig.
3.22. In this example, maximum muscle tension is achieved at an AP rate of
about 80/s, leading to a condition of maximum fusion termed tetanus. Gra-
dation of muscle tension results from fusion of individual twitch quanta and
from recruitment (the excitation of additional neurons). Normally, the APs
in individual motor neurons are asynchronous, and muscle tension can be
finely graded. However, with stimulation by externally applied currents of
a repetitive or oscillatory nature, the twitch quanta of various muscle fibers
may be synchronized. Experimental data show that tetanus of the muscle
group occurs at current levels that are only moderately above the twitch
threshold value (Oester and Licht, 1971).
Externally applied electric currents can stimulate muscle by exciting
motor neurons, or the muscle fibers themselves (Fig. 8.3). Direct stimula-
tion of muscle fiber requires much higher currents than does stimulation of
enervated muscle (Harris, 1971; Walthard and Tchicaloff, 1971). As a re-
sult, stimulation of muscle by external electric currents will usually take
Muscle Function 99
place most efficiently through neural excitation. See Chapter 8 for details of
electrical stimulation of muscle and motor neurons.
The propagating excitation wave along the muscle, when recorded
through cutaneous electrodes, is called an electromyogram (EMG). The
EMG signal has been used in various experimental and clinical applications
((((((((((~~ }O-100 Ilm
Fiber
Fibril
FIGURE 3.21. Skeletal muscle and filament structure of striated muscle.

c
o
-
·Ui
c
C1l
C1l
&i::::l
:2!:
o 100 200 300 400

Time (msec)
FIGURE 3.22. Effects of AP rate on muscle tension. [Adapted from McNeal and
Bowman (1985a). Reprinted by permission from Neural Stimulation, vol. II CRC
Press Inc., Boca Raton, FL.]
as an indicator of muscle activity, and can often be more easily detected

than direct observation of muscle movement or tension.
3.7 Synapses
General Properties
Excitable cells communicate with one another across junctions called syn-
apses. An action potential that has traveled along the presynaptic cell
affects the postsynaptic cell through the release of specialized chemicals
called neuro transmitters in the case of chemical synapses, or electrically
(ephaptic transmission) across gap junctions in the case of electrical syn-
apses. Chemical synapses are the more sensitive and common of these two
modes of intracellular communication. Figures 3.1 and 3.20 illustrate a
muscle synapse at the motor neuron end plate. Figure 3.1 also illustrates
central synapses at the dendrites within the CNS. The neurotransmitter that
is released from the chemical synapse flows across the gap, where it binds to
receptor sites in the postsynaptic cell and opens ionic channels that alter the
postsynaptic membrane potential. Through these means, a small presynap-
tic process which generates only a weak ionic current can depolarize a large
postsynaptic cell. If the postsynaptic potential is sufficiently depolarized, an
action potential will be launched.
Both temporal and spatial integration can contribute to the postsynaptic
potential (Dudel, 1989). With temporal integration, repeated action poten-
tials from the presynaptic neuron can have an additive effect on the
Synapses 101
50
-> 20
-E
cti
:;:;
10
-
c::
CD
0
5
a.
CD
~
.0.. 2
(/)
(.)
:;:; 1
a.
CO
c:: 0.5
-
>.
en
( /)
0 0.2
D..
0.1
25 50 75 100 125 150 175 200
Presynaptic spike potential (mV)
FIGURE 3.23. Example of relationship between pre- and postsynaptic potentials with
injected current (t = 1 ms). Length of synapse (d-e): O.8mm; (a) current electrode;
(b) pre-recording electrode; (c) post recording electrode. [Adapted from Katz &
Miledi (1967).]
postsynaptic membrane potential. This occurs because the postsynaptic

membrane has a much longer membrane time constant than does the
presynaptic cell. With spatial integration, many individual neurons can
synapse on a single postsynaptic cell. The additive effects of the spatially
integrated action potentials can be either excitatory or inhibitory in the
postsynaptic cell.
Experiments using the giant squid axon demonstrate the relationship
between pre-and postsynaptic potentials in chemical synapses as in Fig. 3.23
(Katz and Miledi, 1967). The insert in the figure shows the experimental
arrangement. In this example, a change in the presynaptic spike potential
from 55 to 65mV (an 18% change) results in a ten-fold change in the
postsynaptic potential from 1 to lOmV.
Synaptic Interactions with In-situ E-fields

Polarization of presynaptic processes due to an in-situ electric field can
result in enhancement or inhibition of postsynaptic action potentials. An
......
0
N
rod
~
t!1
('D
~
::to
n
e;.
'"C
::to
bipolar cell
5.
'E.
('D
ell
a
z
!:!l<:
('D
'"
8-
s;::
ell
=
f2.
('D
~
To {
optic a.
nerve 0
='
light
~ ~ ~
FIGURE 3.24. Structure of the retina showing various layers of neurons consisting of photoreceptors (rods and
cones), horizontal cells, bipolar cells, amarcrine cells, and ganglion cells. (Adapated from Tessier-Lavigne,
1991; Dowling and Boycott, 1966).
The Spinal Reflex 103
example of this effect is attributed to the phenomenon of electro- and

magnetophosphenes, which are the visual effects resulting from electric
currents or magnetic fields applied to the head. Experimental evidence
discussed in Sect. 9.8 suggests that phosphenes are generated through modi-
fication of synaptic potentials in the receptors or neurons of the retina. The
retina is rich in the synaptic processes of photoreceptors and neurons that
comprise a visual processing system. Figure 3.24 (adapted from Tessier-
Lavigne, 1991; Dowling and Boycott, 1966) illustrates the organization of
neurons within the retina, showing neuronal structures that have both
tangential and radial orientations with respect to the retina. Phosphenes are
most sensitive to current or an in-situ E-field that is oriented in a radial
direction. This finding suggests that the E-field interacts with radially
oriented neurons in accord with cable theory (Sect. 4.3).
The electrical thresholds and the temporal response of synaptic effects
within the retina differ significantly from corresponding properties of nerve
and muscle excitation. Considering the observed magneto-phosphene
thresholds at the most sensitive frequency (20Hz), the minimum in-situ
electric field that produces phosphenes is approximately a factor of 100
below the rheobase thresholds for nerve excitation (refer to Chapter 4).
Also, the strength-duration time constant for phosphene stimulation is
approximately 100 times greater than that for nerve excitation and 10 times
greater that for excitation of muscle cells. It does not necessarily follow that
the low thresholds associated with phosphenes necessarily apply to other
neural synapses because of the highly specialized configuration of neurons
in the retina. However, experimental evidence discussed in Sect. 9.8 shows
that short-term CNS responses can be seen with in-situ E-fields that are
below the threshold of neural excitation. Although synaptic polarization
could have important consequences in electrical stimulation of the CNS,
the implication of these findings on neural reactions to externally applied
currents or fields has been little studied.
3.8 The Spinal Reflex

Muscle movement can be electrically stimulated through excitation of mo-
tor neurons or direct excitation of muscle fibers. Reflex activity represents
a third mechanism whereby muscle movement may result from electrical
stimulation. Here, stimulation may be initiated at cutaneous or muscle
sensory afferents, with muscle movement ensuing through the reflex arc.
When we touch a hot object, for example, our hand jerks away before we
have consciously appraised the situation. Afterward, deliberate action takes
place. The initial response, in this case, takes place outside of conscious
control through the spinal reflex.
Figure 3.25 illustrates organizational features of the spinal reflex arc.
Sensory inputs, originating from cutaneous or muscle receptors, communi-
Sensory afferents
o ~
S ner ists
m~s~le {
efferents ~-A-n-ta-g-o-n-is-ts--------_../
FIGURE 3.25. Organizational features of reflex arc. Small circles represent receptors;
triangles represent synaptic terminals. Synaptic summation occurs at dendritic
processes within spinal column.
cate with muscle efferents through synaptic processes in the spinal column.
Intermediate neurons (interneurons) may also exist within the reflex arc.
Inputs from higher centers may also be included. These modify the reflex
action based on central processes-for example, we can deliberately inhibit
the reflex action to maintain contact with a hot object. The synaptic summa-
tion may combine inputs as either inhibitory or excitatory, such that the
presence of a synaptic output will depend on a weighted summation of
inputs, thus forming a computing system within the spinal column. The
output of this system consists of excitatory and inhibitory signals to muscle
groups to cause coordinated, patterned movement.
The knee-jerk reflex is a simple example of a monosynaptic type in that
it omits the interneurons such that the synapses of sensory inputs directly
activate motor neurons. The motor neurons, in turn, produce AP signals
that cause contraction in synergistic muscle groups, and inhibit AP signals
to cause relaxation in antagonistic muscle groups.
More complex reflex activity is conditioned by experience. Some reflex
actions are shared by individuals, regardless of prior experience, such as
deep tendon reflexes, the eye blink response, and startle reflexes. Section
7.11 discusses experimental data on startle reactions to electrical stimuli.
4
Excitation Models
4.1 Introduction
Variations in the stimulus waveform or in the electrode arrangement can
result in vastly different stimulation thresholds. To fully understand the
factors responsible for electrical stimulation, computational models must
connect features of the stimulus current with properties of excitable tissue
described in Chapter 3. However, a simpler approach based on a linear,
spatially limited membrane can provide insight into underlying electrical
processes, and also provide closed-form mathematical expressions for some
excitation relationships. This chapter begins with simple linear analysis
models, and proceeds to a more complex, spatially extended nonlinear
model.
The models treated here are essentially electrical cable representations
of single fibers. Clearly, electrical stimulation normally involves a much
more complex macroscopic system. Both sensory and motor responses
depend on the spatial and temporal patterns of stimulation; cardiac re-
sponses involve a three-dimensional dynamic system. Nevertheless, the
response due to large-scale excitation is a result of individual excitable
elements. As will become apparent in subsequent chapters, electrical
responses attributable to individual fibers can explain many observed
macroscopic responses. Thus, single-fiber excitation models provide power-
ful analytic tools for evaluating a wide range of electrical stimulation
properties.
4.2 Linear Strength-Duration Model

A simple analysis model treats an isolated segment of an excitable fiber as
a linear electrical circuit as shown in Fig. 4.1. The membrane is assumed to
consist of capacitance em and resistance r m. The simplified analysis circuit is
presumed to apply to a small patch of excitable membrane. The stimulus
current flowing across the membrane is depicted as originating from a

106 4. Excitation Models
i (t) ---+
c R
FIGURE4.1. Simplified linear circuit model for isolated patch of excitable

membrane.
current source. The resting potential is treated as OV, and depolarization is

analyzed as the voltage rise across the electrical network. In a simplified
analysis, r m is treated as a constant up to the threshold of the action poten-
tial. In actuality, because of the nonlinear membrane properties described
in Chapter 3, rm is relatively constant only up to about 80% of the action
potential (AP) threshold (McNeal, 1976).
The equations governing the response of the RC model are
i{t) = ic{t) + iR{t) (4.1)
~
c
f((t)dt =
m
iR{t)R (4.2)
where i(t) is the total current, ic(t) is the capacitive displacement current,
and iR(t) is the current flowing in the resistance R. Consider that this simple
circuit is excited by a step current pulse having the form
i{t) = {Io t? 0
t< 0
(4.3)
where I is the peak current. Equations (4.1), (4.2), and (4.3), may be readily
solved (such as with Laplace transform methods) as
(4.4)
where Tm = rmCm; Tm is called the membrane time constant.
Assume that there is a single depolarization voltage, V r. needed for
excitation. This is a simplification, since the threshold depolarization volt-
age rises for very short pulses (e.g., <10,us) (Dean and Lawrence, 1983), or
Linear Strength-Duration Model 107
for biphasic stimuli (Reilly et at, 1985). For a pulse of duration t, the
maximum depolarization voltage is evaluated from Eq. (4.4). The threshold
current (IT) required to drive the transmembrane potential to V T can be
derived from Eq. (4.4) as
(4.5)
The threshold current attains a minimum value (10) for an infinitely long
pulse. As t ~ 00, I ~ 10 = VIR, and Eq. (4.5) may be expressed as
(4.6)
The threshold charge, Qn necessary for achieving the depolarization

voltage, V n is
(4.7)
The threshold charge attains a minimum value Qo, for short-duration

pulses. As t ~ 0,
(4.8)
The expression for normalized depolarization charge is
QT = tl1:m (4.9)
Qo 1 - e-tlTm
The energy dissipated by the rectangular stimulus is given by I2Rt (or,
equivalently, IQR), where R is the resistance in the current path. It follows
that the threshold energy, En is given in normalized form as
ET
IgR
= t
(1 - e-tlTm r (4.10)
As demonstrated by Pearce and colleagues (1982), the stimulus energy

attains a minimum value, Eo, when t = 1.251:,
Eo = 2.46I~1: e (4.11)
Figure 4.2 illustrates the normalized threshold current, charge, and
energy versus the normalized duration (tl1:m ).
A parallel treatment has been derived (Blair, 1932a, 1932b; Reilly et at,
1983) for exponentially decaying current pulses of the form
i(t) = {Ie-tiT for t? 0 (4.12)

o fort< 0
In Eq. (4.12), 1: represents the time constant of the exponential current
pulse. Such exponential currents are encountered with capacitive discharge
UJ
o
d
...:
i 10
.~
iii
E
o
z
1~~~~~~~~~~~~~~~~~~~~
0.01 0.1 10 100
Normalized stimulus duration (V1e)
FIGURE 4.2. Calculated strength-duration relationships for square-wave mono-

phasic current.
stimuli (see Chapter 2). The solutions for normalized threshold current and
charge are
(4.13)
(4.14)
where p = rh:m • When rectangular and exponential stimulus criteria are

compared, Eqs. (4.9) and (4.14) both converge to the same minimum charge
thresholds for short durations, and Eqs. (4.6) and (4.13) converge to the
same minimum peak current threshold for long durations (Reilly and
Larkin, 1983). This observation is equally valid when a more complete
nonlinear model is used to represent the excitable fiber (see Fig. 4.14).
The simple linear model points out some important features concerning
electrical thresholds for monophasic currents. For short-duration stimuli,
the amount of charge transferred is the fundamental quantity defining
Linear Strength-Duration Model 109
stimulus potency. For brief monophasic transients, the details of the wave-
shape are relatively unimportant both rectangular and exponential stimuli
converge to the same minimum threshold charge. For long-duration
monophasic currents, the peak current is the main determinant of electrical
thresholds. Stimulus energy alone is not a meaningful index of excitability,
as has been supposed by some in the past.
Empirical Strength-Duration Relationships

An empirical strength-duration (S-D) relationship for excitation of nerves
was first derived by Weiss (1901). We can marvel today at the ingenuity of
this early researcher, who, in those days of only crude electronics and
measurement devices, was able to perform delicate and precise experiments
that retain validity today. Expressions equivalent to the Weiss formulation
can be stated as
I = 1(1 + !L)
TOt
(4.15)
and
(4.16)
where IT and QT are threshold current and charge, respectively, 10 is the

minimum threshold current for long pulses (t ~ (0), Qo is the minimum
threshold charge for short pulses (t ~ 0), t is the duration of the pulse, and
re is an experimental parameter related to the time response of the tissue
being excited. Equation (4.16) is the expression derived by Weiss; Eq. (4.15)
follows from the relationship QT = Irt. Equation (4.15) is sometimes re-
ferred to as a "hyperbolic" relationship.
Lapicque (1907) studied the data of Weiss and proposed what he called
a "logarithmic" relationship. Lapicque's formulation can be expressed as
(4.17)
Equation (4.17) is now referred to as an "exponential" relationship. It is

the same as that derived for the simple linear model [Eq. (4.6)], with the
exception that Eq. (4.17) uses the empirical time constant, reo rather than
the membrane time constant, rm' The value of the time constant derived
from the simple product of membrane resistivity and capacity, r mCm' is very
different from the value determined by fitting experimental data to Eq.
(4.17). More will be said about experimental time constants.
Lapicque introduced the term rheobase to describe the minimum thresh-
old current achieved for very long pulses. He also defined chronaxie as the
duration of a threshold current having a magnitude twice rheobase. These

terms have become standard lexicon today. In both the parabolic and
exponential formulations,
Rheobase = 10 (4.18)
Chronaxie may be related to fe in the two formulations as
Chronaxie = fe (hyperbolic formula) (4.19a)
and
Chronaxie = fe In 2 = 0.693fe (exponential formula) (4.19b)
In both formulations, the parameter fe may be determined by the ratio of
minimum charge to minimum current:
(4.20)
Equation (4.20) suggests a simple and practical means of determining the

parameter fe in both the hyperbolic and exponential relationships. It
suggests that One may determine fe without having to trace out the entire
S-D curve-rather, it is sufficient to evaluate the stimulus threshold at only
two durations (one very long and One very short).
Figure 4.3 compares the hyperbolic and exponential expressions for cur-
rent thresholds. A third curve applies to a myelinated nerve model, which
will be described presently. Both hyperbolic and exponential forms of the
S-D relationship have been used to represent experimental data. Experi-
mental curve fits for the two formulas have been compared for nerve
(Bostock, 1983) and cardiac (Mouchawar et aI., 1989) excitation. Over the
range of stimulus durations studied, the two formulas provide reasonably
good curve fits, although one or the other may be marginally preferred in
specific instances. One attraction of the exponential formulation is that it
can be related to the linear RC network model (Fig. 4.1), which provides a
heuristic, albeit crude, physiological model of neural excitation. A more
accurate analysis model would include the nonlinear and spatially extensive
properties of the excitable membrane as described in the next section.
fe can depend strongly on the spatial distribution of stimulus current. The
apparent time constant generally becomes longer as the membrane current
is distributed in a more gradual fashion. This is clearly demonstrated in
Sect. 4.7, and also by Jack et aI., (1983, p. 33) using a linear cable model.
These studies show that the apparent time constant with membrane current
injected at a discrete point may be less than half the value observed when
the current is distributed more gradually along the fiber.
Large variations in 7:e are seen for different types of excitable tissue. An
average for nerve excitation is around 0.27 ms, with a wide experimental
range (see Table 7.1). The average for cardiac excitation is around 3ms, also
with a wide experimental range (see Table 6.2).
Electrical Cable Representations 111
--D- Myelinated model
• Hyperbolic
o
s -0-- Exponential
E
~
:J
U
"0
o
.c
:ll 10
-5
1l
.~
(U
E
o
z
0.01 0.1 10 100

Normalized current duration (If"l"e)
FIGURE 4.3. Strength-duration curves for neural excitation by rectangular mono-

phasic current pulses. Hyperbolic and exponential curves are from Weiss and
Lapicque formulations; thin curve applies to myelinated nerve model.
Experimental data using cutaneous stimulation show that chronaxy val-

ues or S-D time constants obtained with a regulated current stimulator can
be significantly greater than those obtained with a regulated voltage device
(Harris, 1971). It is likely that this difference can be explained on the basis
of the electrical response of the skin as explained in Chapter 2. When using
a constant voltage device, one observes the current waveform to exhibit
a leading spike due to the capacitive coupling of the skin and a nonlinear
response in which impedance drops as voltage is increased. Both of these
phenomena would tend to distort a voltage-regulated S-D curve relative to
a current-regulated curve.
4.3 Electrical Cable Representations

While the simple linear model of the previous section provides useful
insight into the neural excitation process, it is of limited value in attempting
to describe the full range of interrelations between the excitable membrane
and stimulus current. The linear model does not account for the effects of
the spatial distribution of stimulus current, differential sensitivity to anodal
versus cathodal currents, response to biphasic currents, or response to
sequences of pulsed stimuli. To study these factors adequately, we need a
more complex model involving the spatial and temporal interrelationships
of the excitable membrane.
One-Dimensional Cable Models

Excitation of nerve and muscle cells can be analyzed using electrical cable
theory. The cable equations to be applied to this problem were originated
by O. Heaviside in 1876 in connection with the analysis of the first trans-
Atlantic telegraphy cable (Nahin, 1987). This section briefly introduces
cable theory as applied to the excitable membrane. More extensive treat-
ments of this subject can be found in other publications (Rall, 1977; Jack
et al., 1983; Plonsey and Barr, 1988).
An assumption of cable theory is that at any longitudinal position x along
the cable, the inside and outside potentials do not vary with the position
along the cable's circumference. Considering the small diameters of the
cells to be studied «20.um), this assumption is a reasonable one for biologi-
cal applications, and it permits the cable to be modeled in a one-
dimensional fashion, as in Fig. 4.4. The illustration shows the inside as
having a resistance per unit length r i (Q/cm), a distributed membrane resis-
tance from inside to outside having unit length value rm (0 cm), and a
distributed length-proportional capacity em (.uF/cm). Ie and Ii are the cur-
rents flowing externally and internally, Ve and Vi are the external and
internal potentials, re is the external resistance per unit length of the cable
because of the surrounding medium, and im is the membrane current.!
The currents and voltages are assumed to vary with the longitudinal
distance along the cable. The cable equation is expressed in differential
form as
(4.21)
where V m in the membrane voltage defined by Vi Ve. An additional

relationship is that membrane current consists of both ionic and capacitive
components; that is,
1 Unit length resistances are defined such that the internal and external resistance's
increase with length, L, in accordance with R; = r;L and R. = R.L, and leakage
resistance decreases with length as Rm = r,,/ L (Both R; and Rm have units
of ohms). The capacitance between inside and outside over the length L is given by
em = cmL.
~Ie
~Ii
FIGURE 4.4. Electrical model for linear cable.
(4.22)
For a linear cable, I j = Vml'm, where'm is treated as a constant, independent

of V m • When the cable equations are applied to an excitable membrane at or
near the excited state, the ionic term is instead described by the nonlinear
equations introduced in Chapter 3.
An alternative form of cable equation is obtained by combining Eqs.
(4.21) and (4.22), and expressing in normalized coordinates as
a2 vm _ V _ aVm = 0 (4.23)
ax2 m aT
where X = xlA. and T = tlrm are dimensionless space and time variables
scaled to the space constant A. and time constant "Cm of the cell membrane.
The scaling constants are given by A. = ['ml('e + 'j)pl2, and"Cm = 'mCm. Since
, m has units of (Q cm), and'e and'j have units of (Q/cm), it follows that A. has
units of centimeters; "Ce has units of seconds. A. is also called the electrotonic
distance of the membrane; it defines the distance along the membrane that
a steady-state voltage disturbance due to point current injection will decay
to e- 1 of the value at the disturbed location. "Cm similarly defines the time
response of an isolated piece of membrane when a step voltage is applied.
In general, 'j > > 'e and it follows that A. = J'ml'j. Taking an example from
Jack et aI., (1983, pp. 23-24), for a 20-.um-diameter fiber with'j = 30MQ/cm
and'm = 1.6 MQ cm, the space constant is calculated to be 0.23 cm. Space
constants for invertebrate nerve are in the range 0.23 to 0.65 cm (RaIl,
1977).
Consider a cable of finite length 2L placed in a longitudinal static field
of strength E. The steady-state solution for membrane voltage in response
to a nonftuctuating field is given by (Sten-Knudsen, 1960)
V (X) = -EA.( sinhX ) (4.24)

m coshL/A.
1.0r---~---,----~--,---~==-..---~---r--~----'
0.8
~
LlJ
E 0.6
>
a>
0>
<0
"'"0>
Q)
c
e
.s:J
0.4
E
Q)
::2
2 3 4 5
Distance from center ( X = xIA. )
FIGURE 4.5. Normalized membrane voltage of a finite cable immersed in a DC field

of strength E. Cable length = 2 L. Voltage has odd-function symmetry about X = o.
where X = 0 is taken as the center of the fiber, and the ends are at ±L.
Transient solutions of the membrane voltage are presented by Rall (1977).
Figure 4.5 illustrates Eq. (4.24) as a function of distance for several cable
lengths. Vm has odd-valued symmetry about X = 0, and only one-half of the
function is displayed in Fig. 4.5. The symmetrical property means that the
fiber is hyperpolarized along its anode-facing half, and is depolarized along
its cathode-facing half. The maximum membrane voltage is attained at the
ends of the fiber, and has the value
L
V
m
= -EA. tanh-
A. (4.25)
For very long cells (L ~ (0), the terminus membrane voltage is EA.. But even
for fibers of only modest length, that value is closely approached. For
example, with LIA. = 2 (total length = 4A.), the membrane voltage at the
two ends is ±0.964EA.. For an ideal nonconducting membrane, A. ~ 00, and
Eq. (4.25) reduces to Vm = EL, which is the maximum possible voltage
that can be developed across the membrane of an elongated cell of length
2L by a static or low frequency extra cellular electric field. For an ideal

spherical cell of radius r, the membrane voltage is V m = 1.5 Er (Carstensen,
1987).
Figure 4.6 gives a physical interpretation of the distribution of current
flow around an elongated cell that is placed in a medium having a uniform
electric field (i.e., uniform current density). The fiber is presumed to be
oriented parallel to the undisturbed field. The flux lines indicate that the
current through the membrane is greatest near the ends of the fiber. At a
sufficient distance from the ends, internal and external fields become equal,
and there is no further current crossing the membrane.
Results from cable theory show that a longitudinal electric field (or
current flow) is required to excite a nerve fiber. Other theoretical studies
(McNeal, 1976) demonstrate that current flow oriented perpendicular to
the long axis of a nerve fiber is relatively ineffective compared with a
longitudinal orientation. These expectations are verified by a variety
of experimental studies (Ranck, 1975; Bawin et ai., 1986) showing that
both peripheral and CNS neurons are very much less sensitive to a trans-
verse field, as in Fig. 4.7a, as compared with a longitudinal field, as in
Fig.4.7b.
Excitatory effects with CNS neurons are also found to occur with a
parallel alignment of the field and the dendrosomatic axis, but to be insen-
sitive to a perpendicular orientation (Bawin et ai., 1986). However, the
same experimenters found that synaptic interactions do not depend on
the orientation of the dendrosomatic axis, presumably as a result of the
branching structure of the dendritic tree. This finding stands in contrast to
electrical stimulation of synapses within the retina, in which a parallel
orientation of the field and cell axis is required to produce phosphenes
(see Sect. 9.8).
: : :
: : :
FIGURE 4.6. Representation of current flow around elongated cell placed in a me-
dium having a uniform electric field (i.e., uniform current density). The membrane
is assumed to be semipermeable to current flow.
Axon
(a) Longitudinal
(b) Transverse
FIGURE 4.7. Longitudinal and transverse current excitation. An excitable cell is

much less sensitive to a transverse field (b) as compared with a longitudinal field
(a).
Nonlinear Models
Various computational models have been used to study the excitation prop-
erties of nerve fibers. Excitation properties have been studied for a spatially
isolated segment of membrane, in which the current density crossing the
membrane is the driving force. In these models, the membrane conductance
is modeled by the Hodgkin-Huxley or Frankenhauser-Huxley equations
that were introduced in Chapter 3. Despite simplifications, isolated mem-
brane models have provided substantial insights into neuronal excitation
by externally applied currents (Btitikoffer and Lawrence, 1978; Dean and
Lawrence, 1983, 1985; Motz and Rattay, 1986).
More complete representations include assumptions about mutual inter-
actions among adjacent segments of the excitable membrane. One such
model, developed by Cooley and Dodge (1966), uses a discrete cable model
to represent an unmyelinated fiber, with stimulation by an intracellular
electrode. In this model, membrane conductance is governed by the
Hodgkin-Huxley equations. Myelinated fiber models studied by Fitzhugh
(1962) and Bostock (1983) include in a cable model the passive circuit role
of the myelin internode, in addition to the active FH conductance's at the
Myelinated Nerve Model 117
nodes. These models have also been configured to study excitation by

current injection at a single point on the axon.
A difficulty with the aforementioned nonlinear models is that they pre-
sume knowledge of the current waveform and density crossing the mem-
brane. In electrical stimulation problems we may be able to calculate the
current within the medium containing the neuron, but not necessarily
the current crossing the membrane. The waveform of current crossing the
membrane can differ substantially from that in the surrounding medium
(McNeal, 1976). Furthermore, the force driving current into the membrane
is the external field distribution along the axon, which cannot be described
by the current density at a single point. These difficulties have been re-
moved in the myelinated fiber model of McNeal, described in the following
section.
4.4 Myelinated Nerve Model

Myelinated A fibers are distinguished from unmyelinated C fibers by faster
conduction rates, shorter action-potential (AP) durations, and lower elec-
trical thresholds (Ruch et aI., 1968). Because of the lower thresholds, the
myelinated fiber is a good choice for electrical stimulation studies.
Figure 4.8 illustrates electrical stimulation of a myelinated nerve fiber.
The current emanating from the stimulus electrode through the conducting
medium causes external voltage disturbances (Ve,n) at the nearby nodes.
These disturbances force current across the membrane.
Current
Stimulating ,stimulus
"~~,_-
Etf~~~Y
~ Receptor
Myelin
sheaths Synaptic
or terminals
nerve
terminus 1 in spinal
column
Conducting
medium
FIGURE 4.8. Representation of electrical stimulation of myelinated nerve. The cur-

rent stimulus results in voltage disturbances (Ve) at the individual nodes. These in
turn cause a local depolarization ofthe nerve membrane. (Adapted from Reilly and
Larkin, 1984.)
Figure 4.9 illustrates the representation of the myelinated nerve as for-

mulated by McNeal (1976). The individual nodes are shown as circuit
elements consisting of capacitance (Cm ), resistance (Rm), and a potential
source (Er) , which maintains the transmembrane resting potential. The
voltages Ve,n are the external nodal voltages as in Fig. 4.8. In the McNeal
representation, the myelin internodes are treated as perfect insulators. This
framework could be expanded to include passive myelin properties; how-
ever, such expansion would add significantly to the complexity of the
model. The model is therefore a compromise between relatively simple
single-node models and a more complete node-plus-myelin representation.
Despite the compromise, the model is able to account for a variety of
sensory and electrophysiological effects.
In Fig. 4.9, the current emanating from the nth node is the sum
of capacitive and ionic currents, and is related to internal axonal currents
by
(4.26)
where Cm is the membrane capacitance of the node, Vn is the trans-mem-

brane potential at the nth node, li,n is the internal ionic current flowing in the
nth node, and Vi,n is the internal voltage at the nth node. In this expression,
Vn is taken relative to the resting potential and positive Vn applies to
membrane depolarization. Further relationships are given by
7rd 2
G =-- (4.27)
a 4(}iL
(4.28)
(4.29)
Ve ,n-1
Outside
Inside Ra Ra
FIGURE 4.9. Equivalent circuit models for excitable membranes. The response near
the excitation threshold requires that the membrane conductance be described by a
set of nonlinear differential equations. (After McNeal, 1976.)
where d is the axon diameter at the node, (!j is the resistivity of the axo-
plasm, L is the internodal gap, gm is the subthreshold membrane conduc-
tance per unit area, Cm is the membrane capacitance per unit area, and W is
the nodal gap width.
In Equation (4.26), Vn is the voltage difference across the membrane:
(4.30)
where Vj,n and Ve,n are the internal and external nodal voltages, respectively,
with reference to a distant point within the conducting medium outside the
axon. Substituting Eq. (4.30) into Eq. (4.26) results in
d~n = d m
[Ga(Vn- 1 - 2Vn + Vn+1 + Ve,n-l - 2V.,n + V.,n+l) - Ij,n]
(4.31a)
Equation (4.31a) may be analogously expressed in continuous form as
7: av _ A2 a2 v + V = A2 a2 v. (4.31b)
maT ax 2 ax 2
where V and Ve are membrane voltage relative to the resting potential and
external voltage respectively at longitudinal position x. A form of Eq.
(4.31b), with the constants 7: and A explicitly expressed for a myelinated
fiber, has been presented by Basser and Roth (1991). The main way that Eq.
(4.31b) differs from the cable equation (4.23) is the inclusion of the right-
hand term, in which the external field within the biological medium acts as
a driving force on the membrane voltage.
Equation (4.3lb) can be derived from first principles, or can be obtained
from (4.31a) by substituting em = cmTCdtlx, G a = TCd2/(4{!jtlx), Gm = gmTCdtlx,
where d is the fiber diameter, ~x is the longitudinal increment, {!j is the
axoplasm resistivity (in ncm) internal to the fiber, Cm is capacitance per unit
area, and gm is conductance times unit area. Continuous and discrete spatial
derivatives are connected by iivlax2 = (Vn- 1 - 2Vn + Vn+l)ltlx2; a2v;ax2 =
(Ve,-l - 2Ve,n + V e,n+l)ltlx2; 7:m is the membrane time constant given by cmlgm ;
A is the membrane space constant given by A = (rmlry/2 = (d{!ml(4{!j»1I2, and
{!m is the membrane specific resistance (in ncm2). An additional relationship
is I~n = VIGm •
If one treats A as a constant, then (4.31b) describes the membrane
response only during its sub threshold (linear) phase. For membrane
depolarization approaching the threshold of excitation, membrane conduc-
tance of ionic constituents becomes highly nonlinear, leading to nerve
excitation.
One conclusion that can be drawn from Eqs. (4.31a) and (4.31b) is that a
second spatial derivative of voltage (or equivalently a first derivative of the
electric voltage) must exist along the long axis of an excitable fiber in order
to support excitation. The second spatial derivative of the external voltage
has been included in an "activation function" in order to emphasize this

essential aspect of stimulation (Rattay, 1986, 1989; Plonsey and Barr, 1995).
In a typical analysis, the fiber is considered long and straight. However,
excitation is nevertheless possible where the fiber is terminated or where it
bends within a locally constant electric field (see Sect. 4.5). The orientation
change or termination creates the equivalent of a spatial derivative of the
applied field. Stimulation at "ends and bends" can, in fact, be the dominant
mode of excitation in many cases (see Sect. 4.5).
The ionic current term in Eq. (4.31a) can be expressed for either a linear
or a nonlinear membrane:
(4.32a)
(4.32b)
Equation (4.32a) is a simple statement of Ohm's law for a linear conductor.
Equation (4.32b) applies to the nonlinear ionic current expressions for the
FH membrane given by Eqs. (3.28) to (3.31).
In Eq. (4.31a), the V.,n values are specified from knowledge about the
stimulus current distribution, and the Vn values are unknowns for which
solutions must be found. Procedures for evaluating Eq. (4.31a) typically
consider V•. n as being independent of the membrane currents, that is, the
current crossing the membrane is assumed not to perturb the voltages on
the exterior of the fiber. For subthresholid conditions this is a reasonable
assumption, but in an excited state, the membrane currents can be large
enough to perturb the external voltages. As a consequence, the model will
be less accurate in the excited state. Despite this limitation, the model
accurately reproduces a wide range of excitation phenomena.
The most accurate representation would treat all the nodes in the model
as nonlinear. However, such treatment can result in excessive computing
time, which can be reduced by limiting the number of nonlinear nodes. In
McNeal's original work, he studied an ll-node array with one central node
as nonlinear and all the others as linear. In his study, excitation current was
introduced via an electrode near the central nonlinear node. He defined
excitation as occurring when the nonlinear node reached a peak depolariza-
tion value of 80mV. For the range of stimuli studied by McNeal, this
arrangement was entirely satisfactory. However, for a more general range
of stimulus parameters, some modifications are required.
The model described in this chapter extends the McNeal model by in-
cluding FH nonlinearities at each of several adjacent nodes. Additional
modifications include a test for excitation based on AP propagation,z the
2The use of a single depolarization voltage is not always an adequate indicator of

excitation when brief oscillatory stimuli are used. In that case, a threshold test based
on propagation is needed (Reilly et aI., 1985).
TABLE 4.1. Base-case parameters for SENN

model.
Fiber diameter (D) Variable
Axon diameter (d) O.7D
Nodal gap (G) 2.5mm
Axoplasmic resistivity (e,) 100Qcm
External medium resistivity (ee) 300Qcm
Membrane capacity (cm ) 2#F/cm2
Membrane conductivity (gm) 30.4mS/cm2
Internodal distance (L) lOOD
Source: McNeal,1976.
ability to model arbitrary stimulus waveforms, representation of stimula-

tion at the neuron terminus, and representation of stimulation by uniform
electric fields. This modified representation of McNeal's model has been
referred to as a spatially extended nonlinear nodal (SENN) model (Reilly
et aI., 1985). Unless specifically noted, myelinated nerve model parameters
used to obtain the data in this chapter are as noted in Table 4.1.
To exercise the model, the spatial distribution of voltage along the axon
as a result of the stimulating current must be specified. For example, con-
sider the voltage distribution for isotropic current propagation from a point
electrode placed in a uniform medium. The "indifferent" electrode is taken
to be in the conducting medium, far from the axon. The voltage at a radial
distance, r, from the electrode is given by
V(r) = ~
41U
(4.33)
where e is the resistivity of the medium. In the general case, I and V are
functions of time that follow the stimulus waveshape.
The examples presented in this section are for a point electrode stimulat-
ing a 20-,um-diameter fiber having an internodal spacing of 2 mm; the elec-
trode is assumed to be located one internodal distance (2,um) from the
fiber, and centered above the central node. For fibers smaller than 20,um,
excitation thresholds (at a given electrode distance) would be greater as
noted in Sect. 4.7. Section 4.5 presents additional results for excitation by a
uniform electric field within the conducting medium.
Threshold Criterion
The membrane response of the myelinated nerve model to a rectangular
current stimulus is illustrated in Fig. 4.10. The example is for a small
electrode that is 2mm radially distant from a 20-,um fiber. The transmem-
brane voltage, AV, is scaled relative to the resting potential. The solid
curves show the response at the node nearest the stimulating electrode.
Responses to three different cathodic pulse magnitudes, all of the same

duration (100,us), are depicted. Response a is for a pulse at 80% of
the threshold current, stimulus pulse b is at threshold, and pulse c is 20%
above threshold. The threshold stimulus pulse in this example has an ampli-
tude (IT) of 0.68mA. Figure 4.10 also shows the membrane response to a
threshold pulse at the node nearest the electrode and at the next three
adjacent nodes (broken lines). The time delay from node to node implies a
propagation velocity of 43 m1s.
It is possible for a developing AP to be reversed and entirely abolished by
the phase reversal in a biphasic waveform (Btitikoffer and Lawrence, 1978).
For this reason, a simple membrane voltage test for AP excitation can be
misleading, particularly for stimulus durations less than 100,us. The diffi-
culty is illustrated in Fig. 4.11, which shows the response to a single cycle of
a sinusoidal stimulus having a period of 20,us and an initial cathodic phase.
100
>
E
(1)
Cl
C
'"
.r:
u
(1)
Cl
~
"0
> 50
Q.l
C
'"
'"
.0
E
Q.l Stimulus magnitude
E (a) 0.8 IT
c'"
'"
~
(b) IT
(e) 1.2 IT
(d) - (f) IT
0
0 0.2 0.4 0.6 0.8 1.0

Time (msec)
FIGURE 4.10. Response of myelinated nerve model to rectangular monophasic cur-

rent of 100,us duration, 20-,um-diameter fiber, point electrode 2mm from central
node. Solid lines show response at node nearest electrode for three levels of current.
IT denotes threshold current. Broken lines show propagated response at next three
adjacent nodes for a stimulus at threshold. (From Reilly et aI., 1985.)
100
>
E
CIl
Cl
t:
CtI
.r:.
U
CIl
...
0
Cl
CtI
50
>
CIl
c
...
CtI (a) Stimulus peak = 9.1 mA
.D
E
CIl
E
---
CI)
...
t:
CtI
I- 0 ---------
(b) Stimulus peak = 8.5 mA
Time (msec)
FIGURE 4.11. Nerve model response to one cycle of a sinusoidal pulse with initial
cathodic phase and a period of 20.us. Stimulus (a) is suprathreshold, stimulus (b) is
subthreshold. (From Reilly et aI., 1985.)
Stimulus pulse a, having a peak amplitude of 9.1 rnA, is above threshold: a

propagated AP develops after about 0.3 ms. Stimulus pulse b, having a peak
amplitude of 8.5 rnA, is below threshold: the membrane response decays
rapidly to the cell's resting potential (OmV). Both responses, however,
exceed a depolarization voltage of 80mV, at a time corresponding to the
peak of the cathodic phase of the stimulus.
In the SENN nerve model, excitation is tested on the basis of unambigu-
ous AP propagation. The computer algorithm recognizes this condition by
testing for adequate depolarization (80 mV) that propagates to the third
node beyond the point of initial excitation. The threshold current value
is determined by iterating between a level causing excitation with a level
not causing excitation. The iteration is continued until the threshold and
no-threshold levels differ by no more than 1 %.
Position and Number of Nonlinear Nodes

The most general form of the nerve model would invoke the PH equations
at each of a large number of nodes. However, to reduce computation time,
6
en
.'=c
::::I
...> 4
.......
«I
:c...
~
...c 2
......
CII
::::I
Co)
Iii 0
"0
0
z
-2
0 5 10 15 20
Node position
FIGURE 4.12. Steady-state nodal current distribution for myelinated nerve model.
Stimulating electrode centered over node 11. For cathodic stimulation, positive
nodal current indicates anionic current efflux (cationic flux is in opposite direction).
Yo denotes distance between electrode and axon as a multiple of internodal distance.
(From Reilly et aI., 1985.)
one can minimize the number that are nonlinear, that is, those governed by
the FH equations. The necessary number of linear and nonlinear nodes
depends on the stimulus waveform, the electrode/neuron geometry, and the
extent to which AP propagation is modeled.
In previous work, McNeal (1976) evaluated a cathodic point electrode
placed one-half an internodal distance radially from the axon. Using one
nonlinear node closest to the electrode and five linear nodes on either
side, he found that the excitation threshold could be determined within
a percent or so relative to a much longer axon. However, this arrangement
is not acceptable for anodic or biphasic stimuli, or for more distant
electrodes.
To understand the number of nodes required in the model, consider the
distribution of membrane current resulting from a stimulus. Figure 4.12
shows a spatial pattern of nodal current in response to a cathodic stimulus.
The current distribution plotted here represents a steady-state condition,
that is, the convergent response to a long stimulus pulse. Current efflmc is
represented by the positive current axis, and current influx by the negative
3In this chapter, unless otherwise specified, "current" flow refers to anionic current
according to conventional engineering usage. Cationic current flow is, however, in
the opposite direction.
axis. For an anodic stimulus, the figure would be inverted. An action poten-
tial results only from the depolarizing effect of nodal current efflux, repre-
sented by the positive direction of the vertical axis.
In Fig. 4.12, maximum current efflux with cathodic stimulation occurs at
the node nearest the electrode. This point of maximum response is called
the excitation node, that is, the node where an AP would be initiated with
a threshold-level stimulus. By inverting Fig. 4.12, it can be seen that there
are two potential excitation nodes for an anodic stimulus; these move to
more distant locations along the axon as the electrode is positioned farther
away. In addition, the nodal current distribution becomes more gradual. In
modeling anodic stimulation, it is necessary to ensure that nonlinear nodes
are included at the excitation location.
Care must also be exercised in the selection of the total number of nodes.
The number needed for modeling anodic stimulation is greater than with
cathodic stimulation. If there are not enough nodes, the nodal current
distribution can be altered relative to a longer axon, and significant errors
can be induced. These distortions can be illustrated by calculating the
polarity sensitivity ratio (P), defined as the ratio of threshold current for
anodic to that for cathodic stimulation (absolute values). Figure 4.13 illus-
trates P as a function of the electrode placement Xo measured longitudinally
from the terminus of the node array, for various radial distances Yo from the
axon. The array is assumed to be terminated at a node. The parameters
Xo and Yo are expressed as multiples of the internodal spacing. With this
normalization, P is only weakly dependent on the ratio of nodal to axonal
resistivity (Rm and Ra in Fig. 4.9), and is otherwise independent of param-
eters related to fiber size. The pulsewidth is assumed to be long; this incurs
no loss of generality.
The value of Xo at which P oscillates about fixed limits can be interpreted
as the array length at which the axon model appears to be infinitely long.
This point occurs when Xo is 9, 13, and 37 nodes from the axon terminus
for values of Yo equal to 2, 4, and 8, respectively. Because these distances
represent one half the necessary extent of the axon model, we form the rule
of thumb that for general applications, the model should be at least nine
times longer than the radial distance to a longitudinally centered electrode.
For the computations reported in this chapter, the array length is at least 10
times Yo but not less than 21 nodes. To study AP propagation or related
phenomena (such as AP collision effects), the model requires nonlinear
nodes at least as far as propagation is to be simulated.
Suprathreshold Response
An assumption in the myelinated nerve model is that the voltages external
to the nodes are defined solely by the stimulus current. However, once the
nerve is excited, these external voltages will be perturbed by the nodal
current from the propagating AP, and these perturbations are not included
,., ,.
./
".
".
Q. 8 ./
.Q ./
,-
T§ ./
>. ./
:t::
.::: ./
./
.~ ./
c: ./
Q)
\ ./
,-'"
rJj
>.
rn
."!::
aQ.
Lateral distance from end of axon, Xo (internodal units)
FIGURE 4.13. Ratio of anodic to cathodic threshold current when stimulating elec-
trode is near one terminus of a 91-node axon model. Yo denotes distance between
electrode and axon as a multiple of internodal distance. Computations were done at
multiples of half the internodal distance, resulting in a sawtooth pattern. (From
Reilly et aI., 1985.)
in the model. Nevertheless, these perturbation potentials will fall to

unexited levels at those nodes where the propagating AP has not yet ar-
rived, or has already passed. Consequently, many aspects of suprathreshold
behavior can be adequately addressed with the model, such as repetitive or
sinusoidal stimulus effects, which will be discussed below.
4.5 Response to Monophasic Stimulation

Strength-Duration Relations
Figure 4.14 illustrates thresholds evaluated with the myelinated nerve
model for anodic and cathodic rectangular stimuli with pulse durations
of from l.us to 10ms. Figure 4.14 also shows the S-D curve for a cathodic
exponential stimulus having the form Ie- tlr • The curves in Fig. 4.14 are
similar to those for the simplified linear model, described by Eqs. (4.6) and
Response to Monophasic Stimulation 127
(4.9) for rectangular stimuli, and by Eqs. (4.13) and (4.14) for exponential
stimuli. The threshold charge reaches a minimum at small stimulus dura-
tions for both exponential and rectangular stimuli; for long-duration
stimuli, the peak current (equivalent to rheobasic current) is minimized.
The exponential and rectangular stimuli have the same minimum charge
threshold, indicating that, at brief durations, sensitivity is not affected by
the fine structure of the monophasic pulse waveform.
The shape of the S-D curve for a linear RC network model of a single
node can be characterized by the membrane's exponential time constant,
given by the product RmCm as noted in Sect. 4.2. The myelinated nerve
model also has a response time that depends on both linear and nonlinear
membrane properties and on the spatial distribution of the electric field
external to the axon. Consequently, no single linear circuit parameter speci-
fies the curves in Fig. 4.14. It is therefore useful to define an equivalent
strength-duration time constant ('t.) as the RC time constant of the linear
single-node model having an S-D curve shape that best matches the shape
of a given empirical curve. The value 't. = 92.3#s is obtained using a linear
least-squares fit of Eq. (4.6) to the rectangular current threshold curve in
10 , 100
,
/
\ - - - Cathodal rectangular /
---- Anodal rectangular /
\
, ,.........
Cathodal exponential /
/
.. ,, /
,
/
(}
.. «
.=: ,,
/
/
.. E
CI>
Cl
~
.. , /
/ 10
...c
<tI CI>
..c. ··...Current' "- / tharge./
...
~
tJ /
~
"- :::l
"0
"0 .. ... ..- ..-
....
>( tJ
..c. ;... ...... ::-.:.... "0

"0
III
CI>
~ - ---- ...... ..c.
III
..c. CI>
I- 0.1
~
..c.
I-
...... ......
0.01 L...--'---~L....L..J_.l...-....L.-""""'_.l...-....L.-L...I....L_~"""""'~ 0.1

1 10 102 103 104
Stimulus duration (Ilsec)
FIGURE 4.14. Myelinated nerve model strength-duration curves for monophasic

stimuli. The left vertical axis indicates threshold charge for AP initiation. The right
vertical axis indicates threshold current. The horizontal axis represents pulse dura-
tion for a rectangular stimulus or decay time constant for an exponential stimulus.
(From Reilly et aI., 1985.)
Fig. 4.14. This value differs from the simple product of membrane capaci-
tance and resistance in the subthreshold region of linear response, for which
the product RmCm is only 66f-ls. As suggested by Eq. (4.20), we can more
simply estimate 'te as the ratio Qoflo, which results in 'te = 92f-ls. This estimate
agrees quite well with the value obtained by a least-squares fit of the entire
S-D curve.
The value of 'te cited above applies to a point electrode whose radial
distance from the axon is equal to an internodal space. As noted in Sect. 4.3,
'te is expected to increase as the membrane current is distributed more
gradually along the axon. We would therefore expect 'te to be larger at
increased electrode distances. This expectation is verified with the model
results reported in Sect. 4.7. When the distance to a point electrode is varied
from 1/2 to 4 internodal spaces, the observed value of the 'te ranges from 56
to 128f-ls.
Values of 'te obtained with the myelinated model agree quite well with in-
vivo experimentation of animal nerve (Reilly et aI., 1985), and is within the
range of experimental time constants reported for electrocutaneous nerve
stimulation (see Table 7.1). Nevertheless, experimental time constants vary
over a large range, many being more than twice the values predicted with
the nerve model. It is hypothesized that electrocutaneous stimulation may
originate at neural end structures (receptors, free nerve endings, and motor
neuron end plates). We will further consider models for neural end
structures.
Polarity Sensitivity
The vertical separation between anodic and cathodic stimulation in Fig.
4.14 reflects the polarity sensitivity ratio (P) introduced in Fig. 4.13. In
Fig. 4.14, P ranges from 4.2 at a pulse duration of If-ls to 5.6 at a pulse dura-
tion of 10ms. As noted in Chapter 7, polarity ratios for electrocutaneous
stimulation are closer to 1.3.
It is difficult to reconcile P values from sensory experiments with those
obtained with the myelinated nerve model as long as the stimulating elec-
trode is positioned far from the end of the model axon. As seen in Fig. 4.13,
the minimum value of P is about 4.5 for a variety of electrode/neuron
geometries when the electrode is distant from an axon terminus. But as the
electrode moves closer to the terminus, the model predicts smaller polarity
ratios, and even predicts ratios less than unity in some cases.
Ranck (1975) discusses a qualitative model for stimulation near the ter-
minus of a CNS neuron and shows that it is possible for anodic excitation
to occur at lower stimulus levels than with cathodic excitation. I have not
encountered polarity ratios less than 1 in electrocutaneous experiments.
Nevertheless, Ranck's analysis and Fig. 4.13 suggest that the modest polar-
ity ratios found in electrocutaneous experiments may occur because the
principal sites of stimulation are near terminal structures (receptors, free
nerve endings). Figure 4.13 shows that P ranges from about 4.5 to less than
unity, depending on the electrode/ neuron configuration. With cutaneous
stimulation, there will be a mix of orientations of excitable structures in the
skin beneath the electrode with a spectrum of polarity ratios. The modest
ratios observed in sensory experiments may represent a macroscopic
average of this spectrum. A further analysis of this issue is presented in
Sect. 4.7.
Current Density and Electric Field Relationships

Recall from cable theory (Sect. 4.3) that a longitudinal electric field is
necessary to support excitation. The electric field must also have a spatial
gradient, as can be appreciated with reference to Eq. (4.31), which shows
that second differences4 of the external voltages are the driving forces for
changes in membrane potential. If the electric field were uniform and the
axon were infinitely long in both directions, there would in theory be zero
net current transfer at every node. However, the field within the biological
medium is never uniform. Furthermore, an effective field gradient will be
realized if the orientation of the axon changes with respect to a locally
uniform field, or if the axon is terminated in the field (as with receptors, free
nerve endings, or nerve connections at muscle fibers).
Figure 4.15 illustrates three modes whereby a nerve fiber may be excited
by an external electric field. We designate these as end, bend, and spatial
gradient modes. End and bend modes of stimulation have been analyzed by
several researchers (Reilly and Bauer, 1987; Coburn, 1989; Nagarajan et al.,
1993; Rubinstein, 1993; Struijk et al., 1993; Abdeen and Stuchly, 1994).
The myelinated nerve model can be used to analyze these excitation
modes. For a straight fiber terminated at a node and oriented along a
uniform electric field of magnitude E, the nodal voltages are
V.,n = V.,l + ELn (4.34)
where V.,l is a reference voltage at the first node, L is the internodal
separation, and n is the node number. The membrane response is indepen-
dent of V.,l, which may be taken as zero for convenience. Equation (4.34)
applies if the first node faces the cathode of the current source. Table 4.2
lists monophasic excitation thresholds for unbent fibers of 5, 10, and
20.um-effectively covering the diameter spectrum of myelinated fibers.
The thresholds are inversely proportional to fiber diameter for this mode of
stimulation. For long stimulus pulses (;:::lms), the threshold converges to a
minimum E-field value; for short duration pulses (t ::5 5.us), the threshold
converges to a minimum value of the product of Et (field times duration).
These convergent measures, listed in Part A of Table 4.2, are analogous to
4A second difference is calculated as (Vn- 1 - V n) - (Vn - V n+1) = V n- 1 - 2Vn + Vn+h

where Vn is the voltage at the nth node.
the rheobase current and charge noted in Fig. 4.14 for a discrete current
stimulus. Although the in-situ electric field is the primary force governing
stimulation, current density is perhaps a more frequently cited stimulus
parameter. In Part B of the table, the thresholds are given as current density
values, assuming a bulk conductivity a = 0.2 S/m. For the simulations repre-
sented in Table 4.2, the S-D time constant was found to be 'te = 120.us, as
determined by a linear least square fit of the threshold data to Eq. 4.6. The
thresholds listed in Table 4.2 bracket experimentally determined values if
one properly accounts for waveform and geometric factors (see Table 9.7
and related discussion). In all cases, excitation was initiated at the terminus
facing the cathode of the electric field source.
We also determined minimum thresholds of excitation for a bent fiber.
For a 900 bend at node N, we applied a constant voltage to the first N nodes,
and incremented the voltage as in Eq. (4.34) at each successive node above
N. For a 1800 bend at node N, we first decremented the voltage at successive
nodes up to N, and then incremented the voltage for nodes above N. By this
method, it was possible to simulate a sharply bent fiber within a uniform E-
field. Table 4.3 lists threshold requirements associated with end and bend
Bend
Current Depolarization
---,.
~
"~, .-..-..'..'.-.:.~'....:::-
" ,ff....
::;~::::
Spatial
gradient
"---"V~-~~ '---"v,,-~-
Depolarization Hyperpolarization
FIGURE 4.15. Modes of neural stimulation. Excitation is initiated at points of maxi-

mal current efflux across neural membrane. Potential excitation sites consist of fiber
terminals, sharp bends, and maximal spatial gradient of E-field.
TABLE 4.2. Minimum stimulus thresholds with uniform field excitation;

single monophasic stimuli.
Fiber diameter (urn)
5 10 20
A. Field strength criteria
1. E min (Vim) 24.6 12.3 6.2
2. (Et)min (Vs/m) 2.98 X 10- 3 1.49 X 10- 3 0.75 X 10-3
B. Current density criteria
1. 1min (Nm 2 ) 4.92 2.46 1.23
2. qmin (Clm 2 ) 6.0 X 10-4 3.0 X 10-4 1.5 X 10- 4
Thresholds A.1 and B.1 apply to long pulses (t 2: 1 ms). Thresholds A.2 and B.2
apply to short pulses (t ~ 5#s). Current and charge density determined for
conductivity a = 0.2 S/m.
Source: From Reilly (1988).
modes of stimulation. The first column indicates the bend angle. The second
column lists the node at which the bend occurs. The first listed case with a
bend angle of 00 applies to a terminated fiber, which is the standard case
used to evaluate MRI excitation thresholds (see Sect. 9.7). The third col-
umn lists the rheobase excitation threshold. The last column lists the
strength duration time constant, as determined by the ratio (Et)minIEmin
obtained with short- and long-duration stimuli.
For the cases listed in Table 4.3, the lowest thresholds apply to a fiber
with a sharp 1800 bend at a location distant from the terminus. However,
such a condition would not be realistic for practical fiber trajectories. If we
examine the remaining cases, we see that the straight, terminated fiber
provides the lowest practical threshold. For the bent fiber cases, excitation
was initiated at the bend node. Note that gradual bends would necessitate
TABLE 4.3. Excitation requirements for end and

bend modes of stimulation.
Bend angle Bend node E threshold I,
(deg.) (#) (Vim) (#s)
0 1 6.21 128.2
90 2 8.55 126.0
90 4 9.84 114.3
90 6 9.96 112.9
90 8 9.96 112.9
180 2 6.56 101.4
180 4 5.45 105.0
180 6 5.10 110.3
180 8 5.04 111.6
Thresholds apply to 20-#m nerve fiber within constant

E-field that is oriented parallel to the nerve beyond the
bend point.
higher excitation thresholds, since the second derivative of voltage would

necessarily be lower as compared with a sharp bend.
As will be apparent in succeeding chapters, the predicted rheobase for a
20-.um fiber in the terminated mode closely corresponds to thresholds ob-
tained in many experiments with humans and animals. For instance, Havel
and colleagues (1997) reported perception thresholds with magnetic stimu-
lation through a coil encircling the arm of human subjects. At the threshold
of perception, the rheobase E-field induced in the periphery of the arm was
determined to be 5.9V/m-which is quite close to the theoretical value of
6.2V/m for a 20-.um myelinated fiber.
4.6 Response to Biphasic and Repetitive Stimuli

Strength-Duration Relationships
The current reversal of a biphasic pulse can reverse a developing AP that
was excited by the initial phase. As a result, a biphasic pulse may have a
higher threshold than a monophasic pulse. Figure 4.16 shows S-D curves
from the myelinated nerve model for three types of stimuli: a monophasic
constant current (rectangular) stimulus, a symmetric biphasic rectangular
stimulus, and a sinusoidal stimulus. The data apply to stimulation via a
point electrode 2mm radially distant from a 20-.um-diameter fiber. Stimuli
consist of a single biphasic current with an initial cathodic phase followed by
an anodic phase of the same duration and equal magnitude. The phase
duration indicated by the horizontal axis is that for the initial cathodic half-
cycle. Stimulus magnitude is given in terms of peak current on the right
vertical axis, and in terms of the charge in a single monophasic phase of the
stimulus on the left vertical axis. The charge is computed by Q = Itp for the
rectangular waveforms and from Q = (2ht) Itp or the sinusoidal waveforms
(I is threshold current amplitude and tp is phase duration).
The current reversal in the biphasic waveforms increases the threshold
for a propagating AP. This situation can be seen in Fig. 4.16 by comparing
the monophasic and biphasic rectangular stimulus thresholds. For long
durations, the threshold current is the same for the two stimuli. However, as
the stimulus duration becomes short relative to the S-D time constant
(about l00.us), the biphasic current has an elevated threshold. The degree
of elevation is magnified as the stimulus duration as reduced.
The thresholds illustrated in Fig. 4.16 apply to stimuli with an initial
cathodic phase. Thresholds are greater if the initial phase is anodic, but only
if the phase duration is less than lOO.us. For a single cycle of a sine wave, the
model shows that if the initial phase of the stimulus is anodic, thresholds are
greater than initial cathodic thresholds by 5% at a phase duration of lOO.us,
10% at 50.us, 45% at lO.us, and 60% at 5.us.
Figure 4.17 illustrates threshold multipliers (M) based on the myelinated
nerve model for biphasic rectangular pulses with uniform field excitation
Response to Biphasic and Repetitive Stimuli 133
+--+ Monophasic cathodal ...r1...-

G---£) Biphasic rectangular ~
tp ..::I¥'./:
«
10 ..§.
................* ..•.. .,"*

--"-"::=-=:...=--~-41
0.01 '--...L-....1-L..L..L----'L-.L.-J...-'-'---'----L-L..JL..L..---L...--'-....L..L..L-....1---L....J....L.J O. 1
1 10 10 2 10 3 10 4 10 5
Stimulus phase duration, tp (~sec)
FIGURE 4.16. Strength-duration relationships derived from the myelinated nerve

model: current thresholds and charge thresholds for single-pulse monophasic and
for single-cycle biphasic stimuli with initial cathodic phase, point electrode 2 mm
distant from 20,um fiber. Threshold current refers to the peak of the stimulus
waveform. Charge refers to a single phase for biphasic stimuli. (From Reilly et aI.,
1985.)
(Reilly, 1988). The vertical axis gives the threshold multiplier for a double
pulse relative to a single pulse. The portion of the figure above M = 1
applies to a biphasic pulse doublet, where the current reversal has the same
magnitude and duration as the initial pulse. The portion of Fig. 4.16 below
M = 1 is for a monophasic pulse doublet. Figure 4.17 applies if the initial
pulse is cathodic. Stimulation is also possible with an initial anodic pulse,
but the thresholds are elevated.
According to Figure 4.17, biphasic thresholds are elevated by an amount
that depends on the pulse duration and the time delay before current
reversal. Thresholds are most elevated when the pulse is short and the
current reversal immediately follows the initial pulse. If the phase reversal
is delayed by lOO,us or more, there is little detectable effect on the thresh-
old. An implication of the results shown in Figs. 4.16 and 4.17 is that the
membrane integrates the stimulus over a duration roughly equal to the S-D
5
' "I " " I
4
-jlpr-
S}u-
2:
..:
.!!!
Jtil-
0.
E
:l
E
"C Biphasic doublet
"0
.r:
en
~
.r:
I-
200
100
50
20
0.5 Tp = 2.5 MS ~
Monophasic doublet
I , I , Ii' I , I , , , ,I I ,,
0 5 10 50 100 200
ti
00
Time delay between pulses, (MS)
FIGURE 4.17. Threshold multipliers for biphasic and monophasic pulse doublets:
Uniform field excitation of terminated axon. (From Reilly, 1988.)
time constant. This integration is nonlinear: the biphasic waveforms all

inject zero net charge, but have finite threshold magnitudes. Various fea-
tures observed with biphasic stimulation of the nerve model are also seen
experimentally in cardiac (Chapter 6) and nerve excitation (Chapters 7 and
8), as long as appropriate adjustments are made for experimental strength-
duration time constants.
Sinusoidal Stimuli
Figure 4.16 illustrates excitation thresholds for sinusoidal stimuli consisting
of a single cycle. The excitation threshold also depends on the number of
stimulus cycles. Figure 4.18 illustrates the response of the myelinated nerve
model to threshold-level sinusoidal stimuli at a frequency of 5 kHz (point
electrode 2mm distant from 20-,um fiber). The stimulus is composed of
a single cycle in example a, two cycles in b, and three cycles in c. The
threshold current requirement is reduced according to the data above the
figure as the number of stimulus cycles is varied from one to three.
Figure 4.19 illustrates the relationship between threshold and stimulus
duration for a sinusoidal current with frequency from 25 to 400 kHz; the
initial phase was cathodic in these examples. When evaluated at half-cycle

multiples, there is an oscillating threshold with minima at odd numbers of
half-cycles and maxima at even multiples. The broken lines in Fig. 4.19
show the duration of the stimuli. For the cases displayed, thresholds at
stimulus durations beyond 1.28ms converge to a minimum plateau that is
within a few percent of the threshold that would apply to a cathodic
monophasic square-wave pulse having a duration equal to the phase dura-
tion (one-half period) of the sinewave stimulus. This observation is an
emphirical one. As noted in Sect. 9.7, threshold behavior similar to that
in Fig. 4.19 has been experimentally observed in human subjects using
variable length sinusoidal magnetic stimulation (Budinger et al., 1991).
The thresholds in Fig. 4.19 oscillate because a sinusoidal stimulus that has
an even number of half-cycles is charge balanced (zero net charge), and one
with an odd number of half-cycles will transfer the maximum net charge.
Threshold current: (a) 0.98 rnA

(b) 0.90 rnA
(c) 0.87 rnA
Frequency =5 kHz
100
>
E
C1>
Cl
C
(1)
.r::
u
C1>
...
Cl
(1)
o> 50
C1>
C
...
(1)
.D
E
C1>
E
'"c
...
(1)
I- o
-20~~--~--~--~~~~--~--~--~~
o 0.2 0.4 0.6 0.8 1.0
Time (msec)
FIGURE 4.18. Response of myelinated nerve model to sinusoidal threshold-level

stimuli at a frequency of 5 kHz. Stimulus durations: (a) one cycle, (b) two cycles, and
(c) three cycles. Peak values of threshold currents are listed above figure. (From
Reilly et al., 1985.)
400
~
.s 100
c:
~::J
0
"0
,, ,
(5
..c:
en
~
£
-"
III
(J)
10
a..
1
.1 10 100 1000
Number cycles of stimulation
FIGURE 4.19. Excitation thresholds as a function of number of cycles of sinusoidal

stimulation. Stimulus duration stepped in half-cycle increments out to four cycles,
and full cycle increments beyond that point. Dashed lines indicate duration of
stimulus. Point electrode 2mm distant from 20.um fiber.
The gradually falling aspect of the threshold is a consequence of the non-

linear conductance of the membrane. Recall from Chapter 3 that as the
membrane depolarization approaches the excited state, ionic conductance
for inward- and outward-going current differs. As a consequence, an
oscillating stimulus will build up a bias voltage on the membrane that
increases with each successive cycle of stimulation, as experimentally
verified with biphasic stimulation of both nerve and cardiac tissue (see
Chapters 6-8).
Single monophasic pulses produce a single AP. An oscillating stimulus
has the ability to produce a train of APs, which can greatly enhance the
intensity of the electrical response in nerve (Chapter 7), cardiac (Chapter
6), and skeletal muscle (Chapter 8) responses. Repetitive responses can be
predicted from the nonlinear models described in Chapter 3 for both myeli-
nated and unmyelinated nerves.
Figure 4.20 illustrates the response of the myelinated nerve model to
continuous sinusoidal stimulation. The axon response is shown two nodes
distant from the excitation node in order to verify AP propagation. The
stimulus levels indicated are multiples of the single-cycle threshold. The AP
repetition rates are 100, 250, and 500 Hz for threshold multiples of 1.0, 1.2,
and 1.5, respectively. In each case, APs are synchronized with the stimulus.
Similar phase locking has been observed in the neural responses of sensory
systems to periodic stimulation (Kiang, 1965). Responses were also studied
with the myelinated nerve model for continuous sinusoidal stimulation at
5 kHz. The AP repetition rates were 320, 470, and 540 Hz at threshold
multiples of 1.0,1.2, and 1.5, respectively. At the 1.5 stimulus multiple, there
was a 4.4-ms refractory period after the first AP during which excitation did
not occur. After this initial refractory period, a steady AP rate of 540 Hz was
produced. At both frequencies there is a reduction in magnitude of the
AP spike as the AP repetition rate increases. The maximum AP rate and
100
> 50
E
Ql
Cl
c:
a
co
.r::.
u
Ql
Cl
100
....co
"0 50
>
Ql
c:
~
..0
a
E
Ql
E 100
'"c
...
co
I-
50
a
a 5 10 15 20
Time (msec)
FIGURE 4.20. Model response to continuous sinusoidal stimulation at 500 Hz. The
lower panel depicts the response to a stimulus current set at threshold level (IT) for
a single-cycle stimulus. Upper panels show response for stimulation 20% and 50%
above the single-cycle threshold. (From Reilly et ai., 1985.)
reduced depolarization voltage are consistent with the experimental

A-fiber response illustrated in Fig. 3.11.
Sinusoidal threshold response can be represented by strength-frequency
(S-F) curves as shown in Fig. 4.21. The horizontal axis in Fig. 4.21 is the
inverse of twice the phase duration in Fig. 4.16. Figure 4.21 also shows a
threshold curve for continuous sinusoidal stimulation.
Figure 4.21 includes experimental threshold curves for human perception
and muscle contraction (Dalziel, 1972; Anderson and Munson, 1951). The
experimental curves have been arbitrarily scaled on the vertical axis to
facilitate comparison of the curve shapes. The shapes of the experimental
data and continuous stimulation model results correspond reasonably well
considering that continuous simulation was used in the cited experimental
studies. Sinusoid thresholds rise below 40 Hz-at low frequencies, the slow
/
/
Perception /
(Dalziel, 1972) ~ /
/
Perception /
10 (Anderson & /
Munson, 1951) ~/
-~ Muscle
tetanus
II
II /
,'"
(Dalziel, 1972) ............... / : / ..... /
/~/
"
------- - - -""
/"
1 .~-
0.1
10 1
Frequency (Hz)
FIGURE 4.21. Strength-frequency curves for sinusoidal current stimuli. Dashed
curves are from experimental data. Solid curves apply to myelinated nerve model.
Experimental curves have been shifted vertically to facilitate comparisons. (From
Reilly, 1988.)
rate of change of the sinusoid prevents the membrane from building up a

depolarizing voltage because membrane depolarization is counteracted by
membrane leakage. Square-wave biphasic stimuli, in contrast, do not have
a rate of change that depends on frequency. Consequently, there is no
increase in thresholds at low square-wave frequencies.
To a first approximation, a functional S-F relationship might be obtained
from a monophasic S-D curve by considering that a sinusoidal half-cycle
corresponds to a single monophasic pulse. This simplification would, how-
ever, fail to account for the increased sinusoidal thresholds at both high and
low frequencies relative to monophasic stimulation. An empirical fit to S-F
data from the myelinated nerve model is given by
(4.35)
10
where It is the threshold current, is the minimum threshold current, and
f.andfoare constants that determine the points of upturn in the S-F curve at
high and low frequencies, respectively. An upper limit on the low frequency
term (second bracket) of K DC is assumed in Eq. (4.35) to account for the fact
that excitation may be obtained with finite direct currents. Equation (4.35)
has the asymptotic form
It = Io(J. J for f » fe (4.36)
r
and
It = 10(; forf« to (4.37)
Sinusoidal thresholds were obtained with the myelinated nerve model for
stimulation by a point electrode positioned 2mm radially distant from a
20-llm-diameter fiber. An empirical fit of Eq. (4.35) to the model thresholds
indicates b = 0.8, and that below 80kHz, a = 1.45 for single cycle stimula-
tion, and a = 0.9 for continuous stimulation. From 80 to 400kHz, a = 1.7 for
single cycle, and a = 1.0 for continuous stimulation. Frequency constants
for the nerve model were f. = 5,400Hz and fo = 10Hz. Experimentally
derived values off. and fo encompass wide range, as noted in Chapters 6 and
7. fo in the range 10 to 50Hz is observed for nerve stimulation, and around
10Hz for cardiac stimulation. Experimental geometric mean values off. are
about 500Hz for nerve excitation (Chap. 7), and about 115Hz for cardiac
excitation (Chap. 6). Differences in model an experimental values reflect
the dynamic membrane responses of the different types of tissue being
stimulated, as well as the distribution of stimulating current in an experi-
mental situation.
Equation (4.35) has features in common with an S-F expression proposed

over sixty ago by Hill and colleagues (1937), which can be equivalently
stated as
(4.38)
The expression of Hill et al. involves upper and lower transition

frequencies, as does Eq. (4.35), and has the asymptotic forms of Eqs.
(4.36) and (4.37) with a = b = 1. The lower transition frequency was
described by Hill et al. as arising from the neural property known as
accommodation, that is, the adaptation of a nerve to a slowly varying or
constant stimulus.
The frequency limits for which Eq (4.35) is valid are not known. While
human sensory experiments with sinusoidal currents closely agree with the
form of Eq. (4.35) up to at least 100kHz (see Sect. 7.5), one could infer
experimental verification up to perhaps lOMHz based on sensory thresh-
olds with pulsed stimuli-as noted in Chap. 7, the strength-duration law
applies to human sensory data for capacitor discharges as brief as 0.1.us.
Experiments with rats show reasonable correspondence up to 1 MHz,
as seen in Fig. 4.22 (LaCourse et aI., 1985). The figure plots current thre-
sholds (peak-to-peak) for excitation of rat's tibialis caudalis muscle,
gastrocnemium muscle, or the innervating nerve of the gastrocnemius.
Electrodes were wire loops placed directly on the muscle or nerve; excita-
tion was determined by contraction of the affected muscle. The curve plots
means for 7 to 10 subjects. Also plotted on the figure is a curve in which
thresholds rise directly with frequency, as would be expected from the
theoretical SENN model described previously. The curves taken together
demonstrate a reasonable correspondence with the theoretical model. De-
viations from the curve might be due to experimental difficulties, including
the fact that the stimulus wire produced significant heating-the experi-
ments were taken to the limits beyond which tissue destruction would have
occurred as a result of thermal damage. The authors could have tested
electrical stimulation without excessive tissue heating by limiting the dura-
tion of the sinusoidal current to a few milliseconds, although this apparently
was not done.
Electrical stimulation of nerve has been observed during electrosurgery
using devices that operate at frequencies in the vicinity of 500 kHz and
above (LaCourse et aI., 1985). For such high frequencies, one might expect
that it would require unrealistically high current density based on the
strength-frequency relationship for nerve excitation. A more likely expla-
nation is that excitation during electrosurgery is due to a nonlinear sparking
process at the cutting site. This process effectively rectifies the high fre-
quency current, producing DC or low frequency current in the tissue that is
effective for nerve excitation (Tucker et aI., 1984; LaCourse et aI., 1988,
Slager et aI., 1993). Others have suggested that nonlinear impedance of
1000
--+- Gastrocnemius muscle

100 ······0····· Tibialis caudalis muscle
---0--· Gastrocnemius nerve
-c
~
.....
Theoretical slope
:J
C)
"0
(5
..c
C/)
10
~
:5
.::,t:.
ct!
Q)
Co
I
.::,t:.
ct!
Q)
a..
.1
.1 10 100 1000
Frequency (kHz)
FIGURE 4.22. Excitation thresholds with sinusoidal current, 100 Hz-1 MHz. Stimu-
lation via wire electrode contacting muscle or nerve. (Data from LaCourse et ai.,
1985).
tissue might provide a demodulation of low-frequency amplitude fluctua-

tions on the otherwise high-frequency stimulus (Geddes et aI., 1975).
Repetitive Stimuli
Repetitive stimuli can be more potent than a single stimulus through
threshold reduction or through response enhancement due to multiple AP
generation. In both cases an integration effect of the multiple pulses occurs.
In the first case, the integration takes place at the membrane level. In the
second case, response enhancement takes place at higher levels within the
central nervous system for neurosensory effects, and at the muscle level for
neuromuscular effects. The following considers membrane effects using
the myelinated nerve model. Other integration effects are described in
Chapters 7 and 8.
......
.j::>.
N
L500lls 500lls
",uu IlS
f>-
200lls 200lls 100 s tI1
~
100 Il.S ra·
I>l
50lls g.
100 Ilsi \\ "\. ::t
~ ~
,.: 50 IlS 0
.!!! 10 IlS Q.
g.
0. 0.6
:E 1 \\ \.. 1 I ~ '"
::l
50 IlS
E
:!:! 10 IlS
o
~ 0.4
Q) TP-tlt- 2 3 ... N
... 0=0
.s::.
r- n n I l . ... n.
~ol-
10 IlS
0.2
0=0
(a) Tp = 10 IlS (b) Tp = 50 IlS (c) Tp = 100 IlS
-
O~I----~----~----~----~----L---~
1 2 4 8 16 32 64 2 4 8 16 32 2 4 8 16 32
Number of pulses, Np Number of pulses, Np Number of pulses, Np
FIGURE 4.23. Threshold multiplier for repetitive pulse sequences-thresholds evaluated at Np = 1,2,4,8, 16, 32, and 64. Uniform field
excitation. (From Reilly, 1988).
The lower section of Fig. 4.17 (M < 1) illustrates multiple-pulse threshold

effects for two pulses of the same polarity. The figure gives the threshold
multiplier when there are two pulses of stimulation relative to that for a
single pulse. The effects are most pronounced for short pulses and short
interpulse delays. (A delay of zero is the same as a single pulse of twice the
duration.) The nerve model was also exercised to evaluate the threshold
modification for sequences of pulses, N p , numbering 1, 2, 4, 8, 16, 32, 64,
and 128. Figure 4.23 illustrates the results for tp = 10,50, and lO0.us and for
<5 = 10,50, 100, 200, and 500.us. As in Fig. 4.17 the vertical axis gives the
threshold multiplier relative to a single pulse. The curve labeled <5d =
applies to continuous stimulation over a time period corresponding to Np"
°
Pulse integration reduces thresholds increasingly so as the pulse duration
and delay time are shortened. At a delay of 500.us, the nerve model shows
no measurable pulse integration effect.
Nonlinear Interaction of Multiple Waveforms

A linear system is one in which its response is proportional to the magni-
tude of an input stimulus. Clearly, the neural membrane is highly nonlinear
when depolarized near its excitation threshold, where a slight change in
polarization can trigger a propagating action potential. Even at potentials
well below the action potential threshold where the membrane is only
slightly nonlinear, one can observe nonlinear electrical response. For in-
stance with an oscillatory stimulus, the membrane potential achieves a bias
potential that asymptotically increases with each cycle of stimulation due to
the differing membrane conductivity to inward and outward current. Non-
linear membrane properties are responsible for the variation of membrane
potential and excitation thresholds with respect to the duration of a sinusoi-
dal stimulus as seen in Figs. 4.18, 4.19, and 4.21.
As with all nonlinear systems, the response of the neural membrane to
multiple stimuli will exhibit properties not observable in the response of the
individual waveforms. That is why it is invalid to evaluate neural response
to a multispectral stimulus waveform in terms of its response to the indi-
vidual Fourier components of the waveform. While such "superposition"
techniques are valid for linear systems, they are incapable of accounting for
the response of a nonlinear system.
Nonlinear response of the neural membrane to multiple waveforms can
best be studied with a nonlinear electrodynamic model, such as the SENN
model described previously. As an example of such behavior, Figure 4.24
illustrates the threshold response to a dual stimulus consisting of a 0.2ms
square-wave "conditioning pulse" (CP), and a sinusoidal function. The
simulation applies to a terminated 10-.um myelineated nerve in an E-field
that is aligned with the fiber. The ordinate gives the peak threshold value of
a sinusoidal E-field having a duration of 1 ms, and a frequency indicated by
the abscissa. The CP is applied at a subthreshold level either preceding the
300
200
150
f1mUUUI-
~1.oms-.1
0.2
E T=18.75 VIm for square-wave

conditioning pulse
~ 100
"C
"6 90
..c
In
80
~ 70
-="ijj 60
"C
·0 50
In
:::I
r::: 40
en
30
20
5 10 20 40 80 100
Sinusoidal frequency (kHz)
300
200
150
E
~ 100
"C 90
"6 80
..c
In
~
70
-="ijj 60
"C
·0 50
In
:::I 40
r:::
en
30
20
5 10 20 40 80 100
Sinusoidal frequency (kHz)
FIGURE 4.24. Excitation thresholds with dual-function stimulation of lO-.um fiber.

CP = conditioning pulse at various fractions of excitation thresholds. Upper panel:
sequential stimuli; lower panel: concurrent stimuli.
Parameter Variation Effects 145
sinusoid (upper panel), or conjointly with it (lower panel). The parameter

values indicate the strength of the CP as a fraction of its threshold value of
18.75 Vim if presented alone. For instance, CP = 0.5 means that the condi-
tioning pulse attains an E-field of 9.37V/m. The curve CP = 0 indicates
the strength-frequency response of the neuron with a pure sinusoidal
stimulus.
It can be seen that the CP can drastically lower the excitation threshold
of the sinusoidal stimulus, especially if the two functions occur concur-
rently. For instance at frequencies above 10kHz, a concurrent CP at 0.5,
0.75, and 0.9 will lower the excitation threshold of the sinusoidal stimulus by
factors of 0.60, 0.47, and 0.29 respectively. These properties can be used
to advantage in the design of focal magnetic stimulators, as described in
Sect. 9.9.
Other aspects of nonlinear response have been demonstrated in experi-
ments of human visual and auditory response to stimuli consisting of pairs
of frequencies (Adrian, 1977). In these experiments, subject response de-
pended on the difference frequency of two sinusoids mixed together, even
though the individual frequencies had much higher thresholds if presented
singly. These experiments are discussed in greater detail in Sect. 9.8.
Sensitivity to difference frequencies are a characteristic of nonlinear sys-
tems, and that is a likely explanation for the observations in these experi-
ments. The site of the nonlinear action was not explored in these
experiments. It would be useful to examine whether the electrodynamics of
the nerve membrane are responsible for such results.
4.7 Parameter Variation Effects

To further explore mechanisms for electrocutaneous stimulation, a param-
eter variation study was conducted with the myelinated nerve model (Reilly
and Bauer, 1987). The membrane parameters varied were: fiber diameter
(d), nodal gap width (W), internal axonal conductivity (ga) , membrane
capacity (em), and membrane conductance (gm). These were individually
varied by factors of one half or two times the base-case value, while holding
all other base-case parameters constant. For this study, the base-case fiber
diameter was 1O.um; other parameters were as in Table 4.1. Geometric
factors varied relative to the base case were radial distance from the central
node and longitudinal distance between the electrode and the fiber
terminus.
Four aspects of model response were evaluated in the parameter varia-
tion study:
1. Minimum charge (Qo)-the AP initiation threshold in units of charge of
a short duration pulse (l.us). Charge is determined by the product of
pulse width and current magnitude.
2. Minimum current (/o)-the AP initiation threshold in units of peak

current for a long pulse (2ms).
3. S-D time constant (Te)-determined by a least-squares fit of the thresh-
old currents for AP initiation to the theoretical S-D curve for an ideal
linear membrane. (See Reilly and Larkin, 1983).
4. Polarity sensitivity ratio (P)-the absolute value of the AP initiation
thresholds for anodic versus cathodic stimuli. P was determined sepa-
rately for stimulus durations of 10 and 1,OOO,us.
Table 4.4 summarizes results from the parameter variation study. The
first row lists data for the base case in absolute units. The remaining entries
in the table are expressed as multiples of the base-case values. Part A lists
results for variations of fiber parameters that were set to one half or two
times the base case value. In varying D, we kept the electrode distance
equal to one internodal space. As a result, electrode separation, expressed
in absolute distance units, varied with D. Part B lists results for variations in
the radial distance (yo) between the electrode (positioned over the center
node) and the fiber; Yo is expressed as a multiple of internodal units and is
a unitless quantity. Part C lists results for variations in longitudinal distance
with respect to the truncated end of the axon; Xo = -1 means that the
electrode was positioned one internodal unit beyond the terminus; Xo = 2
and 5 means that the electrode was positioned above the second and fifth
nodes, respectively, from the end.
In Part A of Table 4.4, we see that P is quite insensitive to the choice of
membrane parameters. And except for em, the membrane parameter varia-
tions have only a modest effect on Le. In Part B, we see that the radial
separation of the electrode from a central node has a significant effect on Le.
The fact that Le falls with reduced distance is consistent with the theoretical
expectations discussed in Sec. 4.3. In Part C, we see that the value of P can
be significantly lowered when the electrode is placed near the truncated end
of the axon. The degree of this reduction places the model values of P much
more in conformance with experimental data from electro cutaneous sen-
sory experiments.
Only passive membrane parameters were varied in the study. Bostock
(1983) also examined the effects of variations in active membrane param-
eters. That study used a model for current injection at a single point on the
model axon, rather than excitation through an external electrode as in the
myelinated nerve model. As a result, it is difficult to compare Bostock's
results with the myelinated nerve model. Despite differences in model
assumptions, Bostock's results provide some guidance on the potential
sensitivity of Te to active membrane parameter variations. Of the param-
eters examined by Bostock, the sodium activation rate constant had the
greatest effect: for a factor of two reduction in the activation rate constant,
there resulted at most a 21 % increase in the value of Le.
Parameter Variation Effects 147
TABLE 4.4. Summary of parameter variation effects.

P for P for
Qo 10 Te t = O.01ms t = 1.0ms
Base case 15.9nC 0.18mA 92.3,us 4.66 5.53
A. Variation of membrane parameters
D xli, 0.50 0.50 1.00 1.00 1.00
x2 2.00 2.00 1.00 0.99 1.00
G x II, 0.76 0.82 0.93 0.98 1.10
x2 1.45 1.28 1.13 0.95 0.96
ri X II, 0.70 0.99 0.71 1.00 1.00
x2 1.51 1.11 1.50 0.99 0.98
Cm X II, 0.70 0.99 0.71 1.00 1.00
x2 1.51 1.11 1.50 0.99 0.98
gm X II, 1.27 1.12 1.13 0.98 0.98
x2 0.82 0.87 0.93 0.99 1.01
B. Variation of radial distance-electode centered along axon
Yo = 1/4 0.12 0.14 0.61 1.06 1.24
Yo = II, 0.31 0.34 0.90 1.06 1.18
Yo = 1 1.00 1.01 1.00 1.00 1.00
Yo = 2 4.28 3.58 1.17 0.92 0.83
Yo = 4 23.0 15.9 1.39 1.03 0.94
C. Variation of longitudinal distance-electrode near truncated end
Xo = -1, Yo = 1 1.78 1.65 1.08 0.77 0.87
Xo = 2, Yo = 1 1.12 1.24 0.93 0.53 0.70
Xo = 5, Yo = 1 1.00 1.01 0.99 0.77 0.48
Xo = -1, Yo = 2 4.48 3.79 1.17 0.95 0.89
Xo = 2, Yo = 2 5.38 4.92 1.12 0.69 0.56
Xo = 5, Yo = 2 4.31 3.67 1.15 0.32 0.25
First row lists values in absolute units. Other table entries list values as a multiple of the
base-case datum. Xo and Yo are dimensionless quantities (normalized to node spacing).
D = 10,um for base case.
Source: From Reilly and Bauer (1987).
We note that stimulation near the end of the truncated model axon
results in values of P that are more nearly in line with experimental data.
A possible hypothesis for experimental observations is that the principal
site of transcutaneous sensory stimulation may be at or near neural end
structures.
5
Electrical Properties of the Heart
HERMANN ANTONI
5.1 Cardiovascular System: General Anatomical

and Functional Aspects
The heart provides the main driving force for the movement of blood
through the vessels. It is composed of two hollow organs-its right and its
left half-with muscular walls (Fig. 5.la). Each half comprises an atrium
(Ra, La) and a ventricle (Rv, Lv). The right half receives oxygen-depleted
blood from the body and expels it via the pulmonary artery (Pa) to the
lungs, where it is reoxygenated. Then the blood returns to the left half of the
heart and is thence distributed via the aorta (Ao) to the organs of the body
(Fig. 5.lb). The movement of the blood from the right to the left heart, by
way of the lungs, is called the pulmonary circulation. Its movement to and
from the rest of the body is the systemic circulation. Strictly speaking, the
two constitute a single pathway of blood movement, with the propulsive
force provided at two points by the two halves of the heart (Fig. 5.lb).
The pumping action of the heart consists of a rhythmic sequence of
relaxations (diastole) and contractions (systole) of the chambers. During
diastole, the ventricles fill with blood, and during systole, they expel it
into the large vessels (aorta and pulmonary artery). Backflow from these
arteries to the ventricles is prevented by the valves at their openings.
Before entering the ventricles, the blood passes from the large veins into
the associated atria, which act as booster pumps to help fill the ventricles.
An additional pair of valves between the atria and the ventricles prevents
backflow of blood during the ventricular systole. In the systemic circulation
the arteries carry oxygenated blood, and in the pulmonary circulation the
oxygenated blood is carried by the veins.
Because the demands made on the circulating blood vary with time, the
heart must be able to adjust its activity over a wide range. For example, the
volume of blood expelled per minute by one ventricle (cardiac output) is
about 5 liters for an adult person at rest and rises to almost 30 liters during
hard physical work. Because the two ventricles are arranged in series, their
outputs must be nearly the same at each beat. This requires a mechanism
148
Cardiovascular System: General Anatomical and Functional Aspects 149
for precise adjustment of the outputs of the two ventricles. Moreover, when
the resistances to flow in the systemic or pulmonary circulation increase-
for instance, because of extensive vasoconstriction-the ventricles quickly
adapt to the changed conditions by contracting more strongly and raising
the pressure sufficiently to propel the same volume of blood. Likewise,
changes in venous return and diastolic filling are compensated by adjust-
ment of the cardiac output. This astonishing adaptability of the heart arises
from both intracardial (auto-), and extracardial regulation.
The former is brought about by intrinsic properties of the myocardium,
mainly its ability to respond to increased extension (greater end-diastolic
volume) with the development of a higher contractile force or of a higher
stroke volume. The latter mechanism operates under the control of the
endocrine and autonomic nervous systems. It enables the heart either to
overcome a higher pressure or to eject a larger stroke volume without
increased muscular extension, that is, without increase in the end-diastolic
volume. Under normal conditions the autoregulation is brought into play
when changes in filling occur without a general increase in physical activity,
for instance because of changes in the position of the body that affect
venous return.
The energy the heart requires for its mechanical work comes primarily
from the oxidative decomposition of nutrients. In this regard cardiac and
skeletal muscle differ fundamentally, for the latter can obtain a large part of
the energy needed to meet short-term demands by anaerobic processes, and
the "oxygen debt" that is built up can be repaid later. This is not possible
with the heart, because it is exclusively dependent on oxidative energy
supply. Although the weight of the heart of an adult is only about 0.5% of
the total body weight, the O 2 demand of the heart is about 10% of the total
resting O 2 consumption. When the body is performing hard work, the O 2
consumption of the heart can rise to four times the resting level.
The coronary vessels, which supply the heart, are part of the systemic
circulation. There are two coronary arteries, both arising from the base of
the aortic root (Fig. 5.2). The right coronary artery supplies most of the
right ventricle; the rest of the heart is supplied by the left coronary artery.
Venous drainage is mainly through the coronary sinus, which flows into the
right atrium. At rest the total coronary blood flow amounts to about 250 to
300ml/min; this is about 5% of the cardiac output. During normal resting
activity the heart withdraws more oxygen from the blood than do the other
organs. Of the 20rnlldl of O 2 in the arterial blood, the heart extracts around
14 mlldl. Therefore, when the load on the heart increases and more oxygen
is required, it is essentially impossible to increase the rate of extraction.
Increased O 2 requirement must be met primarily by increased blood flow,
brought about by dilation of the vessels. The strongest stimulus to dilation
of the coronary vessels is O 2 deficiency.
Because cardiac metabolism relies so heavily on oxidative reactions to
provide energy, a sudden interruption of circulation (ischemia) results in
(a)
Pulmonary circulation
------~--
/'..
(b)
150
Origin and Spread of Excitation 151
FIGURE 5.2. Human heart in a frontal view with the main coronary arteries originat-
ing from the aortic root. Venous return occurs via the coronary sinus to the right
atrium.
extensive loss of function within a few minutes; contractions grow progres-

sively weaker, a marked dilation develops, and after about 6 to lOmin the
heart stops beating. Because coronary circulation is normally maintained by
the pumping action of the heart itself, a strong diminution of the contractile
force of the organ must also impair its blood supply, thus, in turn, weaken-
ing contraction still more. The same sequence of events takes place with
ventricular fibrillation (see Sects. 5.8 and 5.10). When a circulatory break-
down affects the entire organism, the brain suffers irreversible damage after
ischemia lasting only 8 to lOmin.
5.2 Origin and Spread of Excitation

Myocardial fibers, like nerve or skeletal muscle fibers, are excitable struc-
tures. The cell boundaries, which can be seen in the microscope as interca-
lated discs, offer no obstacle to the conduction of excitation. Because the
FIGURE 5.1. (a) Simplified anatomy of the opened human heart in a frontal view.
(b) Schematic diagram of the connections of the two halves of the heart with
the pulmonary and systemic circulations. Ra = right atrium; La = left atrium; Rv =
right ventricle; Lv = left ventricle; Ao = aorta; Pa = pulmonary artery.
152 5. Electrical Properties of the Heart
musculature of the atria and ventricles forms a netlike structure, it behaves

as a syncytium in which excitation arising anywhere in the atria or ventricles
spreads out over all the unexcited fibers. This property provides the expla-
nation of the all-or-none response of the heart; that is, when stimulated the
heart either responds with excitation of all its fibers or gives no response.
In a nerve or skeletal muscle, by contrast, each cell responds individually,
so that only those fibers exposed to suprathreshold excitation discharge
conducted impulses, while the others remain at rest.
Autorhythmicity and Geometry of Propagation

The rhythmic pulsation of the heart is maintained by excitatory signals
generated within the heart itself. Under suitable conditions, therefore, a
heart removed from the body will continue to beat at a constant frequency.
This property is called autorhythmicity. Ordinarily, the spontaneous rhyth-
mic triggering of excitation is performed exclusively by the specialized cells
of the pacemaker and conducting system. The various elements in this
system are diagrammed in Fig. 5.3.
Normally, the heartbeat is initiated in the sinoatrial (SA) node, in
the wall of the right atrium near the superior vena cava. When the body is
at rest, the SA node drives the heart at a rate of about 70 impulses/min.
From the SA node the excitation first spreads over the working myocar-
dium of both atria. The only pathway available for conduction to the
ventricles is indicated in Fig. 5.3. All the rest of the atrioventricular
boundary consists of nonexcitable connective tissue. As the excitation
propagates through the conducting system it is briefly delayed in the atrio-
ventricular (AV) node. Propagation velocity is high (about 2m/s) through
the remainder of the system-the common bundle, the left and right bundle
Sino-atrial
node _ _+
Common bundle
Bundle branches ~-~~--E-----
and
Purkinje fibers
FIGURE 5.3. Arrangement of the pacemaker and conducting system of the human
heart as seen in frontal section.
Origin and Spread of Excitation 153
branches and their terminal network, the Purkinje fibers-so that the differ-
ent ventricular regions are excited in rapid succession. From the Purkinje
fibers, excitation spreads at a speed of about 1 mls over the ventricular
musculature.
Hierarchy of Pacemaker Activity and

Artificial Pacemakers
The autorhythmicity of the heart is not entirely dependent on the operation
of the SA node, since the other parts of the conduction system are also
spontaneously excitable. But the intrinsic rhythm of these cells becomes
considerably slower, the farther away from the SA node. Under normal
conditions, therefore, these cells are triggered by the more rapid buildup of
excitation in the higher centers, before they have a chance to trigger them-
selves. The SA node is the leading primary pacemaker of the heart, because
it has the highest discharge rate.
If for any reason the SA node should fail to initiate the heartbeat, or
if the excitation is not conducted to the atria (sinoatrial block), the A V
node can substitute as a secondary pacemaker at a frequency of 40 to
60/min. If there should be a complete interruption of conduction from the
atria to the ventricles (complete heart block), a tertiary center in the
ventricular conducting system can take over as pacemaker for ventricular
contraction.
In the case of the above-mentioned complete heart block, atria and
ventricles beat entirely independently of one another, the atria at the
frequency of the SA node and the ventricles at the considerably lower
frequency of a tertiary center (30-40/min). When there is a sudden onset
of total heart block, several seconds can elapse before the ventricular
automaticity "wakes up." In this pre automatic pause an insufficient
supply of blood to the brain may cause unconsciousness and convulsions
(Adams-Stokes syncope). If the ventricular pacemakers fail altogether,
the ventricular arrest leads to irreversible brain damage and eventually to
death.
When conduction along the bundle branches is interrupted, the cardiac
rhythm is not disturbed as long as at least one branch or subdivision of a
branch remains functional. In this case the excitation spreads out from the
terminals of the intact conduction system and eventually covers the whole
ventricular myocardium; the time required for complete excitation is con-
siderably longer than normal. In the absence of autorhythmicity it is there-
fore possible to keep the blood in circulation by artificial electrical
stimulation of the ventricles. Electrical stimulation can sometimes be con-
tinued for years. The stimuli are generated by subcutaneously implanted,
battery-driven miniature pacemakers and conducted to the heart by wire
electrodes.
5.3 Elementary Processes of Excitation

and Contraction
The action potential of the cardiac muscle cells, like that of neurons or
skeletal muscle fibers, begins with a rapid reversal of the membrane poten-
tial, from the resting potential (approximately -90mV) to the initial peak
about +30mV see Fig. 5.4). This rising phase of the action potential (phase
0) lasts only a few milliseconds. During the subsequent period of repolariza-
tion, there are three phases that can be more or less clearly distinguished in
different regions of the heart: Phase 1 is an initial short phase of repolariza-
tion, during which the membrane potential approaches zero. Phase 2 is a
prolonged plateau following the initial peak, which is a very characteristic
feature of cardiac muscle. Phase 3 is the terminal repolarization approach-
ing the resting level (phase 4). The action potential of the cardiac muscula-
ture lasts about 200 to 400ms-more than 100 times as long as that of a
skeletal muscle or nerve fiber (see Chapter 3). The functional conse-
quences, as we shall see, are considerable (see Sect. 5.4).
Ionic Mechanisms of Excitation

The action potential is generated by a complicated interplay of membrane
potential changes, changes in ionic conductivity, and ion currents. Funda-
mentals of the ionic theory of excitation have been discussed in detail in
Chapter 3. Here we shall give only a short recapitulation, with reference to
the specific peculiarities of cardiac muscle (for details see Noble, 1984;
Pelzer and Trautwein, 1987; Carmeliet, 1992; Antoni 1996). The resting
potential of the myocardium is primarily a K+ potential mainly determined
by a specific K+ channel (gKh see Table 5.1) and maintained by the electro-
genic Na+ pump. As in the neuron, the rapid upstroke phase of the action
potential is brought about by a brief pronounced increase in Na+ conduc-
tance, gNa' which results in a massive Na+ influx (see Fig. 5.4). This initial
Na+ influx, as in the neuron, is very rapidly inactivated. Hence, further
mechanisms are required for the considerable delay in repolarization of the
cardiac muscle tissue. These are (1) delayed activation and slow inactiva-
tion of L-type Ca2+ channels (gCaL) which causes a depolarizing influx of
calcium (slow inward current) and (2) a decrease in the resting (inwardly
rectifying) K+ conductance (gKl) due to depolarization, which reduces the
repolarizing K+ outward current.
Repolarization of the myocardium results from a gradual decrease in gCa
and delayed activation of another population of K+ channels (gK) which are
activated by depolarization at the beginning of the action potential. The
decrease in gCa diminishes the slow inward current, and the increase in gK
enhances the K+ outward current. Yet another population of K+ channels
(gTO) are responsible for the Phase 1 repolarization which is differently
+50
-,--- - -i- -- - - - - - - - - - - - --
Initial peak (1)
::;-
.S-
ea
~~ t / Plateau (2)
+=0
c O-.-ea
~c.. ea E
Q)
E ~ Upstroke (0)
/ Repolarization (3)
c ~ c..
~ c.. §
..c -50
E ~ U
<c
Q)
E tiQ)
ex:
I 100 ms (4)
-100
(a) Action potential
5.0
1.0
I " "
,
0.5 "
. ...... ......
,~gca
0.1 ----
(b) conductivity of depolarizing currents
->.
'+=0>
0
5.0
::J
'ON
§ E
o~ 1.0 /
gk
I-- --
Q)C/) gk1
.. .. ............
c E
.... 0.5
--
CIS
..c
E
Q)
E 0.1 --- -- .-.
.-.-.-.-.-~.
(c) conductivity of repolarizing currents
FIGURE 5.4. (a) General form of the cardiac action potential from a ventricular
myocardial cell. (b) Changes in the conductivity of depolarizing ionic currents
during the action potential. (c) Changes in the conductivity of repolarizing ionic
currents. The magnitude of the corresponding ionic currents further depends on the
difference between the membrane potential and the equilibrium potential of the
corresponding ion.
155
TABLE 5.1. Ionic channels participating in excitation of mammalian cardiac cells.

The symbol> g < is used to characterize ionic channels and refers to their proper-
ties as conductances which carry ionic currents.
Notation Characteristics and special functions
Ionic channels mainly carrying inward currents
gNa Activated by depolarization. Inactivation is time- and voltage-
fast Na+ dependent. Responsible for upstroke of the action potential in atrial
channel and ventricular myocardium as well as in His-Purkinje system.
gCaL Activation by depolarization. Inactivation depending on voltage and
L-type Ca'+ on internal Ca2 +. Responsible for slow inward current (plateau
channel phase of the action potential, upstroke in SA node) as well as for
excitation contraction coupling.
gf Nonspecific cationic channel. Carries mainly Na+ inward current when
pacemaker activated by polarization to high membrane potentials. Responsible
channel for diastolic depolarization in His-Purkinje system and partly
responsible for pacemaker activity of the SA node.
gb Voltage independent channel in SA nodal cells carrying Na+ ions. The
background Na+ Na + current is offset by an outward K+ current at the end of the
channel action potential, but as the K+ current decays, it contributes to
pacemaker behaviour.
Ionic channels mainly carrying outward currents
gKI K+ channel responsible for maintaining the resting potential near the
inward K+ equilibrium potential in working myocardium and in His-
rectifier Purkinje system. Shutting off during depolarization, and opening
during repolarization, thus favoring the plateau phase and late
repolarization.
gK K+ channel slowly activating upon depolarization. Mainly responsible
delayed for repolarization of the action potential. Because of its slow
rectifier inactivation partly responsible for pacemaker depolarization. gK
comprises two components gKr (rapidly activating) and gK, (slowly
activating) which are differently developed in different regions of
the heart
K+ channel generating transient outward (=TO) current when
activated by depolarization. Mainly present in atrial and Purkinje
fibers as well as in subepicardial ventricular fibers. Responsible for
the early phase of repolarization (phase 1) in these cells.
Data taken from Noble (1984), Pelzer & Trautwein (1987), Sanguinetti & Jurkiewicz (1990),
Carmeliet (1992), Antoni (1996).
developed in different regions of the heart. When the membrane is at its

resting potential, the depolarizing and repolarizing currents are in balance.
Mechanism of Contraction
While the excitatory processes take place at the surface membrane of the
cardiac muscle cell, the contractile machinery is located in the interior. The
Elementary Processes of Excitation and Contraction 157
main constituents of this machinery are the "contractile proteins": actin

(molecular weight 42kD) and myosin (approximately 500kD). Actin and
myosin form the thin and thick myofilaments of the myofibrils. These are
contractile bundles about l.um in diameter, which are subdivided by the Z
disks into compartments about 2.3.um long, called sarcomeres. The struc-
ture of a sarcomere is illustrated in Fig. 5.5. There is a regular arrangement
of the two filaments, with the myosin filaments forming the A band (isotro-
pic band). On either side of the A bands are regions containing only thin
filaments, which therefore appear light; these isotropic I bands extend to
the Z lines.
,,
~
,,.....
~" ..
~
'>.'..... )
, Sarcomere
2.2J.1m
•• • •• • • •
·
•. · • · •• ·----A-Filaments----·
• • • •
.
• •·•· · · • • (Myosin) ••
• • •
· ·
• · •. · •· • . / Actin)
• ~"'~? I("Filaments
• • • •
• • • • • •
Actin -6(~~rXQ~~~~~~~
Cross-
Myosin -~--rr-4
bridges
10nm
FIGURE 5.5. Top: Electron micrograph of mammalian cardiac muscle representing
the functional unit of a myofibril. Middle: Banded structure of the myofibrils;
arrangement of the myosin and the actin filaments. Bottom: Actin filaments and
myosin filaments with cross-bridges.
The way in which they interact to bring about contraction is described by

the sliding-filament theory (Huxley and Hanson, 1954). In the relaxed
muscle the ends of the thick and thin filaments usually overlap only slightly.
When the myofilaments contract, the thin actin filaments slide over the
thick myosin filaments, moving between them toward the middle of the
sarcomere. During the sliding process neither the myosin nor the actin
filaments themselves shorten.
Excitation-Contraction Coupling
The transmission of an action potential from the excited cell membrane to
the myofibrils in the depths of the cell requires several sequential processes
in which calcium ions initiate contraction. This occurs when the concentra-
tion of intracellular free calcium rises above about 10- 7 mollI. Two intracel-
lular tubular systems are involved in this process: the transverse tubular
system (TTS) and the longitudinal system (sarcoplasmic reticulum (SPR),
see Fig. 5.6). The T system is formed by invagination of the outer mem-
brane and transmits the action potential to the interior of the cell. The
Onset of contraction Relaxation

calcium release -...... calcium elimination ~
Ca"" 3Na'
SL=-===
z
.. .. z
FIGURE. 5.6. Interplay of calcium movements (black arrows = Ca2+ release; shaded
arrows = Ca2+ elimination) and contractile activation during the onset of contrac-
tion (left) and underlying relaxation (right). At the onset of contraction,
transsarcolemmal Ca2+ influx induces a calcium-triggered Ca2+ release from the
sarcoplasmic reticulum (Triggering). During relaxation, Ca2+ is partly eliminated
from the cell and partly stored in the sarcoplasmic reticulum (Refilling). A = actin;
M = myosin; SL = sarcolemma; SPR = sarcoplasmic reticulum; TIS = transverse
tubular system; Z = Z-line of the sarcomere.
Stimulation, Propagation, and Refractoriness 159
longitudinal system which represents an intracellular calcium store is in

close contact with the T system through so-called lateral cisternae. When
the action potential is propagated along the surface and into the T system,
it causes an influx of calcium ions. The accumulation of calcium near
the lateral cisternae triggers a further mobilization of calcium from the
intracellular stores (see Fig. 5.6).
Upon repolarization of the membrane, the transmembrane influx of cal-
cium ions ceases, and the calcium ions are removed by ATP-driven calcium
pumps as well as by exchange mechanisms located in the longitudinal
system and in the surface membrane (Fig. 5.6). Relaxation occurs when the
myoplasmic concentration of activating calcium ions is reduced below
about 10-7 moUI. This reduction inhibits the interaction of actin and myosin
cross bridges, which then detach.
5.4 Stimulation, Propagation, and Refractoriness

Elementary Mechanisms
As in other excitable tissues, action potentials of cardiac cells are elicited by
depolarization of the resting membrane to threshold (Fig. 5.7). Except for
the sinoatrial and the atrioventricular nodal cells, all myocardial fibers
contain fast Na + channels, and the activation of these channels by depolar-
ization is responsible for the initiation of the action potential. Normally, the
spontaneous depolarization of pacemaker cells initiates excitation in the
circumscribed pacemaker area as soon as the threshold is attained (see Sect.
5.5). Excitation is then propagated as a result of depolarization of still-
resting fibers by their coupling to closely adjacent excited ones. Artificial
stimulation works by the same mechanism.
To stimulate the heart as a whole, it is sufficient to stimulate only a
fraction of about 50 closely coupled cardiac cells. Hence, the stimulus
intensity required to do this is about the same for small, isolated myocardial
preparations as for the whole organ. If a single isolated cardiac cell in
culture is stimulated through intracellular electrodes, a current intensity of
about 1 nA (with a duration of 5 ms) is required to elicit an action potential.
If, on the other hand, a myocardial cell is in its normal connection within the
tissue, suprathreshold intracellular stimulation, depending on the coupling
resistance, requires more than 100nA (Johna, 1989).
With extracellular stimulation, as it is applied in experiments on isolated
cardiac tissue or by implanted pacemakers, threshold current strength for a
given pulse duration depends strongly on the geometry of the electrodes
and their position with respect to the cellular alignment in cardiac muscle.
Under optimal conditions (electrodes diameter 0.7-1.0mm; pulse duration
0.5ms), pacemaker stimulation of the heart can be achieved with only 10-
14 fJA from an extracellular electrode (Irnich, 1973).
Cathodal make
: 1 \ Anodal break
! ! \
\ \ \
1
\ \ \
\
\
\
\
\
, 0
\,,~:~;0 mV
-100
FIGURE 5.7. The two types of pulsed DC stimulation; changes in membrane poten-
tial during cathodic make and during anodic break. Action potentials are elicited
when the membrane potential attains the threshold.
A sufficiently strong and long-lasting rectangular current pulse can stimu-

late twice: first, in the region of the cathode when the circuit is closed
(cathodic make), and second, in the region of the anode when the circuit is
opened (anodic break) (Fig. 5.7). This latter effect is in part the result of an
increase in recovery from inactivation of the sodium channels. In part it is
brought about by a decrease in potassium conductance during hyperpolar-
ization. This declines slowly when the stimulus is switched off, and thus
leads to a transient depolarization to threshold (dotted line in Fig. 5.7). For
more details about cardiac stimulation, see Chapter 6.
Conduction of the Action Potential

Conduction in heart muscle occurs by local circulating currents, in the same
fashion that conduction is brought about in nerve (see Fig. 3.10). These
currents are driven by the sodium potential across the membrane of the
excited region. The current displaces the charge stored on the adjacent
membrane, depolarizing the membrane toward its threshold for increasing
sodium conductance. This process will take place more quickly as more
current flows to the resting regions and as the capacity that has to be
discharged becomes smaller (Fozzard, 1979).
The structure and the functional properties of the ventricular conduction
system allow rapid spread of excitation all over the ventricles, thus bringing
about nearly synchronized contraction. This is especially pronounced in the

hearts of large animals, where, as compared with the ordinary myocardium,
the fibers of the conducting system exhibit considerably greater diameters
with a correspondingly smaller internal resistance. The intercalated disks
offer no obstacle to the conduction of cardiac excitation, because of special
structures in the apposed portions of cardiac membrane, the so-called gap
junctions or nexuses, which provide the electrical coupling between cardiac
cells.
Disturbances of propagation (slowing or block) are usually attributed to
a reduction of the excitatory inward currents: the fast sodium inward cur-
rent in the atrial and ventricular myocardium or in the conducting system,
and the slow calcium inward current in the A V node. The sodium inward
current is reduced by depolarization due to inactivation of the voltage-
dependent sodium channels. Depolarization itself may be the result of
various factors, such as loss of internal or increase of external potassium or
decrease of potassium conductance. However, the fast sodium inward cur-
rent can also be reduced in the absence of depolarization by, for example,
drugs with local anesthetic effects, which are frequently used as anti-
arrhythmics. In the A V node, where fast sodium channels are absent,
inhibition of the slow inward current by certain calcium antagonists (for
instance, verapamil) can increase the normal AV delay.
If the excitatory inward currents are critically reduced, the ability of
propagation will also depend on the geometry of the fibers: Normally, the
wave of excitation will proceed without difficulty in a direction that
branches off from the main direction at an angle of more than 90°. Under
less favorable conditions, block of conduction can preferably occur at such
branchings (Spach et aI., 1982). Moreover, block of conduction may also be
the result of an increase of the internal longitudinal resistance when con-
ductance through the gap junctions becomes reduced. Such a change occurs
when the intracellular concentration of free calcium rises above the maxi-
mal physiological concentration of about lO-smolli. This can happen when
calcium leaks into damaged cardiac cells. A similar effect is initiated by a
rise of the intracellular concentration of hydrogen ions (acidosis) (D61eze,
1970; de Mello, 1972).
Definition and Mechanism of Refractoriness

When an action potential has been elicited in a cardiac muscle cell, this cell
cannot be reexcited until its membrane potential has repolarized to a cer-
tain level. During the absolute refractory period, the cell is inexcitable; and
during the subsequent relative refractory period, excitability gradually re-
covers as indicated in Fig. 5.8. Thus a new action potential can be elicited
sooner with a stronger stimulus. Action potentials generated very early in
the relative refractory period do not rise as sharply as normal action poten-
tials and have a lower amplitude and a shorter duration (Fig. 5.8).
mV
Maximal rate of depolarization +20
'7
Vis o
200 -20
-60
100
...................
~
.... -100
o~~~--~~~~~ .. relative
-20
mV 1----------1I+--l .. total
Membrane potential 1 - - - - - - - 1 H · · .... effective
FIGURE 5.8. Left: Dependence of the maximal rate of depolarization of the action
potential (as an indirect measure of the availability of the fast sodium channels) on
the membrane potential prior to excitation. Right: Different types of refractoriness
as related to the cardiac action potential.
The term effective refractory period is used to define the phase from the
beginning of the excitatory cycle during which no propagated excitation can
be elicited. This period lasts slightly longer than the absolute refractory
period. When the threshold is not exactly determined, but instead stimuli of
a constant suprathreshold diastolic strength (for instance, twice diastolic
threshold) are used to determine the refractory period, its duration will
comprise the absolute and part of the relative refractory period, during
which the stimulus remains ineffective. This phase is usually called the
functional refractory period.
The chief cause of refractory behavior is the inactivation of the fast Na +
channels during prolonged depolarization. Not until the membrane has
repolarized to approximately -40mV do these channels begin to recover.
The duration of the refractory period is therefore, as a rule, closely related
to the duration of the action potential. When the action potential is short-
ened or lengthened, the refractory period changes accordingly. But drugs
that act as local anesthetics, inhibiting the initial Na+ influx or retarding its
recovery after inactivation, can prolong the refractory period without
affecting action potential duration. In the absence of drugs, a similar behav-
ior is observed in the A V nodal cells, which, after termination of an action
potential, remain nonexcitable for a short time. The refractory period of
these cells is thus not only voltage- but also time-dependent.
Another phenomenon that might be considered in this context is the
supernormal period. This appears as a transient decrease in threshold for
stimulation below its diastolic level immediately following the relative
refractory period. This phase coincides with the terminal phase of repolar-
ization, when the membrane potential has not yet fully recovered. It can
explain the response to electrical stimuli at this instant which is otherwise
ineffective. The decrease in threshold (increase in excitability) during the
supernormal period can be attributed to the fact that the distance between
actual membrane potential and threshold potential is reduced because the
membrane retains moderate depolarization. This is more pronounced when
the time course of the terminal phase of repolarization is slow, and it
disappears if repolarization occurs abruptly.
The prolonged refractory period protects the musculature of the heart
from too-rapid reexcitation, which could impair its function as a pump. At
the same time, it prevents recycling of excitation in the muscular network of
the heart, which would interfere with its rhythmic activity. Because the
refractory period of the excited myocardial cells is normally longer than the
time taken for spread of excitation over the atria or ventricles, a wave of
excitation originating at the SA node or a heterotopic center can propagate
over the heart only once and must then die out, for it encounters refractory
tissue everywhere. Reentry (defined in Sect. 5.9) thus does not normally
occur.
An action potential triggered immediately following the relative refrac-
tory period of the preceding impulse is normal, as Fig. 5.8 shows, in
upstroke rate and amplitude. Its duration, however, is distinctly less than
that of the preceding action potential. In fact, there is a close relationship
between the duration of an action potential and the interval that preceded
it, and thus between duration and repetition rate. The main cause of this
phenomenon is an increase in the potassium conductance, which outlasts
the repolarization phase of the action potential and returns only gradually
to the basal level. When the interval between action potentials is short, the
increased K+ conductance accelerates repolarization of the next action
potential.
When cardiac muscle is exposed to 50- or 60-Hz alternating current, it
will respond in a way illustrated by Fig. 5.9. The first depolarizing AC half-
wave triggers an action potential and then remains ineffective until the end
of the corresponding (functional) refractory period, when a new action
potential starts. For the reasons mentioned above, the duration of the
second action potential is shorter, and so is its refractory period. Thus, 60-
Hz AC, when applied to cardiac muscle, leads to a rhythmic response, but
at a frequency increased in inverse proportion to the refractory period.
5.5 Regular and Ectopic Pacemakers

Elementary Events in Impulse Formation
The working myocardium of atria and ventricles is not automatically active:
as outlined above, action potentials are generated by spread of excitation.
Stimulation with alternating current
+
, t
! Il
0
,
I
--.I
1s -100
mV
FIGURE5.9. Electrical response of a single myocardial cell to stimulation with 50-Hz
AC for 1.3 s. Microelectrode recording from isolated myocardium of rhesus
monkey.
In all cardiac muscle cells that are capable of autorhythmicity, depolariza-

tion toward the threshold occurs spontaneously. This elementary process
of excitation can be observed directly by intracellular recording from a
pacemaker cell. As shown in Fig. 5.10, the repolarization phase of such an
action potential is followed-beginning at the maximal diastolic potential-
by a slow depolarization which triggers a new action potential when
the threshold is reached. The slow diastolic depolarization (pacemaker
potential) is a local excitatory event, not propagated as is the action
potential.
Actual and Potential Pacemakers

Normally, only a few cells in the SA node are responsible for timing
the contraction of the heart (actual pacemakers). All the other fibers in
the specialized tissue are excited in the same way as the working muscula-
ture, by conducted activity; that is, these potential pacemakers are rapidly
depolarized by currents from activated sites before their intrinsic slow
diastolic depolarization reaches threshold. Comparison of the two pro-
cesses (Fig. 5.10) shows how a potential pacemaker can assume the leading
role when the actual pacemaker ceases to function. Because the slow
diastolic depolarization of the potential pacemaker takes longer to
reach threshold, its discharge rate is lower. In the working myocardium,
there is no automatic depolarization; the upstroke of the action potential
triggered by the imposed current rises sharply from the resting-potential
baseline.
Regular and Ectopic Pacemakers 165
Working Actual Potential

mV myocardium pacemaker pacemaker
20
0+---~r----3~-------.--+-~~--~-------+--~~~~
20
40
60
Changes in frequency
due to influence on :
~
(1) Slow diastolic depolarization
(2) Threshold potential
(3) Maximal diastolic potential
FIGURE 5.10. Top: Typical forms of action potentials in working myocardium as

compared with actual and potential pacemakers. Bottom: Different mechanisms by
which changes in frequency of pacemaker cells are brought about.
According to current opinion, the slow diastolic depolarizations of the

SA node are mainly caused by the slow decline of the K+ conductance that
had increased during repolarization. This requires a relatively high back-
ground Na+ conductance to produce depolarization. In the ventricular
conducting system, the background Na+ conductance is normally low.
Cathodal polarization
-.-J L
m{_~~~l
-100 I 15
FIGURE 5.11. Induction of automatic activity in a cell of working myocardium
(cat ventricle) by cathodic polarization.
Therefore, the membrane potential reaches relatively high levels just after
the action potential, which permits extensive recovery of the rapid Na+
system. The subsequent diastolic depolarizations involve a special ionic
channel that does not operate in the SA node.
Ectopic Pacemakers
The capacity for spontaneous excitation is a primitive function of myocar-
dial tissue. In the early embryonic stage, all the cells in the heart are
spontaneously active. As fetal differentiation proceeds, the fibers of the
prospective atrial and ventricular myocardium lose their autorhythmicity
and develop a stable, high resting potential. But the stability of the resting
potential can be lost under various conditions associated with partial depo-
larization of the membrane (catelectrotonus, stretching, hypokalemia, Ba2 +
ions) (see Fig. 5.11). Then the affected fibers can develop diastolic depolar-
izations like those of natural pacemaker cells and, in some circumstances,
can interfere with the rhythm of the heartbeat. On the other hand, depolar-
ization due to elevated K+ does not produce autorhythmicity, because a
concomitant rise in K+ conductance inhibits spontaneous activity. A center
of autorhythmicity apart from the regular pacemaker tissue is called an
ectopic center or ectopic focus.
5.6 Effects of Autonomic Nerves and of Changes

in Electrolyte Composition
The cardioregulatory centers in the brain exert a direct influence on the
activity of the heart, by way of sympathetic and parasympathetic nerves.
This influence governs the rate of beat (chronotropic action), the systolic
contractile force (inotropic action), and the velocity of atrioventricular
conduction (dromotropic action). These actions of the autonomic nerves
are mediated in the heart, as in all other organs, by chemical transmitters-
Effects of Autonomic Nerves and of Changes in Electrolyte Composition 167
acetylcholine in the parasympathetic system and noradrenaline in the

sympathetic system.
Parasympathetic and Sympathetic Innervation

The parasympathetic nerves supplying the heart branch off from the
vagus nerves on both sides in the cervical region. The fibers on the right
side pass primarily to the right atrium and are concentrated at the SA
node. The A V node is reached chiefly by cardiac fibers from the left
vagus nerve. Accordingly, the predominant effect of stimulation of the
right vagus is on heart rate, and that of left vagus stimulation is on
atrioventricular conduction. The parasympathetic innervation of the
ventricles is sparse; its main influence is indirect, by inhibition of the
sympathetic action.
The sympathetic nerve supply, unlike the parasympathetic, is nearly
uniformly distributed to all parts of the heart. The sympathetic cardiac
nerves come from the lateral horns of the upper thoracic segments
of the spinal cord and make synaptic connections in the cervical and
upper thoracic ganglia of the sympathetic trunk, in particular the stellate
ganglion. The postganglionic fibers pass to the heart in several cardiac
nerves. Sympathetic influences on the heart can also be exerted by cat-
echolamines (mainly adrenaline) released from the adrenal medulla into
the blood.
Chronotropy, Inotropy, and Dromotropy

Stimulation of the right vagus or direct application of acetylcholine to the
SA node causes a decrease in heart rate (negative chronotropy); in the
extreme case, cardiac arrest can result. Sympathetic stimulation or applica-
tion of noradrenaline increases the heart rate (positive chronotropy).
When vagus and sympathetic nerves are stimulated at the same time, the
vagus action usually prevails and is followed by a delayed sympathetic
influence when stimulation is terminated (Fig. 5.12). Modification of the
autorhythmic activity of the SA node by these autonomic inputs occurs
primarily by way of a change in the time course of the slow diastolic
depolarization. Under the influence of the vagus, diastolic depolarization is
retarded, so that it takes longer to reach threshold (Fig. 5.10). In the
extreme case, diastolic depolarization is eliminated, and the membrane
actually becomes hyperpolarized. The sympathetic fibers act to increase the
rate of diastolic depolarization and thus shorten the time to threshold (see
Table 5.2).
Because the positive chronotropic action of the sympathetic nerves
extends to the entire conducting system of the heart, when a leading pace-
maker center fails, the sympathetic input can determine when and to what
Atropin treatment
80
(sympathetic effect)
~ 60
>. 40
(.)
c:
Gl
:::I
-...
c-
Gl
CI
c:
100
:;:; Time after stimulation (s)
IV 2
Gl
..Q
.S 40
Gl
CI
c:
IV
60
.s::. Reserpine treatment
0
80 (vagal effect)
100
FIGURE 5.12. Changes in beating frequency of isolated right atrium (guinea pig)
following 50-Hz field stimulation, 10V/cm, 1 s. Dotted line: uninfluenced atria (con-
trols). Vagal effects are derived from the difference between controls and atropine
treatment, sympathetic effects from the difference between controls and reserpine
treatment.
TABLE 5.2. Main action of several influences on the electrical and mechanical
activity of the heart.
Atrial and ventricular
AVnode Purkinje system myocardium
SA node Vm" and Vm"and AP Contraction
PA CV CV PA duration Force
Warming ++ ++ + +
Cooling ++ +
Acidosis 0 0 +
Alkalosis + 0 0 + +
Ke increase
Ke decrease 0 0 0 ++ 0 +
Sympathetic ++ ++ 0 + + ++
(noradrenaline)
Parasympathetic 0 (- ) In atrium In atrium
(acetylcholine)
O 2 deficiency 0
Vm " = maximal rate of depolarization of action potential. +, increase; + +, large increase; -,

decrease; - -, large decrease; 0, no prominent effect. P A = pacemaker activity; CV =
conduction velocity; AP = action potential.
Effects of Autonomic Nerves and of Changes in Electrolyte Composition 169
extent a subordinate center takes over as pacemaker. In the same way,

however, an ectopic focus or rhythmicity can be stimulated to greater
activity, so that the danger of arrhythmia increases.
An influence of the autonomic nerves on the conduction of excitation can
normally be demonstrated only in the region of the A V node. The sym-
pathetic fibers accelerate atrioventricular conduction and thus shorten
the interval between the atrial and ventricular contractions (positive
dromotropic action). The vagus-especially on the left side-retards atrio-
ventricular conduction and in the extreme case can produce a transient
complete AV block (negative dromotropic action). These effects of the
autonomic transmitter substances are associated with a particular feature of
the cells in the A V node. As discussed above, the fibers of the A V node
closely resemble those of the SA node. Because there is no rapid inward
Na + current, the upstroke is relatively slow and hence the conduction
velocity is low. The vagus acts to decrease the rate of rise still further,
whereas sympathetic activity increases it, with the corresponding effects on
the velocity of atrioventricular conduction.
Vagal and Sympathetic Tone

The ventricles of most mammals, including humans, are influenced pre-
dominantly by the sympathetic system. By contrast, the atria can be shown
to be subject to the continual antagonistic influence of both vagus and
sympathetic nerves; this effect is most clearly evident in the activity of the
SA node. It can be observed, for example, by intersecting or pharmacologi-
cally blocking one of the two sets of nerves; the action of the opponent then
dominates. When the vagus input to the dog heart is removed, the rate of
beating increases, from =100/min at rest to ISO/min or higher; when the
sympathetic input is removed, it falls to 60/min or less. This maintained
activity of the autonomic nerves is called vagal and sympathetic tone.
Because the rate of the completely denervated heart (the autonomic
rate) is distinctly higher than the normal resting rate, it can be assumed
that, under resting conditions, vagal tone predominates over sympathetic
tone.
Mechanism of Autonomic Transmitter Actions

The effects of vagal stimulation and application of the parasympathetic
transmitter, acetylcholine, are attributed to one fundamental action-an
increase in the K+ conductance of the excitable membrane. In general, such
an influence is expressed in the tendency of the membrane potential
to oppose depolarization, as is evident in both the retardation of the
slow diastolic depolarization in the SA node, described above, and in a
shortening of the action potential of the atrial myocardium. The reduc-

tion of the rate of rise of the action potential in the A V node can also
be explained by a stronger outward K+ current that counteracts the
slow inward Ca2 + current. In the ventricular myocardium, by contrast, the
sympathetic-antagonistic action dominates-that is, the main action is
inhibition of noradrenaline release from the sympathetic nerve endings.
With respect to the mechanisms by which the sympathetic fibers (or their
transmitters) act, there is convincing experimental evidence that they in-
crease the slow inward Ca2 + current. That is, the contractile force becomes
greater (positive inotropic action) because this effect has intensified the
excitation-contraction coupling. The positive dromotropic action on the
AV node is also likely, in view of the above considerations, to be related to
enhancement of the slow inward Ca2+ current. As yet, there is no satisfac-
tory explanation of the mechanism of the positive chronotropic sympathetic
action. At the SA node, enhancement of the slow inward current is prob-
ably involved. In the case of the Purkinje fibers, however, an influence on a
specific, hyperpolarization-activated pacemaker current is likely.
Autonomic transmitter substances are thought to bind to certain molecu-
lar configurations on the effector cell (the word "receptor" is used both for
these subcellular structures and for sensory cells). The effects of noradrena-
line and adrenaline on the heart, described above, are mediated by
so-called f3 receptors. Sympathetic effects can be prevented by f3 receptor
blockers, such as dichloroisoproterenol (DCI) and pronethalol. In the
heart, as in other organs, the deadly nightshade poison atropine acts as an
antagonist to the parasympathetic effects of acetylcholine.
Effects of the Ionic Environment and of Drugs

Of all the features of the extracellular solution that can affect the activity of
the heart, the K+ concentration is of the greatest practical importance. An
increase in extracellular K+ has two effects on the myocardium: (1) the
resting potential is lowered because the gradient K/IK e+ is less steep, and
(2) the K+ conductance of the excitable membrane is increased-as it is by
acetylcholine in the atrial myocardium. Doubling of the K+ concentration,
from the normal4mmolll to about 8mmolll, results in a slight depolariza-
tion accompanied by increased excitability (threshold for stimulation re-
duced) and increased conduction velocity. Moreover, this influence causes
suppression of heterotopic centers of rhythmicity. A larger increase in K+
(over 8mmol/l) reduces excitability (threshold for stimulation increased)
and reduces conduction velocity. When the extracellular K+ concentration
is lowered to less than 4mmolll, the stimulating influence on pacemaker
activity in the ventricular conducting system dominates. The enhanced
activity of heterotopic centers can lead to cardiac arrhythmias (Table 5.2).
The excitability-reducing action of large extracellular K+ concentrations
is turned to advantage during heart operations to immobilize the heart
Electrocardiogram 171
briefly for the surgical procedures (cardioplegic solutions). While the heart
is inactive, circulation is maintained by an extracorporeal pump (heart-lung
machine). Impairment of cardiac function due to increased blood K+ during
extreme muscular effort or in pathological conditions can be largely com-
pensated by sympathetic activity.
Under normal conditions, changes in extracellular Ca2 + are of limited
interest for the function of the heart, because the neuromuscular excit-
ability is much more sensitive to such influences. Under experimental con-
ditions, changes in the extracellular Ca2+ concentration rapidly affect the
force of cardiac contraction. Complete excitation-contraction uncoupling
can be achieved by the experimental withdrawal of extracellular Ca2 +.
Table 5.2 summarizes the most important physical and chemical influ-
ences on excitation and contraction of the heart; only the dominant effects
are considered.
5.7 Electrocardiogram
As excitation spreads over the heart, an electric field is produced that can
be sensed on the surface of the body. The changes in magnitUde and
direction of this field in time are reflected in alterations of potential differ-
ences measurable between various sites on the body surface. The electro-
cardiogram (ECG) represents such potential differences as a function of
time. It is thus an indicator of cardiac excitation-not contraction!
Because the directly measured potentials usually amount to less than
ImV, commercially available ECG recorders incorporate electronic ampli-
fiers; these contain high-pass filters with a cutoff frequency near 0.1 Hz (a
time constant of 2s). Therefore, DC components and very slow changes of
the potentials at the metal recording electrodes do not appear at the output.
All electrocardiographs have a built-in means of monitoring amplitude in
the form of a 1-mV calibration pulse set to cause a deflection of 1 cm.
ECG Form and Nomenclature; Relation to

Cardiac Excitation
With electrodes attached to the right arm and left leg, the normal ECG
looks like the curve shown in Fig. 5.13. There are both positive and negative
deflections (waves), to which are assigned the letters P through T. By
convention, within the ORS group, positive deflections are always desig-
nated as R and negative deflections as 0 when they precede the R wave or
as S when they follow it. By contrast, the P and T waves can be either
positive or negative. The distance between two waves is called a segment.
An interval comprises both wave and segments (for instance, the PO inter-
val, see Fig. 5.13). The RR interval, between the peaks of two successive R
waves, corresponds to the period of the beat cycle.
h -
c
CI.I
)(
CI.I
Q.. -
c
iii
CI.I
c
0
-
CI.I E E
CI.I CI.I > .-
> C) 0
E >
CII
CII !II
~ ~
-g
CII C)
CI.I U CI.I ~
~ !II (/) !II ::J
I-
r 11. a
11.
a:
a
I-
(/)
Calibr.
1 mvn R
I I
I I
I I
I I
I I
I I
T
I I
I I
+ I I
0-' \".,.,
P
V'- -~ "..- A (U)

,"--,
Q
V
S
<0.1 s < 0.12 s

Duration QT interval
PQ interval frequency-dependent
< 0.2s at 70/min 0.32-0.39 s
FIGURE 5.13. Normal form of the ECG with bipolar recording in the direction of the
long axis of the heart. The times below the ECG curve are important limiting values
for the duration of distinct parts of the curve.
An atrial part and a ventricular part can be distinguished in the EeG.

The atrial part begins with the P wave, the expression of the spread of
excitation over the atria. During the subsequent PQ segment, the atria as a
whole are excited. The dying out of excitation in the atria coincides with the
first deflection in the ventricular part of the curve, which extends from the
beginning of Q to the end of T. The QRS complex is the expression of the
spread of excitation over both ventricles, and the T wave reflects recovery
from excitation in the ventricles. The intervening ST segment is analogous
to the PQ segment in the atrial part, indicating total excitation of the
ventricular myocardium. Occasionally, the T wave is followed by a so-called
U wave; this probably corresponds to the dying out of excitation in the

terminal branches of the conducting system.
The PQ interval is the time elapsed from the onset of atrial excitation to
the onset of ventricular excitation and is normally less than 0.2 s. A longer
PQ interval indicates a disturbance in conduction in the region of the A V
node or the bundle of His. When the QRS complex extends over more than
0.12 s, a disturbance of the spread of excitation over the ventricles is indi-
cated. The overall duration of the QT interval depends on heart rate. When
the heart rate increases from 40 to l80/min, for example, the QT duration
falls from about 0.5 to 0.2 s.
Origin of EeG
As a wave of excitation passes over a cardiac muscle fiber, a potential
gradient dVldx is generated, the magnitude of which depends on the
momentary phase of excitation. At the front of the wave, there is a steep
gradient corresponding to the amplitude of the action potential. During the
repolarization phase, there appear much smaller gradients in the opposite
direction. To a first approximation, the excited myocardial fiber behaves in
the physical sense as a variable dipole, the magnitude and direction of which
are symbolized by an arrow (vector). By definition, the dipole vector points
from minus to plus; that is, from the excited to the unexcited region; an
excited site, as seen from the outside, is effectively electronegative as com-
pared with an unexcited site. At every moment during the excitatory pro-
cess, dipole fields across the surface of the heart sum to an integral vector.
As this occurs, a large fraction of the vectors will neutralize one another,
as observed from outside the system, because they exert equal effects in
opposite directions.
When excitation spreads over the atria (P wave), the predominant direc-
tion of spread is from top to bottom; that is, most of the individual depolar-
ization vectors generate an integral vector pointing toward the apex. When
the atria are excited as a whole, the potential differences disappear tran-
siently, for all the atrial fibers are in the plateau phase of the action poten-
tial (PQ segment). Only when the excitation moves into the ventricular
myocardium do demonstrable potential gradients reappear. Spread of exci-
tation over the ventricles begins on the left side of the ventricular septum
and generates an integral vector pointing toward the base of the heart
(beginning of QRS). Shortly thereafter, spread toward the apex predomi-
nates (largest QRS vector). During this phase, excitation moves through the
ventricular wall from inside to outside. Spread through the ventricles is
completed with the excitation of a basal region of the right ventricle, at
which time the integral vector points toward the right and up (end of QRS).
While the excitation was spreading over the ventricles (QRS), it died out in
the atria. When the ventricles are totally excited (ST segment), the poten-
tial differences disappear briefly, as they did during atrial excitation (PQ
segment).
If repolarization of the ventricles took place in the same sequence as
depolarization and at the same rate, the behavior of the integral vector
during recovery would be approximately the opposite of that during the
spread of excitation. This is not the case because the process of repolariza-
tion is fundamentally slower than that of depolarization and rates of
repolarization are not the same in the different parts of the ventricles.
Repolarization occurs sooner at the apex than at the base and sooner in the
subepicardial than in the subendocardial layers of the ventricles. Thus,
during the ventricular recovery phase (T wave), the direction of the integral
vector hardly changes; it points to the left. The different curve forms
obtained with the arrangement of leads ordinarily used, on extremities
and chest wall, are basically projections of the momentary integral vectors
onto certain lead axes.
ECG Recording
In bipolar recordings, two recording electrodes are placed at defined sites
on the body surface, and the potential difference between these electrodes
is monitored. In unipolar recordings, only one recording electrode is placed
at a defined site, and the potential is measured with respect to a reference
electrode. This electrode can be thought of as positioned at the null point
of the dipole, between positive and negative charge. In clinical practice,
the following recording arrangements are the most commonly used today
(Fig. 5.14).
Limb leads:
Bipolar: standard Einthoven's triangle (leads I, II, III)
Unipolar: Goldberger's augmented limb leads (aVR, aVL, aVF)
Chest leads:
Bipolar: so-called small chest triangle of Nehb (D, A, I), not shown
in Fig. 5.14
Unipolar: Wilson's precordial leads (V1-V6)
Einthoven's Triangle
Because in bipolar recording from the limbs by the method of Einthoven
the arms and legs act as extended electrodes, the actual recording sites are
at the junction between limbs and trunk. These three points lie approxi-
mately on the corners of an equilateral triangle, and the sides of the triangle
represent the lead axes. Figure 5.15 illustrates the way in which the relative
amplitudes of the various ECG deflections in the three recordings are
derived from the projection of the frontal plane vector onto the associated
lead axes.
aVR aVL aVF
-------
III
'--+l-.......J
Einthoven's leads
~
aVF
Goldberger's leads
~
~
~
V4
---L---
~
V6
Wilson's leads
r"'l 100 ms
FIGURE 5.14. Arrangement of ECG leads in common use. Right: Typical curves
recorded from a healthy subject.
Types of QRS Axis Orientation

The direction of the largest integral vector (the chief vector) during the
spread of excitation is called the electrical axis of the heart. When the spread
of excitation is normal, its direction in frontal projection agrees well with
the anatomical long axis of the heart. Therefore, limb recordings can be
used to infer the orientation of the heart. The various categories are based
on the angle a between the electrical axis and the horizontal. In the normal
range (shown at the top in Fig. 5.15), the angle to the horizontal varies
from 0° to 90°. Angles above the horizontal are given a negative sign. The
general categories of QRS axis orientation are: normal range (0° < a <
+90°); right axis deviation (+90° < a < + 180°); left axis deviation (-120°
< a < 0°).
For the construction of the electrical axis from the ECG by means of
Einthoven's triangle (Fig. 5.15, top), two lead pairs suffice, for the third can
be derived from the other two. At each instant during the excitatory cycle,
it holds that: deflection in II = deflection in I + deflection in III (downward
deflections having negative sign).
aVL
aVF
FIGURE 5.15. Top: Einthoven's triangle. The recording sites at the extremities are
represented as the corners of an equilateral triangle, and the sites of the triangle
correspond to the lead axes. Angle a of the QRS electrical axis indicated on the
outer circle. The projection of an integral vector with angle a + 58° on the three
axes is shown, and the magnitude of the various deflections in each axis is indicated
by the customary curves. Bottom: Summary of axis orientations with the unipolar
(Goldberger) and bipolar (Einthoven) limb leads.
In Goldberger's method, the voltage measured is that between one

extremity-for example, the right arm (lead aVR)-and a reference elec-
trode formed by voltage division between the two other limbs (see Fig.
5.14). With aVR recording, the lead axis on which the vector loop is pro-
jected is represented by the line bisecting the angle between I and II in the
Einthoven triangle (Fig. 5.15). The axes for aVL and aVF are found in an
analogous way.
Unipolar Precordial Leads

Whereas the limb leads just described are fundamentally related to the
frontal projection of the vector, the unipolar precordial leads of Wilson
provide information chiefly about the horizontal vector projection. A refer-
ence electrode is produced by joining the three limb leads, and an exploring
electrode records from specific points on the chest at the level of the heart
(see Fig. 5.14). A positive deflection is seen when the instantaneous vector,
projected onto the appropriate axis, points toward the recording site. If it
points in the opposite direction, the deflection is negative.
Use of the ECG in Diagnosis

The EeG is an extremely useful tool in cardiological practice, for it reveals
changes in the excitatory process that cause or result from impairment of
the heart's activity. From routine EeG recordings, the physician can obtain
information pertaining to the following:
Heart rate: Differentiation between the normal rate (60-90/min at rest),

tachycardia (over 90/min), and bradycardia (below 60/min).
Origin of excitation: Decision whether the effective pacemaker is in the SA
node or in the atria, in the A V node or in the right or left ventricle.
Abnormal rhythms: Distinction among the various kinds and sources (sinus
arrhythmia, supraventricular and ventricular ectopic beats, flutter and
fibrillation).
Abnormal conduction: Differentiation on the basis of degree and localiza-
tion, delay or blockage of conduction (sinoatrial block, AV block, right or
left bundle-branch block, fascicular block, or combinations of these).
QRS axis orientation: Indication of anatomical position of the heart; patho-
logical types can indicate additional changes in the process of excitation
(unilateral hypertrophy, bundle-branch block, etc.).
Extracardial influences: Evidence of autonomic effects, metabolic and
endocrine abnormalities, electrolyte changes, poisoning, drug action
(digitalis), etc.
Primary cardiac impairment: Indication of inadequate coronary circulation,
myocardial O 2 deficiency, inflammation, influences of general pathologi-
cal states, traumas, innate or acquired cardiac malfunctions, etc.
Myocardial infarction (complete interruption of coronary circulation
in a circumscribed area). Evidence regarding localization, extent, and
progress
It should, however, be absolutely clear that departures from the normal

EeG except for a few typical modifications of rhythmicity or conduction as
a rule give only tentative indications that a pathological state may exist.
Whether an ECG is to be regarded as pathological or not can often be
decided only on the basis of the total clinical picture. In no case, can one
come to a final decision as to the cause of the observed deviations by
examination of the ECG alone.
5.8 Abnormalities in Cardiac Rhythm as Reflected

in the ECG
A few characteristic examples may indicate how disturbances of rhythmic-
ity or conduction can be reflected in the ECG. The recordings, where not
otherwise indicated, are from Einthoven's limb lead II (see Fig. 5.14). We
first consider an analytical diagram as it is frequently used to illustrate the
origin and spread of excitation in the heart. The bar SA in Fig. 5.16 symbol-
izes the rhythmic discharge of the SA node. The successive stages in the
spread of excitation are shown from top to bottom, with the absolute
refractory periods of the atria (A) and ventricles (V) indicated along the
abscissa by rectangles. Figure 5.16a shows a normal ECG with the pace-
maker in the SA node and the QRS complex preceded by a P wave of
normal shape. Above the ECG trace, the process of excitation is dia-
grammed in the described way.
Rhythms Originating in the AV Junction

A source of rhythmicity in the A V junctional region (the A V node itself and
the immediately adjacent conductile tissue) sends excitation back into the
atria (including the SA node) as well as into the ventricles (Fig. 5.16b).
Because excitation spreads through the atria in a direction opposite to
normal, the P wave is negative. The QRS complex is unchanged, conduction
occurring normally. Depending on the degree to which the retrograde atrial
excitation is delayed with respect to the onset of ventricular excitation, the
negative P wave can precede the QRS complex [Fig. 5.16b, (1)], disappear
FIGURE 5.16. Top: Illustration of the spread of excitation. The rhythmic discharge of
the SA node is symbolized in the bar SA. The successive stages in the spread of
excitation are shown from top to bottom, with the absolute refractory periods of the
atria (A) and ventricles (V) indicated along the abscissa. A V summarizes the total
atrioventricular conduction. Below: Normal time course of cardiac excitation (a)
and of various disturbances (b-g): (b) Excitation generated at various parts of the
AV junctional region. (c) Excitation originating in the ventricles spreads more
slowly, and the QRS complex is severely deformed. (d) Interpolated ventricular
extrasystoles of different origins. (e) Ventricular extrasystole with fully compensa-
tory pause; S = normal SA interval. (f) Supraventricular extrasystole with incom-
plete compensatory pause. (g) Complete (third-degree) AV block.
Abnormalities in Cardiac Rhythm as Reflected in the ECG 179
in it (2), or follow it (3). These variations are designated as upper, middle,

and lower A V junctional rhythms.
Rhythms Originating in the Ventricles

Excitation arising at an ectopic focus in the ventricles spreads over various
paths, depending on the source of the excitation (Fig. 5.16c). Because
myocardial conduction is slower than conduction through the specialized
system, the duration of spread through the myocardium is usually consider-
ably extended. The differences in conduction path can cause pronounced
deformation of the QRS complex.
Extrasystoles
Beats that fall outside the basic rhythm and temporarily change it are called
extrasystoles. These may be supraventricular (SA node, atria, A V node) or
ventricular in origin. In the simplest case, an extrasystole can be inter-
polated halfway between two normal beats and does not disturb the basic
rhythm (Fig. 5.16d). Interpolated extrasystoles are rare, since the basic
rhythm must be slow enough that the interval between excited phases is
longer than an entire beat. When the basic heart rate is higher, a ventricular
extrasystole is ordinarily followed by a so-called compensatory pause.
As shown in Fig. 5.16e, the next regular excitation of the ventricles is
prevented because they are still in the absolute refractory period of the
extrasystole when the excitatory impulse from the SA node arrives. By
the time the next impulse arrives, the ventricles have recovered, so that
the first postextrasystolic beat occurs in the normal rhythm. But with
supraventricular extrasystoles or ventricular extrasystoles that penetrate
back to the SA node, the basic rhythm is shifted (Fig. 5.16f). The excitation
conducted backward to the SA node interrupts the diastolic depolarization
that has begun there, and a new cycle is initiated.
Atrioventricular Disturbances of Conduction

The EeG observed in cases of complete A V block is shown in Fig. 5.16g. As
previously mentioned, the atria and ventricles beat independently of one
another-the atria at the rate of the SA node and the ventricles at the lower
rate of a tertiary pacemaker. Incomplete AV block is characterized by
interruption of conduction at intervals, so that (for example) every second
or third beat initiated by the SA node is conducted to the ventricles (2: 1 or
3: 1 block, respectively).
Atrial Flutter and Fibrillation

Atrial flutter and fibrillation are arrhythmias resulting from an uncoordi-
nated spread of excitation, so that some atrial regions contract at the same
Abnormalities in Cardiac Rhythm as Reflected in the ECG 181
time as others are relaxing (functional fragmentation). Atrial flutter is

reflected in the EeG by waves with a regular sawtooth shape and a fre-
quency of 220 to 350/min, which take the place of the P wave (Fig. 5.17a).
Because of temporary A V block due to the refractory period of the
Defibrillation
f 11 A
~~~ 120 mm Hg
300 ms 80
~:
Blood pressure
FIGURE 5.17. (a-d) ECG changes during flutter and fibrillation. (a) Atrial flutter
with sawtooth-shaped flutter waves and 4: 1 A V conduction. (b) Atrial fibrillation
resulting in absolute arrhythmia of the ventricles. (c) Ventricular flutter. (d) Ven-
tricular fibrillation. Time scale in f valid for all curves. (e, f) Induction and termina-
tion of ventricular fibrillation by electric current.
ventricular conducting system, normal QRS complexes appear at nearly

regular intervals. In the ECG associated with atrial fibrillation (Fig. 5.17b),
atrial activity appears as high-frequency (350-600/min) irregular intervals
(absolute arrhythmia). There is a continuum of intermediate states
between atrial flutter and fibrillation. In general, the hemodynamic effects
of these atrial arrhythmias are only slight, and the patients are frequently
quite unaware of the arrhythmia.
Ventricular Flutter and Fibrillation

With ventricular flutter and fibrillation electrical activity is uncoordinated,
and the ventricles do not fill and expel the blood effectively. Circulation is
arrested, and unconsciousness ensues; unless circulation is restored within
minutes, death results. The ECG during ventricular flutter exhibits high-
frequency, large-amplitude waves (Fig. 5.17c); whereas the fluctuations
associated with ventricular fibrillation are very irregular, changing rapidly
in frequency, shape, and amplitude (Fig. 5.17d). If flutter persists for several
minutes, it changes more and more into fibrillation. Flutter and fibrillation
can be set off by many kinds of heart damage: oxygen deficiency, coronary
occlusion (infarction), cooling, overdoses of drugs or anesthetics, etc.
Ventricular fibrillation is also the most common acute cause of death in
electrical accidents.
5.9 Mechanism of Flutter and Fibrillation

There are two alternative concepts for the mechanisms underlying flutter
and fibrillation. The first, originally proposed by Engelmann (1875) and
later supported by Rothberger and Winterberg (1941), Scherf (1947), and
others, assumes ectopic automaticity as the main cause of fibrillation. The
second concept, proposed by Mines (1914), Garrey (1914), and Lewis, et al.
(1920), postulates that circulating excitation waves (reentry) are responsible.
The first case assumes that there are one or more ectopic foci that, by firing
at high rates, overcome the regular formation and conduction of impulses in
the heart. In the second case, fibrillation is attributed to a primary distur-
bance in the spread of excitation, in which excitatory waves circulate
throughout extended regions of the heart.
It cannot be ruled out at the moment that either of the two mechanisms
might operate under certain conditions. However, most of the phenomena
that accompany the induction of fibrillation by any means can be explained
convincingly only by the hypothesis of reentry. Direct evidence in favor of
this hypothesis was presented by Allessie et al. (1973) for isolated rabbit
atria and by Janse et al. (1980) for the left ventricles of isolated perfused
porcine and canine hearts after coronary occlusion.
Mechanism of Flutter and Fibrillation 183
Conditions for Reentry; Anatomical and

Functional Pathways
Reentry requires interconnected pathways along which excitation can pro-
ceed continuously. Since the days when Mines and Garrey developed the
concept of reentry, many efforts have been made to identify anatomical
structures in the heart that might represent reentry pathways. Rosenblueth
and Garcia Ramos (1947) were able to show experimental evidence for
such structures in the atrial wall surrounding the ostia of the great veins. A
second pathway for reentry was shown by Wit et al. (1972) in meshes
formed by Purkinje fibers (see Fig. 5.18), the excitability of which had been
depressed. Moreover, Allessie et al. (1977) demonstrated reentry in a flat
sheet of the atrial wall without macroscopically visible circuit pathways,
showing that the development of reentry does not strictly depend on any
preexisting ringlike anatomical structure, but may occur when the electro-
physiological properties of myocardial tissue create a functional state
that offers comparable conditions. Thus, a refractory zone may simulate a
nonexcitable anatomical obstacle around which reentry can occur. In any
case, reentry requires an excitation wave with a refractory zone shorter than
the given interconnected pathway.
Length of the Excitation Wave

In a myocardial strand assumed to be infinite in length, excitation would
proceed in a wavelike fashion. The distance between the excited and recov-
ered parts is about 0.3 m, as can be calculated by multiplying the conduction
velocity (=l.Om/s) by the duration of the myocardial action potential
(=0.3 s). Obviously, such a large excitation wave cannot bring about reentry
in a human heart, because it is longer than any conceivable cardiac pathway.
Thus, reentry requires the excitation wave to be shortened, either by
decreasing the velocity of conduction, shortening the refractory period, or
both.
The consistent relationship between heart size and probability of
fibrillation was an early finding of Garrey (1914), who first emphasized
that there must be a critical myocardial mass (or size) for reentry to
occur. Furthermore, this relationship explains why sustained fibrillation
is more likely in large hearts, whereas small hearts tend to defibrillate
spontaneously.
Mechanisms of Abbreviated Refractory Periods

Several mechanisms can lead to an abbreviation of the cardiac excitatory
process: (1) interval-dependent shortening of the cardiac action poten-
tial with increasing frequency of excitations, (2) shortening because of
Normal spread Wave of excitation Wave shortened +

of excitation shortened unidirectional block
.. no reentry .. reentry
• l[
FIGURE 5.18. Representation of a loop-shaped piece of the ventricular conducting

system (top). Absolute refractory zones of the excitation waves are represented in
black, relative refractory zones are dashed. Left: The wave of excitation travels
along the conducting system from top to bottom. Where the wavefronts meet,
they extinguish each other because none of them can proceed into the absolute
refractory zone of the other one. The same events will occur if the waves
become shortened (middle). Right: Reentry is possible only if shortening of
the waves of excitation is combined with transient block of conduction in one
direction. This may occur when the heart is stimulated during its early relative
refractory period .
Mechanism of Flutter and Fibrillation 185
metabolic disturbance caused by ischemia, and (3) shortening because of

the electrotonic interaction between excited (depolarized) tissue proximal
to a conduction block and unexcited tissue beyond the block.
1. An action potential initiated soon after a preceding one will be consid-
erably shorter. The shorter the interval between the two action potentials,
the shorter the duration of the second one. This influence depends on the
rate of recovery from inactivation of the depolarizing ionic currents and
on the rate of inactivation of the repolarizing currents (see Chapter 3).
If an action potential is initiated before these processes are complete, its
duration will be shorter than usual (Carmeliet, 1977).
2. In mild ischemia or in the early phase following coronary occlusion,
there is a marked decrease in action potential duration accompanied
by shortening of the refractory period (Elharrar et aI., 1977). In more
severe ischemia, the action potential duration decreases further, but the
refractory period tends to lengthen, thus displaying the phenomenon of
postrepolarization refractoriness (Lazzara et aI., 1975).
3. If an action potential is propagated in excitable tissue, its shape
depends not only on the active membrane properties of the excited tissue,
but also on the passive membrane properties of the tissue still unexcited.
During the fast upstroke of the action potential, the active membrane area
supplies current to the neighbouring resting area, thus initiating excitation.
This local circulating current, which depolarizes the resting membrane,
simultaneously tends to repolarize the active membrane and to shorten the
action potential. If propagation fails at a site of a conduction block, it can
hasten the repolarization of the active membrane area proximal to the
block. In this way, the duration of the action potential proximal to the site
of a conduction block can be shortened by about 50% (Sasyniuk and
Mendez, 1971).
Mechanisms of Slow Conduction

Very low conduction velocities of about 0.1 m/s or less can be obtained
experimentally in myocardial preparations perfused with solutions contain-
ing both potassium in high concentrations and epinephrine (Cranefield
et aI., 1971). In such a medium, the fast inward current is inactivated by
the potassium-induced depolarization, and simultaneously the slow inward
current is enhanced by epinephrine. The result is an action potential, known
as a slow response, which is slowly propagated with a conduction velocity
of about 0.02 to 0.08m/s. Under these conditions, it is possible to induce
reentry within a loop of Purkinje fibers shorter than 15 mm.
Another possible mechanism of conduction delay occurs as a result
of the electrotonic spread of excitation across a segment of inexcitable
tissue (Jalife and Moe, 1981). An in excitable segment-for instance, an
ischemic zone-may be electrically well coupled with surrounding tissue
but be unable to generate an all-or-nothing response. Thus, the spread of
excitation across such tissue will be determined by the local circuit current
that causes a depolarization in the excitable tissue beyond the inexcitable
zone. A depolarization of sufficient amplitude may elicit an action potential
in the distal tissue, although with marked delay. Such an electrotonically
mediated conduction delay can result in conduction velocity less than
0.01 mls and thus may allow reentry along circuit pathways only approxi-
mately 2 to 3mm in length (Antzelevitch and Moe, 1981). Moreover, if the
conduction delay across an inexcitable gap outlasts the refractory period of
the tissue proximal to the gap, the excitation of the tissue beyond the gap
can reexcite the proximal tissue by electrotonic transmission.
Conduction velocity also depends on the direction of propagation as
related to the fiber orientation: propagation is slower in the direction trans-
verse to the long cell axis than in the direction of this axis. Measurements of
the axial resistance exhibit considerably higher values in the direction trans-
verse to the long cell axis than in the longitudinal direction (Clerc, 1976;
Spach et aI., 1981). A possible explanation for the increased transverse axial
resistance may be given by the fact that a substantial part of the resistance
is located in the cell-to-cell junctions. Because the current in the transverse
direction must pass through a junction every cell width, there are more cell
junctions per unit length in the transverse direction, and hence the effective
axial resistance will be greater.
Unidirectional Block and One-Way Conduction

To accomplish reentry, the excitation wave must travel in only one direc-
tion; that is, conduction must be blocked in one branch of the loop (Fig.
5.18, right). Furthermore, the blocked area must permit conduction in the
opposite direction of the excitation wave arriving via the unblocked branch
of the loop. If the effective refractory period of the area proximal to the
block is shorter than the conduction time along the pathway, it will be able
to reenter the loop. Thus, the occurrence of unidirectional conduction is an
essential condition for reentry.
Probably the most important cause of unidirectional conduction block in
cardiac tissue is given by non uniformity in the recovery of excitability
leading to spatial variations in refractoriness. As illustrated in Fig. 5.18
(right), the two pathways of the loop differ in their refractoriness. An
impulse entering the system is therefore blocked in the right limb because
of the still-existing refractory state of this region, whereas it can travel
freely in the left limb where excitability has already recovered. After some
delay (because of the travel time), the latter impulse can reenter the right
limb, where the refractory period has meanwhile expired. Experimental
evidence for the significance of nonuniformity in excitability was shown
by Allessie et ai. (1976), who found that reentrant tachycardias could be
induced in a sheet of rabbit atrial tissue only if the excitation wave was
elicited at a region of markedly inhomogeneous refractoriness.
Vulnerable Period: Threshold for Fibrillation 187
Unidirectional conduction can also be associated with electrotonic

propagation across a functional obstacle. Successful electrotonic transmis-
sion across an inexcitable gap in Purkinje fiber preparations depends on the
mass of excited tissue proximal to the gap. If this tissue mass is too small to
provide enough current for the excitation of a larger mass of tissue beyond
the gap, conduction will fail. However, conduction in the opposite direction
can succeed because the larger tissue mass is able to exert a sufficient
depolarizing influence (Jalife and Moe, 1976).
Nonuniformity of active or passive cell properties leading to asymmetry
in conduction are inherent features of cardiac tissue. Although not appar-
ent under normal conditions where the safety factor of propagation is high,
this nonuniforrnity can become obvious by the occurrence of unidirectional
block if propagation is critically impaired, for instance, during propagation
of a premature beat.
5.10 Vulnerable Period: Threshold for Fibrillation

Atrial or ventricular fibrillation can be induced by electrical stimulation
coinciding with late atrial or ventricular systole-the so-called vulnerable
period (Wiggers and Wegria, 1939). In the ventricles, the nonuniformity in
recovery of excitability is maximal preceding the apex of the T wave of the
ECG (Fig. 5.19). At this time, electrical stimulation elicits an excitation
wave that encounters some regions fully recovered, others partially recov-
ered, and others still absolutely refractory. Propagation of an electrically
induced wavefront can thereby be initiated preferentially in certain direc-
tions, thus setting the stage for multiple reentry-the electrophysiological
basis of ventricular fibrillation. The normal heart is susceptible to ventricu-
lar fibrillation only during this phase; at other times, stimulation is ineffec-
tive or results only in a single extrasystole.
In addition to the induction of fibrillation by electrical stimulation,
a spontaneously generated extrasystole that occurs during a highly non-
uniform phase of recovery may also induce fibrillation. This is the basis
of the so-called R-on-T phenomenon in the ECG, which is considered a
preliminary symptom of imminent fibrillation. Under certain conditions,
the duration of the vulnerable period may span a longer time should
marked nonuniformity in recovery persist, such as after a premature
beat that exaggerates non uniformity of recovery (Wegria et aI., 1941).
Furthermore, the vulnerable period may extend even beyond the T wave
if the expiration of the refractory period of the myocardium lags behind
the repolarization. This has been demonstrated during ischemia, in
which postrepolarization refractoriness occurs (Williams et aI., 1974).
Under such conditions, late stimulation or a late coupled premature
beat may induce tachycardia or ventricular fibrillation (EI-Sherif et aI.,
1975).
R
Vulnerable period
~
Inexcitable Excitable Excitable

(refractory) with re-entry but no re-entry
FiGURE 5.19. Diagram to explain the vulnerable period of the ventricles. The
triangle-shaped figures below the ECG curve symbolize (as in Fig. 5.18) the
branched network of the conducting system of the myocardium. In the vulnerable
period, the conduction pathway is still partially refractory, so the wave of excitation
generated by stimulation can propagate in only one direction and induce reentry in
the same way as illustrated in Fig. 5.18. The ventricles would still be nonexcitable if
they were stimulated earlier; at a later time, reentry is no longer possible because
normal spread of excitation again takes place.
The Threshold for Fibrillation

The extra nonuniformity required for fibrillation is brought about in part by
the direct influence of the current itself, which favors depolarization below
the cathode and repolarization below the anode (see Chapter 4). If the
inherent nonuniformity of excitability of the heart is low, the required
increase in nonuniformity brought about by the applied current must be
large, and thus the fibrillation threshold will be high. On the other hand, if
the inherent nonuniformity is more pronounced, little additional extra
nonuniformity will be required, and any agency that increases the degree of
asynchrony of excitability recovery in heart muscle will facilitate the induc-
tion of ventricular fibrillation and lower the fibrillation threshold. Such
conditions are given by localized or global myocardial ischemia, hypother-
mia, overdosage of quinidine or of ouabain, as well as by chloroform

anesthesia (Han and Moe, 1964). The temporal dispersion of recovery
of excitability and the degree of ventricular vulnerability are closely
related. Likewise, metabolic acidosis predisposes the heart to ventricular
fibrillation, whereas alkalosis exerts an opposite effect (Gerst et aI., 1966).
In canine experiments with coronary occlusion, the fibrillation threshold
was also lowered if the site of stimulation remained unaffected by ischemia
(Han, 1969). This means that, in such cases, the reduction of the fibrillation
threshold is most probably the result of increased irregularity in the propa-
gation of premature impulses rather than to a change of the excitability
itself. Watson et aI. (1973) determined ventricular fibrillation thresholds in
human subjects in cardiopulmonary bypass using 50-Hz sinusoidal current
applied both through epicardial electrodes and endocardial electrodes,
alternatively. In the former case, the minimum fibrillating current was
735flA RMS; in the latter case, 67 flA. However, in one patient with grossly
abnormal electrical activity of the heart, ventricular fibrillation could be
induced with only 15 flA. This finding stresses the risk of micoelectrocution
in hospitalized patients. In another study, performed in patients at the time
of cardiac surgery, Horowitz et aI., (1979) measured the fibrillation thresh-
old using a lOO-Hz train of rectangular pulses applied to the epicardial
surface of the left ventricle. In 10 patients with normal left ventricles, the
fibrillation threshold was 33.6 ± 9.5mA. However, in 12 patients with 75%
obstruction of the left anterior descending coronary artery, the ventricular
fibrillation threshold was only 18.6 ± 6.9mA. Hence, there is no doubt that
a diseased heart is much more susceptible to fibrillation by electric shock.
Relation to the Threshold for Stimulation

The absolute value of the fibrillation threshold also depends on factors that
determine the stimulation threshold, including the size and position of the
electrodes, the current waveform, the duration of current flow, and the
transition resistance. Only if these circumstances can be kept constant can
relative changes in the fibrillation threshold be attributed to changes in the
vulnerability of the heart. In Fig. 5.20, mean values for both thresholds are
compared as the stimulus duration is increased from 0.5 ms to several milli-
seconds (Younossi et aI., 1973). The values of the stimulation threshold
reflect the strength-duration curve for the isolated whole heart. The curve
drawn from the values for the fibrillation threshold shows an essentially
similar shape, but it is shifted to a higher level by a factor of 10 to 20. Similar
findings have been obtained by in-situ measurements from different species
(Brooks et aI., 1955).
The different levels of the two thresholds can be explained by the differ-
ent conditions under which a cardiac response is elicited. To attain the
stimulation threshold, a local density of current just high enough to stimu-
late a number of cells (approximately 50) with fully recovered excitability is
- - small electrodes
- - - - large electrodes
100 Threshold for fibrillation

", .... ,~ /
50 " ."
~
-E
I I)
CI)
' ..........J:;
" .'.
' . .............. .JJ
.!! ........ , ....
~
Q. ..............
-
.§ 10
o
' . ....................
.... .... ....
..c
0, 5 .... .... ....
- ...
cCI)
...
en
FIGURE 5.20. Relationship between stimulation threshold and fibrillation threshold

in the isolated guinea pig heart (n = 14) for rectangular pulses (duration 0.5-50ms).
Note opposite effect of changes in size of stimulation electrodes on the two thresh-
olds. Large electrodes were about 0.7cm2 ; small electrodes, 0.05cm2 •
required. By contrast, to produce fibrillation, the cardiac response must be

elicited during the early phase of recovery, and that requires a higher
current intensity. Furthermore, to elicit premature responses at different
sites of the heart, the current must spread throughout a certain mass of
tissue, thus increasing the probability of multiple reentry. For the same
reason, the two thresholds respond in an opposite way to changes in the
electrode size (Fig. 5.20). Enlargement of the electrodes leads not only to a
more extensive extrapolar current spread, but also to a decrease in
the current density. The former effect is obviously responsible for the
decrease in the fibrillation threshold; the latter increases the threshold for
stimulation.
Effects of Alternating and Pulsed Direct Current

The fibrillation threshold cannot be predicted from the stimulation thresh-
old when the current covers more than one cardiac cycle. For instance,
the threshold for stimulation with pulsed DC is clearly below the value
obtained with AC (Fig. 5.21). This result can be explained by the higher
slope of the rising phase of the DC pulses compared with the sinusoidal AC
waves. By contrast, the fibrillation threshold for DC is considerably higher
than for AC (Hohnloser et aI., 1982). Clearly, this result must be explained
by causes other than the stimulatory efficacy of the two kinds of current.
These causes can be derived from Fig. 5.9. The single fiber responded to 50-
Hz AC with a series of action potentials at a frequency that depended on
the duration of its refractory period and on the current strength. When the
isolated heart is stimulated under comparable conditions, it shows a series
of extrasystoles. A DC pulse exerts its stimulating effect only during the
make-or-break phase of the current (Fig. 5.7). However, under the influ-
ence of long-lasting suprathreshold DC, the single fiber shows an additional
effect; namely, the development of spontaneous activity (Fig. 5.11). The
myocardial fiber then behaves like a pacemaker cell, exhibiting diastolic
depolarizations and rhythmic firing. The frequency of these spontaneous
action potentials, however, is much lower than during AC flow of compa-
rable strength. Similarly, the frequency of extrasystoles of the whole heart
[mAl
40 Threshold for stimulation Threshold for fibrillation
D de pulse
30 • ae (50 Hz)
20
10
o
FIGURE 5.21. Right: Comparison of fibrillation thresholds for AC (50Hz) and pulsed
DC at a duration of 1 s. Mean values ± SD from guinea pig heart (n = 5). Left:
Corresponding thresholds for stimulation. Size of electrodes was 0.05 cm 2 in all
experiments.
is lower during DC flow than during AC even at a higher current intensity.

It seems that the differences in the fibrillation thresholds for AC and DC
can be related to the different numbers of repetitive extrasystoles that
preceded the onset of fibrillation (see also Sect. 6.5).
Significance of Single and Repetitive Extrasystoles

Wegria et ai. (1941) were the first to discuss a decrease of the fibrillation
threshold after a premature response. They deduced this idea from the fact
that the vulnerable period of a premature beat is prolonged and extends
nearly to the end of the T wave, thus indicating a further state of nonuni-
form excitability. Han and Moe (1964) confirmed this conclusion by means
of multiple-electrode recordings, which showed that the dispersion of the
refractory period of nearby regions of the dog heart increased after a
premature beat and thereby enhanced the intrinsic vulnerability; they
measured a 35% decrease of the fibrillation threshold after a premature
response (Han et aI., 1966). This result was confirmed by Hauf et ai. (1977),
who found the decrease of the fibrillation threshold for a premature
response to be 34 % in the guinea pig heart.
A series of extrasystoles reduces the fibrillation threshold more so than
does a single extrasystole (Roy et aI., 1977). The dependence of the
fibrillation threshold on the number of the repetitive extrasystoles that
precede the onset of the arrhythmia explains why the fibrillation threshold
decreases with increasing stimulus duration; that is, with an increasing
number of pre fibrillating extrasystoles. After the induction of a certain
number of extrasystoles closely succeeding one another, the non uniformity
of excitability is at its maximum and cannot be increased further by induc-
ing a greater number of extrasystoles. Hence, with increasing duration of
AC flow, the fibrillation threshold reaches a minimum current plateau that
is usually not much higher than the stimulation threshold (see Fig. 6.9).
5.11 Electrical Defibrillation

The first observations on electrical defibrillation date from about the same
time as the induction of fibrillation by electric current was first described.
Nevertheless, electrical defibrillation was not performed in humans before
1947 (Beck et aI., 1947). Defibrillation requires short pulses of compara-
tively high intensity, applied through a large area. In case of success,
fibrillation is terminated instantaneously, and normal cardiac activity is
restored, usually without any sign of injury.
Defibrillation can be explained by synchronized stimulation of all excit-
able regions of the fibrillating heart by the electric shock. In this way, the
heart as a whole is converted into an absolutely refractory state which
prevents circus movement from going on. This state is comparable with the
Electrical Defibrillation 193
absolute refractory period at the end of the normal spread of excitation

(Fig. 5.19). When interrupted in this way, reentry does not reappear, pro-
vided there are no premature excitations.
The large area of application of the defibrillation pulse and its compara-
tively high intensity of up to several amperes are required to achieve a
suprathreshold stimulus strength in all regions of the fibrillating heart
where excitable gaps are present at the instant of the shock. Since the
excitable gaps perform circus movement too, defibrillation can also be
effective if comparatively weak shocks are applied as a series of pulses in a
circumscribed area via an intracardiac pacemaker electrode. Under these
conditions, defibrillation occurs through the abolition of the excitable
gaps, not instantaneously, but successively when they pass the region of
stimulation.
Without doubt, electrical defibrillation is the most effective measure
presently available against ventricular fibrillation. However, its application
will lead to full success only when the circulatory arrest due to fibrillation
does not exceed several minutes. Otherwise, irreversible damage of the
brain must be expected. When the circulatory arrest lasts only a few
seconds, unconsciousness may ensue as the first still-reversible disturbance
of the central nervous function. After about 2 min, spontaneous respiration
stops, and, increasing with time, the heart itself becomes more and more
depressed, because its blood supply is interrupted too.
6
Cardiac Sensitivity to Electrical
Stimulation
6.1 Introduction
It is important to understand the relationships between cardiac excitability
and stimulus features such as duration, frequency, repetition pattern, wave-
shape, electrode configuration, current density, and current pathway. Much
of our knowledge of these effects is derived from studies directed toward
therapeutic uses of cardiac stimulation, such as cardiac pacing and defi-
brillation. We can better evaluate cardiac hazards if we understand the
principles of cardiac electrophysiology presented in Chapter 5. The reviews
in Chapters 3 and 4, concerning principles of excitable membranes, will also
be referred to in this chapter.
6.2 Threshold Sensitivity with Respect to

Cardiac Cycle
The sensitivity of the heart to electrical stimulation depends critically on the
timing of the stimulus with respect to the cardiac cycle. Figure 6.1 (adapted
from Jones and Geddes, 1977, and Geddes, 1985) shows the threshold
sensitivity for both excitation and fibrillation as a function of the initiation
time of brief (0.5-5 ms) rectangular current stimuli. The data were obtained
in experiments with 120 dogs, with stimulation of the myocardium via a
helical bipolar electrode. The membrane waveform illustrates the action
potential (AP) that might be measured across the membrane of a single
myocardial cell. The lower waveform shows a corresponding ECG that
might be simultaneously measured with cutaneous electrodes (see Sect.
5.7). The horizontal axis of the figure applies to the initiation time of the
stimulus with respect to an ongoing cardiac cycle, measured as a fraction of
the QT interval of the ECG waveform. The QT interval is equivalent to the
AP duration, as suggested by the drawings below the graph.
The period labeled" ARP" is the absolute refractory period, during which
time the cellular membrane is nearly fully depolarized, and it is impossible

194
Threshold Sensitivity with Respect to Cardiac Cycle 195
to initiate a new AP process, no matter how strong the stimulus. The period
labeled "RRP" is the relative refractory period-this is the repolarization
phase of the AP process, during which the membrane may be reexcited
by a sufficiently strong stimulus. Following the relative refractory period
is the diastolic period, during which the tissue is most sensitive to excita-
tion. The refractory behavior of cardiac tissue is largely a result of the
deactivation variable of the cellular membrane during the AP process (refer
to Sect. 3.2).
The excitation curves shown in Fig. 6.1 reflect the refractory condition of
the excitable cardiac membrane, showing a threshold that declines sharply
during the RRP to a minimum plateau that is sustained until the next
cardiac cycle. Ventricular fibrillation (VF), however, can be induced only in
200
100
\
~
co 50 VwntricUlar fibrillation
Gl
Q.
ci:
ll
.§ 20
OJ
c 10
~
:; 5 Excitation
u
"'0
~"~O.5""
"0
.r:; 2
...
1/1
Gl
.r:;
I- 1.0
0.5 5.0
(%QT Interval)
Membrane Potential
FIGURE 6.1. Excitation and ventricular fibrillation thresholds: upper, stimulation

thresholds; middle, membrane potential; lower, ECG waveform. (Adapted from
Geddes, 1985.)
196 6. Cardiac Sensitivity to Electrical Stimulation
TABLE 6.1. Ratio of VF to excita-

tion thresholds for monophasic
current pulse.
Current duration (ms) Ratio I/IEx
0.5 145 ± 55
1 100 ± 27
2 112 ± 37
3 87 ± 25
4 81 ± 12
5 63 ± 37
10 52 ± ?
IF determined during vulnerable

period; I Ex determined during diastolic
period. Experiments with dogs (n =
120). Stimulation via bipolar helical
electrode sutured to myocardium.
Source: Data from Jones and Geddes
(1977).
a narrow window during the repolarization phase of the heart-this is the

so-called vulnerable period of the heart (see Sect. 5.1). The temporal point
in the heart cycle when an extrasystolic excitation is most likely to cause VF
displays a bell-shaped distribution. The parameters of that distribution
(mean ± S.D.) were found to be 85 ± 9% of the heart's QT duration in
sheep (Ferris et aI., 1936), and 72 ± 9% in dogs (Jones and Geddes, 1977).
Table 6.1 lists for the Jones and Geddes experiments the ratio of the
minimum fibrillation current during the vulnerable period to the minimum
excitation current during the diastolic period. The indicated ratio will
depend on the conditions of stimulation, including the stimulus waveform,
the electrode size, and the locus of excitation. The tabulated values apply to
stimulation of dog hearts through a bipolar helical electrode sutured to the
myocardium.
Only a few cardiac cells need to be excited to produce an extrasystolic
contraction (Geddes, 1985), whereas ventricular fibrillation typically re-
quires several conditions, namely: (1) excitation, (2) during the repolariza-
tion phase, and (3) involving a critical mass of the heart. Ventricular
fibrillation can be initiated by exciting a site that is partially refractory as the
normal AP propagates over the surface of the heart [see Sect. 5.9, and
Antoni (1979)]. Reexcitation at that stage can result in an extraneous AP
that fails to extinguish naturally, but rather circulates in an uncoordinated
fashion called "reentry," that fails to produce pumping. It should not be
assumed that excitation outside the vulnerable period is inconsequential to
cardiac safety considerations. As pointed out in Sect. 6.4, extrasystolic
excitation can significantly increase the susceptibility of the heart to fibrilla-
tion by a subsequent excitatory stimulus.
Threshold Sensitivity with Respect to Cardiac Cycle 197
The heart is more sensitive during the diastolic interval to stimulation at

the cathode electrode than at the anode, as are neurons (see Sect. 4.4).
Nevertheless, cardiac tissue can be preferentially excited with an anodic
stimulus during the relative refractory period (Brooks et aI., 1955; Cranfied
et aI., 1957). Figure 6.2 (from Antoni, 1979) illustrates anodic and cathodic
thresholds. An anodic pulse applied during the relative refractory period
repolarizes the heart to a level that depends on the strength of the stimulus.
If the repolarization is sufficient, the membrane may spontaneously revert
to a depolarized state if the anodic stimulus is terminated within the AP
time frame-the result can be the initiation of an extraneous AP (Hoffman
and Cranfield, 1960). Cranfield and colleagues (1957) also demonstrated
that, in the relative refractory period, anodic break may cause excitation at
some minimum threshold, but may fail to produce excitation above a higher
stimulus value. This phenomenon produces a "dead zone" between the
anodic and cathodic thresholds where excitation does not occur.
The phenomenon of excitation by anodic and cathodic make and break
currents was studied with two-dimensional cardiac simulation models
(Roth, 1995). Although in all cases, excitation was initiated through cellular
depolarization-a process most efficiently enacted beneath a cathode-it
was found that a "virtual cathode" exists somewhat distant from an anode,
much like that illustrated in Figure 4.12 for one-dimensional nerve models.
Ventricular
fibrillation
20
\ Cathode
\
\ Excitation
\
\
,
Anode
°0~--~----2~0-0----L----4~00----~----~600
Elapsed time from Q-wave (ms)
FIGURE 6.2. Excitation and ventricular fibrillation thresholds. [Adapted from

Antoni (1979). Reprinted by permission of VCR Publishers, Inc., 220 East 23rd St.,
New York, NY 10010 from: Progress in Pharmacology (Stuttgart, Gustav Fisher
Verlag), vol. 2/4, pp. 5-12, fig. 3 (page 6).]
Diastolic rheobase thresholds of excitation with this model for an electrode

length of 1mm and a surface area of 1.5mm2 were: 0.038, 0.41, 0.49, and
5.3mA, respectively, for cathode make, anode make, cathode break, anode
break.
When the stimulus sufficiently exceeds the VF threshold, a fibrillating
heart may be defibrillated by wide-scale excitation that places the heart in
a homogeneous state (see Sect. 5.11). There exists a threshold window
within which fibrillation is possible. Below the lower limit of the window,
VF is not possible because the stimulus is too feeble to cause excitation;
above an upper limit, VF is avoided by wide-scale excitation. Consequently,
exposure to high voltage or high current can result in massive injuries,
without death from VF (see Sect. 10.2).
6.3 Strength-Duration Relations for

Unidirectional Currents
An experimental relationship between the threshold intensity of a rectan-
gular current stimulus and its duration defines a strength-duration (S-D)
relationship. We have seen in Chapter 4 that experimental S-D data are
frequently fitted with a hyperbolic (Eq. 4.15) or exponential (Eq. 4.17)
expression. The parameter ie in either formulation may be determined by
the ratio of minimum charge to minimum current:
(6.1)
Equation (6.1) represents a simple and practical relationship for determin-

ing the parameter ie in both the hyperbolic and exponential relationships.
Although the S-D formulas were originally applied to neural excitation
thresholds, these curves have been used to fit empirical data for cardiac
excitation.! For example, the exponential relationship provides an excellent
fit to ventricular excitation thresholds in dog and turtle hearts (Pearce et aI.,
1982) as well as excitation thresholds of sheep cardiac Purkinje fibers
(Dominguez and Fozzard, 1970). Equally good fits have been obtained with
the hyperbolic relation for excitation thresholds of dog hearts (Jones and
Geddes, 1977). Comparison of experimental S-D curves for cardiac excita-
tion in dogs reveals that both formulations give a good fit to experimental
data, with no clear preference between the two (Mouchawar et aI., 1989).
Table 6.2 summarizes strength-duration data derived from a variety of
sources. The table is organized by experimental end points consisting of
excitation, fibrillation, and defibrillation. The column labeled "ie" lists em-
1 As Used here, excitation refers to the production and propagation of an

extrasystolic action potential-usually during the diastolic period of a previous beat.
Strength-Duration Relations for Unidirectional Currents 199
pirical time constants, which in most cases were determined by Equation

(6.1) or as specified in the cited work.
The medium value of 7:e in Table 6.2 is approximately 2.6 ms-a value that
is roughly a factor of 10 above S-D time constants typically observed for
neural excitation (see Sect. 4.5 and Table 7.1). Some general observations
are pertinent to the information presented in Table 6.2.
Electrode Size
Irnich (1980) reports that 7:e increases monotonically with electrode size, as
in Fig. 6.3. The Irnich data can be fitted by a relationship of the form 7:e =
KA 0.56, where A is the electrode area and K is a scaling constant. The
direction of this relationship follows that observed with cardiac models
(Roth, 1995) and for neural excitation (Pfeiffer, 1968), although it has not
been observed by others (Talen, 1975). Smyth et aI.'s data (1976) corre-
spond closely to the Irnich relationship. For instance, for an electrode area
of either 32 or 8mm2 , the values of 7:, indicated by Smyth et ai. are 2.3 and
1.7ms, respectively-values that correspond very closely to the Irnich data.
Theoretical considerations (Jack et aI., 1983, pp. 275, 418) predict that the
S-D time constant should increase with the size ofthe area stimulated. With
a small electrode, charge is applied rapidly to the membrane in a concen-
trated area, but more slowly to distant parts of the membrane. Conse-
quently, with a small electrode, the membrane becomes more efficiently
depolarized, and also acquires a local voltage gradient. These factors con-
tribute to lower excitation thresholds and to shorter S-D time constants. On
the other hand, if a larger area of membrane is excited, the time required
for depolarization increases, and, because the depolarization is more
diffuse, larger stimulus intensities are required for excitation. Jack et aI.,
(1983, p. 418) reported that the size of the stimulated area affects the time
constants much more in muscle than in nerve tissue.
Polarity
With cathodic stimulation, the site of depolarization is concentrated near
the electrode; with anodic stimulation, it is spread more gradually over
distant areas [see Sect. 4.4 and Reilly et aI., (1985)]. As explained above,
both excitation thresholds and S-D time constants are expected to be re-
duced as the area of stimulation is reduced. Consequently, cathodic stimu-
lation should result in shorter time constants and lower excitation
thresholds than anodic stimulation. Thalen et ai. (1975) found that cathodic
excitation thresholds were indeed below anodic thresholds by a factor of
2.6, in agreement with theoretical expectations; however, the observed S-D
time constants with anodic stimulation were smaller than with cathodic
stimulation by a factor of 2.5, a result not in conformance with the above
arguments. Further work is needed to reconcile these experimental obser-
vations with theoretical expectations.
~
0
TABLE 6.2. Strength-duration data for cardiac stimulation by square-wave current waveforms. ?'
(""l
T, (ms) Species Preparation Stimulus locus Electrode Notes Reference ~
....
0..
A. Thresholds of excitation iii·
n
3.6 Rabbit Isolated tissue Papillary muscle Large (E-fie1d) Knisley et al. (1992) en
3.8 Guinea pig Isolated tissue Papillary muscle Small Weirich et al. (1985) 0
~
1.75 Rabbit Isolated heart Left ventricle lcm2 Roy et al. (1985) f!J.
::to
2.86:!:: 0.99 Dog In vivo Trans-chest 8.1 em diam. Voorhes et al. (1983) ~.
2.14:!:: 0.97 Dog In vivo Ventricle Small Pearce et al. (1982) ~
7.31 :!:: 2.22 Turtle In vivo Ventricle Small Pearce et al. (1982)
....0
0.68 to 3.94 * * * 4-94mm2 (a) Irnich (1980) t!l
0
0.206 Dog In vivo Myocardium See note (b) Jones and Geddes (1977) n
....
1.1 Human In vivo * 12mm2 Fozzard and Schoenberg (1972) ~.
2.1 Dog In vivo Ventricle Loop-and- (c) Thalen et al. (1975) ~
catheter en
cathode
§.
0.73 Dog In vivo Ventricle Loop-and- (c) Thalen et al. (1975) E..
~
catheter anode
....
o·
~
3.75 :!:: 0.50 Sheep Long (>8mm) Near end Pointlike (d) Fozzard and Schoenberg (1972)
Purkinje fiber
29.3 Sheep Short Purkinje fiber Near end Pointlike (d) Fozzard and Schoenberg (1972)
2.63 :!:: 0.91 Dog In vivo Trans-thoraz 4.1cm dia. Voorhees et al. (1992)
2.6:!:: 0.88 Sheep Purkinje fiber Near end Pipette (d) Dominguez and Fozzard (1970)
2.3 Human In vivo * 0.32cm2 (e) Smyth et al. (1976)
1.7 Human In vivo * 0.18cm2 (e) Smyth et al. (1976)
1.49 Dog In vivo Myocardium 7.9mm2 Mouchawar et al. (1989)
B. Thresholds of fibrillation
r/l
7.7 Guinea pig Isolated heart Whole heart * Weirich et al. (1985)
(-36.0) Guinea pig Isolated heart * * (f) Hohnloser et al. (1982) §
1.70 Dog In vivo Myocardium See note (b) Jones and Geddes (1977) qs.
::r
C. Thresholds of defibrillation 6
3.90 Dog Isolated heart Whole heart Uniform field (g) Geddes et al. (1985)
2.76 Dog Isolated heart Right ventricle Catheter (g) Wessale et al. (1980)
4.01 Dog In vivo Trans-chest (g) Bourland et al. (1978) t:l
2.98 Pony In vivo Trans-chest (g) Bourland et al. (1978)
i·
:;0
g.
6.50 Dog In vivo Ventricle * Koning et al. (1975)
~.
* Information sketchy or not available. ~
(a) T, monotonically increasing with electrode size. g
(b) Electrode pair (I-em separation) wrapped around 2-mm-OD tube and sutured to myocardium.
(c) Average of loop and catheter electrode time constants.
e
2.
(d) T, strongly dependent on length of excised fiber, intracellular electrode. 9:
(e) Data from 21-month electrode implant. (;l
(f) Fibrillation induced by anodic break during vulnerable period of extrasystole; stimulus duration 30 to lOOms. Time constant not indicative of single ~.
excitation (see text).
(g) Trapezoidal pulses with slopes not exceeding 15%.
a
~a
'"
~
......
E
gj'"
8
OJ
.§
Cl
(jj
10 100
Area (mm 2 )
FIGURE 6.3. Relationship between S-D time constant and electrode area for cardiac
excitation. (Data from Irnich, 1980).
Size of Tissue Preparation

When the stimulating current is introduced through an intracellular micro-
electrode, the values of ie for small cardiac tissue samples depend critically
on the size of the preparation. In the experiments of Fozzard and
Schoenberg (1972), unusually large (29.3-ms) values of ie were seen with
small preparations. However, long samples (>8mm) of Purkinje fibers
yielded values of ie that appeared typical of whole-heart preparations
(3.75ms). The direction of this effect follows theoretical expectations for
intracellular excitation (Jack et aI., 1983). See Chapter 4 for additional
discussion of relationship between S-D time constants and the size of the
stimulated tissue.
Fibrillation by Prolonged Currents

The S-D curves discussed above apply to single cardiac responses, in which
case we expect thresbolds to follow Eqs. (4.15) and (4.17). However, when
the duration of the stimulating current is prolonged beyond the S-D time
Biphasic and Sinusoidal Stimulation 203
constant, the fibrillation threshold may continue to drop due to the effects
of multiple excitations (see Sect. 6.4). If we attempt to fit such data to the
S-D formula, we may obtain time constants that appear to exceed that
of mycardial excitation. The data presented by Hohnloser et al. (1982),
for example, display an unusually long value of 7:e• The S.D. curve for
these experiments declined steadily with stimulus duration, out to 60ms,
after which an increase in threshold was seen. According to the authors, the
fact that fibrillation could be initiated either by cathodic make or anodic
break accounted for the observed S-D relationship. The timing of
these events with respect to the cardiac cycle is of major importance,
according to the authors. Knickerbocker (1973) also observed fibrillation
thresholds that declined steadily with prolonged stimulation out to 2 s-
the maximum period of stimulation in his experiment (refer to Sect. 6.8 and
Fig. 6.11).
Consistency of Data
There is a wide spread of reported 7:e values, with a medium of about 2.6ms.
The thresholds of fibrillation listed in Table 6.2 demonstrate a particularly
wide variance in comparison with the other categories, although it is diffi-
cult to judge the extent to which this is simply a consequence of relatively
few data points. In general, it is more difficult to establish minimum
fibrillation thresholds than excitation thresholds, because the former de-
pend more critically on the on and off times of the stimulus current with
respect to the cardiac cycle, as well as on the number of prior extrasystolic
excitations (refer to Sect. 6.5).
6.4 Biphasic and Sinusoidal Stimulation

Before considering how oscillatory currents produce cardiac excitation, it is
helpful to make comparisons with corresponding properties of neural exci-
tation. Figure 4.16 illustrates S-D curves from a neuroelectric model for
monophasic and biphasic waveforms. Compared with the S-D curve for a
monophasic pulse, thresholds for sinusoidal stimuli rise at durations below
the S-D time constant. This upturn occurs because the current reversal of
the sinusoidal cycle tends to reverse an action potential that was started
by the initial phase of the stimulus. To compensate for this reversal, the
biphasic wave requires a higher current to achieve threshold. This cancella-
tion effect is most effective as the phase duration tp is reduced. For long tp'
thresholds depend more on the rate of rise and peak value of the current.
It should not be construed that thresholds for all oscillatory stimuli are
necessarily elevated above monophasic stimuli of the same phase duration.
Indeed, if the sinusoidal cycle is repeated as an oscillatory waveform,
thresholds may decrease with successive oscillations, as noted in Fig. 4.19.
In Chapter 4, we presented a functional relationship for strength fre-

quency (S-F) curves obtained from monophasic S-D curves by considering
that a sinusoidal half-cycle corresponds to a single monophasic pulse. This
simplification would, however, fail to account for high- and low-frequency
upturns in S-F curves relative to S-D curves. A better empirical fit to S-F
data from the myelinated nerve model is described by Eq. (4.35), which
involves high- and low-frequency transition parameters Ie and 10' respec-
tively. While the equation was derived from a neuroelectric model, it pro-
vides a reasonably good fit to cardiac stimulation data if the equation
parameters are chosen correctly.
Table 6.3 includes Ie values determined from S-D curves published by
several sources listed in the right-hand column. The table is organized by
beat rate effects, excitation, and fibrillation. Data apply to continuous cycle
stimuli except for rows 2 and 4, which apply to single-cycle stimulus wave-
forms. The data listed under the heading Ie were determined by a rough
curve fit of the experimental data to Eq. (4.36). Excluding item 1, the
geometric mean of the Ie values in Table 6.3 is 115 Hz. Item 1 was excluded
from this average because it was felt that the mechanism affecting the
beating rate in this experiment is fundamentally different from the other
two categories by virtue of the unusually small value Ie and the sharply
defined frequency sensitivity of the threshold current. The average value
1/(2/e) = 4.3ms compares favorably with the S-D time constant ie = 2.6ms
derived from the data in Table 6.2. This comparison appears adequately
consistent considering the great diversity of experimental methods repre-
sented in Tables 6.2 and 6.3.
Figure 6.4 plots a number of experimental threshold data points taken
from S-D curves published by several of the sources listed in Table 6.3. The
plotted data have been normalized by the minimum threshold current in the
various experiments. Figure 6.4 also shows a plot of the empirical formulas
defined by Eq. (4.35), in which the following relationships have been used
for cardiac excitation: Ie = 115Hz,/o = 10Hz, a = 1.45 for single-cycle
stimulus and 0.9 for continuous-cycle stimulus, and b = 0.8. The value of Ie
has been taken from the geometric mean of the data listed in Table 6.3. The
other values have been taken from the myelinated nerve model for neural
excitation (refer to Chapter 4). Of the data plotted in Fig. 6.4, that of
Weirich et al. (1985) apply to single-cycle stimuli; the other data apply to
continuous-cycle stimulation. The empirical curves plotted in Fig. 6.4
provide a reasonably good fit to the combined experimental data; better fits
could be obtained if the parameters Ie and 10 were determined separately for
each individual experiment. The low-frequency term, Ku should have an
upper limit, K DO that accounts for the finite threshold observed with DC
stimulation. Fibrillation thresholds for DC currents have been reported by
several authors. The data of Ferris et al. (1936) indicate K DC values of about
5.2; Biegelmeier (1987) reported values of K DC in the range of 3 to 4 for
durations of several seconds and a value of about 1.0 for durations of a
fraction of a second.
TABLE 6.3. Strength-frequency data for cardiac stimulation by sinusoidal current waveforms.
I, (Hz) Species Preparation Stimulus locus Electrode Notes References
A. Thresholds of beat rate change
3 Frog Isolated heart Whole heart Uniform field (a) Kloss and Carstenson (1982)
B. Thresholds of excitation
278 Guinea pig Isolated tissue Papillary muscle AgAgCl (b) Weirich et al. (1985)
50 Guinea pig Isolated heart Aorta/apex AgAgCl (c) Weirich et al. (1983)
C. Thresholds offibrillation
250 Guinea pig Isolated heart Whole heart AgAgCl (b) Weirich et al. (1985)
72 Guinea pig Isolated heart Aorta/apex AgAgCl (c) Weirich et al. (1983) I:IJ
79 Dog In vivo Ventricle 3.1cm' Kuge1berg (1976) -g:
148 Human In vivo * 4-mm diam. Kugelberg (1976) ~
100 Dog In vivo Various Various Geddes et al. (1971) o·
100 Dog In vivo Trans-chest 8 x 10cm (d) Geddes and Baker (1971) §
0..
120 Dog In vivo Trans-chest 1I2-in. wire (e) Geddes and Baker (1971) IJ)
§.
* Information sketchy or not available. '"o
(a) Thresholds for changing beat rate by 30%. 0.:
(b) Single-cycle stimulus. a
IJ)
(c) Current duration = Is.
(d) Current flow transverse to long axis of animal. -[
(e) Current flow longitudinal to long axis of animal.
g.
::s
~
1000
• Weirich '85
0
A , J:.
Weirich '83
e,o Kugelberg '76
_,0 Geddes '71
E
~
:s 100
0
"0
(5
.c
<Jl
~
.c
I-
ID
>
'cii
Q5 10
a:
• Continuous
t>
1
1 10 100 1000 10000
Frequency (Hz)
FIGURE 6.4. Strength-frequency data for cardiac stimulation by sinusoidal currents.

Points are examples from published experimental data, solid curves represent ana-
lytic expression for single-cycle (upper) and continuous-cycle (lower) stimuli. Two
examples are provided for each source, as indicated by open and filled symbols.
6.5 Duration Sensitivity for Oscillatory Stimuli

Cardiac sensitivity to oscillating currents is affected by a complex of factors
not present with single DC pulses, including: (1) A current reversal can
reverse an excitation process begun on an initial stimulus phase; (2) the
excitation threshold resulting from a prolonged AC current steadily drops
with the current duration; (3) an AC current may promote a train of
extrasystolic excitations, which successively make the heart more suscep-
tible to ventricular fibrillation. The first two of these effects was demon-
strated in Chapter 4 for the neural membrane-Fig. 4.19 shows a threshold
that oscillates with successive current phases, but nevertheless steadily
declines on the average. The third effect has been developed in sect. 5.10.
Excitation Sensitivity
The cancellation effect of a biphasic reversal has been observed for cardiac
stimulation. Figure 6.5 (adapted from Green et aI., 1985) shows the ven-
tricular fibrillation threshold versus the duration of a 50-Hz sine wave
Duration Sensitivity for Oscillatory Stimuli 207
initiated at its peak. A threshold minimum occurs at approximately the

point of current reversal (Sms). Although the current dip is small in com-
parison with the variance of the measurements, it was consistently repro-
duced in several experiments. Furthermore, when the stimulus was initiated
at the zero crossing of the current waveform, the minimum was shifted to
10ms. The effects of current reversal on the fibrillation threshold shown in
Fig. 6.5 is relatively weak. But, by analogy with neural stimulation, a stron-
ger effect is expected at higher stimulus frequencies. Indeed, biphasic
current reversal of a I-kHz waveform has been shown to require current
levels that are significantly elevated relative to a monophasic waveform
(refer to discussion concerning Fig. 6.18).
Stimulation by an alternating current involves a nonlinear integration
of the current that crosses the excitable membrane. The membrane con-
ductance is approximately constant well below the excitation threshold.
However, as the depolarization voltage approaches the value needed
for excitation, the conductance for membrane current influx and efflux
becomes increasingly asymmetric. As a result, membrane depolarization
caused by one phase of cyclic stimulus may be only partially offset by a
current reversal on a succeeding half-cycle. This asymmetric process can
result in a buildup of the membrane depolarization on successive cycles of
an oscillatory stimulus. This process of membrane depolarization was dem-
onstrated in the cardiac system with stimulation of the papillary muscle of
rhesus monkeys by 50-Hz alternating currents (Antoni et aI., 1969). In that
study, AC depolarization effects were observed experimentally and were
6
enc.
E 5
~'"
'" 4
(l)
E:
~
0
.L: 3
en
e
...
.L:
u.
2
>
00 5 10 15 20
Stimulus duration (ms)
FIGURE 6.5. Ventricular fibrillation threshold versus duration of 50-Hz sinusoidal

stimulus initiated at peak. Elecrodes right fore limb to left hind limb of dogs (n =
12). Means and S.D. shown. (From Green et aI., 1985.)
also verified with computer simulation of the electrical response of Purkinje

fibers.
Fibrillation Sensitivity
As with nerve stimulation (Fig 4.20), oscillatory stimuli can produce a
sequence of excitations in the heart-an effect that can significantly
enhance the biological response that might otherwise result from a single
excitation (see Sect. 5.10). This phenomenon has important safety conse-
quences, because the fibrillation threshold falls steadily with increasing
number of extrasystolic excitations. Figure 6.6 illustrates this effect with
experimental data from in-vivo excitation of dog hearts (from Sugimoto et
aI., 1967)-the vertical axis gives the VF threshold to a 60-Hz current; the
horizontal axis gives the number of prior ventricular responses produced by
a conditioning 60-Hz current with variable duration. The duration of cur-
rent for six responses was about 1 s. Over the range of one to six ventricular
responses, the average VF threshold drops by a factor of 30. When the
conditioning current was changed to a train of unidirectional rectangular
pulses and the VF stimulus was a single DC pulse, a response curve nearly
identical to that shown in Fig. 6.6 was obtained. Clearly, the ability of a
Number of responses
FIGURE 6.6. Ventricular fibrillation thresholds as a function of number of excitations

from 60-Hz stimulation. Experiments with dogs (n = 6); mean ± standard deviation
plotted. Needle electrodes inserted into epicardial surface. (Data from Sugimoto et
ai., 1967.)
TABLE 6.4. Comparison of AC and DC excitation

and fibrillation thresholds-experiments with dogs.
DC (t = lrns) AC (50Hz, t = 5s)
(rnA-peak) (rnA-peak)
Ex. thresholds 0.10 + 0.05 0.11 ± 0.03
VF thresholds 15 ± 4.3 0.26 ± 0.08
Ratio VF/Ex. 150 2.4
VF stimuli applied during vulnerable period to in-vivo dog

hearts (n = 14); 25-gauge bipolar needle electrodes spaced
5 mm apart, inserted 2 rnrn into epicardial surface. ± values
indicate standard deviation.
Source: Data from Sugimoto et al. (1967).
stimulus to produce multiple cardiac excitations significantly affects its

potency for causing fibrillation.
Table 6.4 compares excitation and ventricular fibrillation thresholds for
in-vivo stimulation of dog hearts. The VF threshold exceeds the excitation
threshold by a multiple of 150 when the stimulating current is a single DC
pulse; with 60-Hz stimulation of 5 s duration, the multiple is only 2.4. Nev-
ertheless, the 60-Hz excitation threshold is nearly the same as that for a
single DC pulse. Table 6.5 summarizes AC and pulsed DC thresholds from
the data of Hohnloser et al. (1982). With DC stimulation of 1 s duration, the
ratio of the VF threshold to the excitation threshold is 23: 1; with AC
stimulation, the ratio is only 4.2: 1.
Both alternating and direct currents can disrupt the normal rhythm of the
heart and promote multiple excitations (c.f. Fig. 3.15). The greater fibrilla-
tion efficacy of 60Hz relative to a DC pulse arises because the AC stimulus
is a more efficient promoter of multiple ventricular responses. The lowering
of the fibrillation threshold by an extrasystolic excitation arises because
extra beats create a higher degree of nonhomogeneity within the heart,
making it more susceptible to fibrillation (see Sect. 5.10). A burst of closely
TABLE 6.5. Comparison of AC and DC excitation

and fibrillation thresholds-experiments with
guinea pig hearts.
DC(t=ls) AC (50Hz, t = 18)
(rnA-peak) (rnA-peak)
Ex. thresholds 1.3 ± 0.6 2.6 ± 0.8
VF thresholds 30 ± 8.5 11 ± 2.2
Ratio VF/Ex. 23.1 4.2
Isolated perfused guinea pig hearts (n = 3); ring-shaped

electrode attached to apex of heart.
Source: Data from Hohnloser et al. (1982).
spaced extrasystoles especially enhances this effect (Hohnloser et aI., 1982).

The maximum achievable rate of extra-systolic excitation is enhanced by
the fact that the AP duration of an extrasystole can be significantly shorter
than the normal AP duration. This phenomenon was demonstrated by
simulation of Purkinje fiber response such as seen in Fig. 3.14.
Figure 6.7 illustrates the relationship between current magnitude and the
number of extrasystolic excitations induced by a 3-s burst of 50-Hz alternat-
ing (solid dots) or direct (open dots) current introduced into isolated guinea
pig hearts (Hohnloser et aI., 1982). The terminal values of each curve show
the fibrillation threshold. It is evident that both AC and DC stimulation are
capable of producing multiple excitations, but the AC stimulus is much
more effective.
Figure 6.8, from Weirich et ai. (1983), demonstrates the relationship
between the extrasystolic rate and the fibrillation threshold for sinusoidal
40
30
III
. 20
~
Q)
.0
VF threshold
E
:J
Z
10
°0~----------~~----------2~0~------3~0
Current magnitude (rnA)
FIGURE 6.7. Relationship between current strength and number of extrasystoles for
DC and 50-Hz AC stimulation of 3-s duration. Terminal values represent ventricu-
lar fibrillation thresholds. Perfused guinea pig hearts; ring electrode at apex of heart.
(From Hohnloser et aI., 1982.)
200
-c: 100
I Fibrillation threshold
Co /
« / /
E ,~ / /
t:20'"
Q)
t
'7'
' ... 8'" / /
/ /
/;/
B 10 '6",9" ... 1110.9/8

"0 .... . , ,'" 10 _... ""
o
..c 5
..... 5 " ... , , ' .......... ..,"'"
" , ............ _ " . ;
~ ----4 .... ' " ' ..... --..,.,.,,,.
..c
J- .... ,,"
................ ---*"
2 ---- Diastolic threshold
5 10 100 1000
Frequency (Hz)
FIGURE 6.8. Threshold cardiac stimulation by sinusoidal current of 1-s duration.

Ring electrode at heart apex. Isolated, perfused guinea pig hearts (n = 4). Upper
solid curve shows VF threshold; lower solid curve shows excitation threshold during
diastolic interval; broken curves show contours of constant estrasystole number.
Mean thresholds shown. (From Weirich et aI., 1983.)
currents of 1 s duration with frequency ranging from 1 to 1,000 Hz. The

upper solid line represents the fibrillation threshold; the lower solid line
represents the threshold for inducing an excitation during the diastolic
interval of the heart; the broken lines represent current levels at which the
extrasystolic rate is constant. The thresholds in Fig. 6.8 have a minimum
around 30Hz. Above that frequency, fibrillation and diastolic thresholds
follow constant excitation rates of 11 and 8 per second, respectively. Below
30Hz, fibrillation and diastolic thresholds also rise, but are associated with
decreasing excitation rates. The upturn in thresholds above 30 Hz may be
explained by the fact that as the AC frequency is increased, the alternate
phases of alternating current have an increasing tendency for mutual can-
cellation. This explanation is strengthened by the observation that thresh-
olds do not rise when the sinusoidal stimulus is replaced by a train of 1-ms
unidirectional pulses (Weirich et aI., 1983). Figure 6.8 also shows a low-
frequency upturn that is associated with decreasing excitation rate as
the frequency is lowered. Although a low-frequency upturn is expected for
an excitation curve (see Fig. 6.4), it is not easy to explain the association of
VF thresholds with decreasing excitation rates because it becomes less
probable that an excitation will coincide with the vulnerable period of the
heart.
The Z Relationship for Fibrillation by AC Currents

Various experiments have demonstrated that when VF thresholds are plot-
ted versus duration of an AC stimulus, the result is a sigmoidal-shaped
curve with an asymptotic maximum at short durations (a few cycles of
stimulation) and an asymptotic minimum at long durations (several sec-
onds). This so-called Z relationship was evident in early experiments of
Ferris et al. (1936) and was later substantiated by others.
Figure 6.9 illustrates the Z relationship for fibrillation by AC stimulation
as determined by Roy et al. (1977). In these experiments, stimulation was
via catheter electrodes of various areas, introduced into the right ventricle
of dog hearts. The authors obtained an empirical fit of thresholds with the
following equation:
(6.2)
where 10 is the minimum current corresponding to the longest duration

(1,000 cycles of 60Hz); Kn and b are empirical constants given in Table 6.6.
The value of Kn listed for an area of 500mm2was obtained by this writer
using logarithmic extrapolation of the Kn values at the three areas tested (A
= 0.22, 14, and 90mm2). This extrapolated value represents a rough esti-
mate of the threshold dependence for cardiac excitation over a large area.
Equation (6.2) provides an excellent fit to Roy et al.'s experimental data,
provided that N ;::: 3; for N :5 3, measured thresholds reach an asymptotic
maximum, whereas Eq. (6.2) continues to rise, thereby overstating the
experimental thresholds. These experiments show that the threshold de-
pendence on electrode area is a strong function of the duration of the AC
stimulus. At 1,000 cycles of 60-Hz stimulation (t = 17 s), there is an 18: 1
ratio of thresholds between the largest and smallest electrodes tested (90
and 0.22mm2); with 1 cycle of current, the ratio of thresholds is perhaps 2: 1.
The final column of Table 6.6 lists the ratio of the upper and lower plateaus
of the Z curves, based on the data from Roy et al. The upper plateau has
been determined from Eq. (6.2) at N = 3. It can be seen that the plateau
ratio diminishes as the area is increased. The value listed for A = 5OOmm2
is an estimate for large-area contacts or possibly for whole-heart
stimulation.
Figure 6.10 illustrates VF thresholds in guinea pigs versus the duration of
AC stimulation having frequencies of 5, 50, and 200 Hz. The curves for
50 and 200 Hz differ by a nearly constant multiple that reflects the fre-
quency sensitivity of the excitation process (refer to Sect. 6.4). The curve at
5Hz, in contrast, converges to the 50-Hz thresholds at short durations and
to the 2oo-Hz thresholds at long durations. The indicated VF thresholds
depend in a complex way on the individual extrasystolic excitations, their
rate, and their timing with respect to the cardiac cycle. At 1 s stimulation,
the difference in thresholds for the 50- and 200-Hz curves closely corre-
200~-'-'''''",-''-'-''nTrr-.-.-OT''''
100
~
'"
OJ
Q.
0:<:
E 10
....c
~
:::J
U
"0
0
.r:
Vl
~
.r:
J-
Number of 60 Hz cycles
FIGURE 6.9. Relationship between VF threshold and number of 60-Hz cycles. Sepa-
rate curves apply to area of catheter electrodes inserted into right ventricle. Mean
and S.D. values shown for 50 dogs. (From Roy et aI., 1977, © 1977 IEEE.)
sponds to the differential sensitivity of diasystolic stimulation (see Fig. 6.8).

The 5-Hz VF threshold is, however, elevated relative to that at 50Hz
because of its lower rate of extrasystolic production. As the duration of
stimulation is shortened, the extrasystolic production rate for the three
TABLE 6.6. Empirical constants for Eq. (6.2).

Area(mm2) Kn b Ima/l m;n
(500) (3.7 x 10- 3) (1.5) (50)
90 1.8 x 10- 3 1.5 90
14 8.3 x 10- 4 1.5 200
0.22 1.1 x 10- 4 2.0 1,000
Empirical constants for A = 90, 14, and 0.22mm2

were determined by Roy et al. (1977). Values for A
= 500mm2have been extrapolated from Roy et a!.'s
data. Values listed for Ima/lm;n represent ratio of
upper and lower plateaus of fibrillation thresholds.
200
.--.....•
100
~-~
80 5 Hz '"
" .
0--=:"::'::':8 •
\~\200H'
<t
E 60
"0
(5
.
or: 40
II>
...
~
o ,,~
,,
~n
or:
u.
> ,
...
20 50 Hz ... 6:-::.!=~::'
0
"--0-0-0-
10
2000
Duration of current (ms)
FIGURE 6.10. Fibrillation threshold versus duration of current for three frequencies.
Data from isolated perfused guinea pig hearts. (Adapted from Antoni, 1985.)
curves becomes less disparate. Consequently, for short stimulus durations,

VF thresholds more closely reflect diastolic excitation sensitivity at all three
stimulation frequencies.
Figure 6.11 illustrates VF thresholds for 20-Hz AC and DC currents
applied to dogs (left foreleg to right hind leg). Both AC and DC stimuli
result in steadily declining thresholds for increasing stimulus duration, al-
though the rate of decline in AC thresholds is much greater. As explained
above, the differential rate of decline arises because the extrasystolic rate
associated with the AC waveform is greater than that for the DC waveform.
(see also Fig. 5.21).
Considering the mechanisms discussed above, we can represent the dura-
tion sensitivity of 50/60-Hz fibrillating currents as shown in Fig. 6.12. It is
assumed that short-duration stimuli coincide with the vulnerable period of
the cardiac cycle. The thresholds in regions (1) and (2) follow a strength-
duration curve for pulsed stimuli-the assumed S-D time constant is 3ms,
as suggested in Sect. 6.3. The lower curve in region (1) is drawn under the
assumption that the stimulus current initiation time results in a unidirec-
tional (monophasic) pulse, in which case the threshold converges to a rl
relationship. If the initiation time is near a current reversal such that the
current pulse is biphasic, the threshold will be elevated as indicated by the
upper curve in region (1). For a balanced biphasic waveform, the relation-
ship will be approximately r1.s, as suggested by the discussion in Sect. 6.4
and by analogy with neural membrane properties (see Chapter 4). For
intermediate cases of biphasic asymmetry, the threshold would fall in be-
tween the two curves. In region (2), the stimulus duration is longer than the
S-D time constant (indicated as to in Fig. 6.12), and the threshold curve
converges to a constant-current relationship. In region (3), the stimulus
consists of more than one AC cycle, leading to multiple extrasystoles. The
VF threshold drops with each successive excitation, reaching a minimum
plateau in region (4).
The ratio, R, of the plateau thresholds in regions (2) and (4) and the
transition duration, t1 and t2, vary with the experimental conditions, includ-
ing the tested species. Table 6.7 summarizes empirical values of R, t 1, and t2
applying to 50/60-Hz exposure to the whole heart or through a large-area
cardiac contact electrode.
In applying animal data to humans, Biegelmeier and Lee (1980) argue
that differences in heart beat rate between humans and animals should be
taken into account, such that the ratio R and upper transition duration t2
would be increased in a human relative to an animal with a much faster beat
rate. This idea is bolstered by the data in Table 6.7 for guinea pigs-the
transition time t2 and the ratio R are both quite short with respect to the data
for larger animals having longer cardiac cycles. Biegelmeier proposes
human VF thresholds as indicated in the lower section of Table 6.7. The
2000~----~~--~------~~--~~----~
~ 1000
§
«
.E.
Q)
...
"C
::J
C.
E
...c'"
...::J~
U
100
50L-L-----L-----L-----__L-____L-____ ~
50 100 200 500 1000 2000

Current duration (ms)
FIGURE 6.11. VF thresholds versus duration for DC and 20-Hz AC currents; in-vivo
stimulation of dogs (n = 10-25). Error bars show 95% confidence limits. (From
Knickerbocker, 1973, © 1973 IEEE.)
Constant
energy
criterion
10
Duration of 60 Hz stimulus (s)
FIGURE 6.12. VF threshold versus 60-Hz stimulus duration. Numbered portions of

the curve represent regions of differing sensitivity (see text). Short-duration stimuli
are assumed to coincide with vulnerable period.
TABLE 6.7. Z-curve parameters for 50/60-Hz currents.

Threshold Transition
ratio time (s)
Species Current path R t1 t2 Reference Notes
A. Measured relationships
Sheep Right fore/ 10 0.27 1.4 Ferris et al. (a)
left hind (1936)
limb
Pig 10 0.03 4.0 Jacobsen et al.
(1975)
Dog Ventricle 50 0.05 2.0 Roy et al. (1977) (a)
(5 cm2 elect.)
Dog Fore/hind limb 25 0.02 2.0 Biegelmeier and (b)
Lee (1980)
Guinea pig * 6 0.20 0.80 Antoni (1985)
B. Postulated relationships for humans
Human Hand/foot 20 0.20 5.0 Beigelmeier and (c)
Lee (1980)
Human Hand/foot 20 0.10 2.0 Beigeimeier (c)
(1987)
* Information sketchy or not available.

(a) Area relationship extrapolated to 5cm 2 (see text).
(b) Derived from data of Kouwenhoven et al. (1959), Scott et al. (1973), and Kiselev (1963).
(c) Postulated from animal data; adjustment for beat rate; 50% probability curves.
Energy Criteria and Impulse Currents 217
bottom entry represents the recommendations of the International

Electrotechnical Commission (IEC) for human safety criteria (see Fig.
11.10).
The extent to which Fig. 6.12 would apply to frequencies other than 50/
60Hz has not been established, although its general features are expected
to be applicable to a much wider frequency range. The data of Antoni
(1985) (Fig. 6.10) indicate that over the time interval of 0.1 s to 1 s, the Z
curve for 200 Hz is shifted above the curve for 50 Hz by an approximately
constant multiplicative factor, consistent with the frequency sensitivity dis-
cussed in Sect. 6.4. If the duration of stimulation were to be reduced below
0.1 s to a fraction of a 200-Hz cycle, we would expect the 50-Hz and 200-Hz
curves to converge to the same curve indicated in regions (1) and (2).
6.6 Energy Criteria and Impulse Currents

Engineers in North America have traditionally evaluated safe current
thresholds for 60-Hz exposure using the so-called electrocution equation of
Dalziel (1960, 1968):
(6.3)
Where If is the fibrillation threshold, t is the duration of exposure, and K is
a scaling constant that depends on body weight and the assumed percentile
rank of sensitivity. The limits on t are from 8.3 ms to 5 s. Equation (6.3) is
fundamentally an energy criterion in which tt is a constant. Values of K for
various body weights and percentile ranks are given in Table 6.8, based on
Daiziel's criteria (see Fig. 6.15). The values listed under the category
"MNF" were said to apply to a "maximum nonfibrillating" current. The 50
percentile rank was also represented as an average in some of Daiziel's
writings. Daiziel concluded that VF thresholds are linearly related to body
weight, as discussed in Sect. 6.7.
TABLE 6.8. Values of K for electrocution

equation.
IF (rnA)
for percentile rank
Body weight (kg) 50% 0.5% MNF'
20 177 78 61
50 368 185 116
70 496 260 156
'MNF indicates "maximum nonfibrillating

current."
Source: Based on data from Dalziel (1968).
o Experimental minima
• Extrapolated 0.5% value
o Z-curve
~--,,-------o-~~
,, 0
\
~\
,
0.01 0.1 10
Shock duration (5)
FIGURE 6.13. Minimum VF thresholds obtained in experiments with 45 dogs using

data of Ferris and Kiselev. (Adapted from Dalziel, 1968, © 1968 IEEE.)
Daiziel based the time relationship in Eq. (6.3) on his analysis of the 50/
60-Hz fibrillation data from dogs developed by Ferris et a1. (1936) and
Kiselev (1963). Figure 6.13, for example, shows one of Daiziel's curve fits to
Ferris's and Kiseiev's combined data involving 45 dogs. The open points
represent the minimum fibrillation thresholds obtained at each tested dura-
tion. The closed points represent Dalziel's estimates of the 0.5 percentile
rank values, obtained by extrapolating the statistical distribution of VF
thresholds pertaining to each tested value of t. The solid line shows Daiziel's
energy fit to the extrapolated points; the broken indicates a Z-curve inter-
pretation of the same data. Except for one extrapolated point at 0.083 s, the
minimum experimental points and the extrapolated points both appear to
fit a Z curve.
It was not unreasonable for Dalziel to have postulated a constant energy
fit to the empirical data available to him. For one thing, the C I12 relationship
does provide a reasonable fit to the data when viewed over the range of
durations from 0.08 to 5 s, as indicated in Figs. 6.12 and 6.13. Also, an energy
criterion has a certain intuitive appeal to many engineers. Furthermore, the
role of extrasystolic excitations in lowering VF thresholds was not widely
known at the time that Dalziel published his work.
Although the electrocution equation may not correspond strictly to theo-
retical expectations, it does give a reasonable fit to the VF threshold/time
Energy Criteria and Impulse Currents 219
relationship for 60-Hz currents with durations exceeding 1 cycle. A diffi-

culty, however, is that some analysts have attempted to ascribe an energy
criterion to a wide variety of waveforms, including short-duration impulse
currents. This view has probably been reinforced by other work of Dalziel
(1953), in which he suggests an energy criterion for short-duration impulse
currents.
It should be clear from preceding discussions that the excitation or fibril-
lation capability of a stimulus is not predicted by its energy content. We
have seen in Sect. 6.3 that short-duration unidirectional currents instead
produce a well-defined strength-duration curve that is unrelated to constant
energy. For durations well below the strength-duration time constant, the
unidirectional stimulus threshold converges to a constant charge criterion
(see Sect. 6.3). If the transient is biphasic, the stimulus becomes less effec-
tive (see Sect. 6.4).
Short-duration charge thresholds for electrodes placed on the heart have
been reported for a variety of open-heart experiments and procedures. For
cutaneous electrodes, direct experimental data on transient thresholds are
lacking, and we need to draw inferences from data applying to longer
stimulus durations. Referring to Fig. 6.12, the charge threshold for a
monophasic current may be estimated as
(6.4)
where Qo is the minimum charge threshold for short durations (t < 3ms),
and IL is the peak current threshold for a long-duration (t = 3s) stimulus; R
and to are defined in Fig. 6.12. As an example, consider the median threshold
for hand-to-foot exposure of an individual weighing 20kg. From Table 6.10,
Dalziel's peak fibrillation threshold is 0.102j2 = 0.144 A. If we use R = 20
and to = 3ms, the mean fibrillation charge threshold from Eq. (6.4) is
calculated as 8.6me. That value compares favorably with the minimum
fibrilliating charge of 5mC observed experimentally by Pelaska (1963)
using capacitor discharges across the thorax of dogs. Cardiac arrhythmia's
were observed by Pelaska with capacitor discharge as low as 0.25 me. It is
probable that the later value does not represent a minimum threshold for
arrhythmia's, because stimuli below 0.25 mC were apparently not tested by
Pelaska.
When a capacitor discharge is lead through an inductance, a decaying
oscillatory stimulus is produced, with a net charge displacement still given
by Q = CV. The excitation potency of an oscillatory discharge can, under
some circumstances, be greater than a purely capacitive discharge, since the
amount of charge initially displaced can exceed the stored charge on the
capacitor (Reilly and Larkin, 1985b). Indeed, fibrillation thresholds have
been observed to be lower when inductance is added to a capacitive dis-
charge circuit (Dalziel, 1953), although this is not always the case (Pelaska,
1965). Neural excitation thresholds, expressed in terms of net capacitor
charge, have also been observed to be lower when inductance is added to
the discharge circuit (Reilly and Larkin, 1985b). To account for the overall
effect of added inductance, we would have to consider the initial displaced
charge and the frequency and decay time constant of the oscillatory current
waveform. The interrelationship among these factors has not presently
been defined.
6.7 Body-Size Scaling

Much of our empirical knowledge of electrical hazards has been derived
from animal experimentation. In interpreting such data for human ap-
plications, we must make some judgments concerning scaling from animal
data. Differences in physiology, anatomy, and size may all have a bearing
on the scaling method. Body size is one variable that has been studied
systematically.
Experimental Body-Size Data

Much of our knowledge of electrical safety is derived from animal experi-
mentation. Early experiments made use of the dog and rat for application to
human safety (Kouwenhoven and Langworthy, 1931; Hooker et aI., 1932;
Kouwenhoven, 1949). Several studies have suggested that average thresh-
olds increase with the size of the experimental animal. In early studies,
Ferris and colleagues (1936) plotted the VF thresholds for several species of
animal (guinea pig, rabbit, cat, sheep, pig, calf, and dog) and postulated a
linear relationship between body weight and the VF threshold, as shown in
Fig. 6.14. The thresholds were also correlated with the weight of the heart
itself, although the relationship appeared to be a nonlinear one. Dalziel
later used Ferris's data, along with Kiselev's (1963) data on dogs to obtain
a linear regression fit:
If = 3.68 W + 28.5 (6.5)
where W is the body weight (kg), and If is the average VF threshold (mA-
rms) applied between fore and hind limbs for a duration of 3 s. The corre-
lation coefficient for Eq. (6.5) was reported as r = 0.74. Figure 6.15 plots
Dalziel's regression line, along with the average thresholds and weights of
the larger animals tested by Ferris and Kiselev; the middle line shows
Dalziel's 0.5 percentile rank, which he obtained by extrapolating the statis-
tical distributions of the VF thresholds. Dalziel proposed a nonfibrillating
current safety criterion for a hypothetical 50-kg person as
If = 116r 1/ 2 (mA-rms) (6.6)
In this formula, the time relationship follows Eq. (6.3), and the body-weight
relationship follows the lower curve in Fig. 6.15. The minimum non-
Body-Size Scaling 221
400
o Sheep o
+ Pig ,
~ Calf o
320 o Dog
+
LL
> 160 +
80
o 20 40 60 80 100
Body weight (kg)
FIGURE 6.14. Ventricular fibrillation thresholds for several species of animal, plotted
against body weight. Large symbols indicate average values. Current path, left fore
limb to right hind limb; duration, 3s. (From Ferris et al., 1936, © 1936 AlEE.)
fibrillating criterion of Dalziel has been widely used for safety analysis in
North America (see Sect. 11.8 for additional details).
The body-weight relationship was also examined by Geddes and col-
leagues (1973). Using data from 104 animals of several species (rabbits,
puppies, one monkey, dogs, goats, and ponies), regression fits were ob-
tained as a power law of the form:
(6.7)
Table 6.9 lists Geddes' regression coefficients for several electrode arrange-
ments. Although these coefficients were obtained with a stimulus duration
of 5 s, it would not be unreasonable to apply the same values over the range
2 to 5 s, since little threshold variation was observed over that range. Varia-
tions in the coefficient K indicate a sensitivity to the current pathway, which
is discussed in some detail in Sect. 6.10.
Table 6.9 indicates that the VF threshold is nearly proportional to the
square root of body weight, rather than the linear relationship given by Eq.
400~--~--~--~~--~--~------~
,
,/
Calves /
l,LRegression line
300 Sheep /
1,/
.t
Pigs
~/
,/
,,
Cii
.
E
« 200
.g /
II) /
5 0.5% min .
...!:- fibrillating
.~Q)-
i l1,/
• A ,,/
VI
01 ,
~
100
lI
i' current
x
o 100
Body weight (kg)
FIGURE 6.15. Relationship between fibrillating current and body weight for 3-s
shocks of 60Hz (from Dalziel, 1968, © 1968 IEEE).
TABLE 6.9. Body-weight regression

parameters.
Electrodes K a r
Fore/fore 69.4 0.533 0.887
Left fore/right hind 29.7 0.510 0.929
Left fore/left hind 33.6 0.437 0.749
Five species of animal (n = 104); 60-Hz current;

applicable durations, 2-5 s.
Source: Data from Geddes et al. (1973).
Body-Size Scaling 223
(6.7). Nevertheless, the two regression formulas are in reasonable agree-

ment over a fairly wide range of body weights, as seen in Table 6.10,
especially for the front-to-hind limb thresholds. For small human body
weights (20kg), the Dalziel formula gives a more conservative estimate of
VF thresholds; at large body weights (100kg), the Geddes relationship is
more conservative.
A body-weight relationship is not universally accepted for the VF thresh-
old in human safety applications. Points of view on the subject were pre-
sented in a specialists' panel meeting (Geddes and Antoni, 1985). In that
discussion, Biegelmeier pointed out that when the VF thresholds of several
species are plotted against body weight (as in Fig. 6.14), one indeed observes
an apparent body-weight relationship. But when the thresholds of a single
species are plotted against body weight, the statistical correlation largely
disappears. Rather, the regression line may represent an interspecies de-
pendency that is dominated by factors other than body weight per se.
As a consequence of the arguments stated above, the IEC does not
include a body-weight law in its safety criteria. Rather, the IEC bases its
criteria on experimental data from dogs (Biegelmejer, 1986, 1987), which it
considers to be conservative in view of the demonstratedly lower VF
thresholds of dogs as compared with humans. Especially pertinent are
experimental thresholds obtained for both dogs and humans using the same
arrangements of electrodes placed in the heart or on the myocardium
(Green et aI., 1972; Rafferty et aI., 1975a, 1975b). Considering several end
points (ECG irregularities, pump failure, and ventricular fibrillation), me-
dian thresholds for humans were observed to be a factor of two or more
above corresponding thresholds for dogs.
A number of researchers have considered the dog heart to provide a
reasonably conservative model for human safety applications. In the experi-
ments of Green and Rafferty cited above, the smallest current having an
observable effect-that of heart rhythm disturbance-occurred at 60#A in
dogs, and at 80#A in humans. Consequently, the authors recommended for
TABLE 6.10. Median fibrillation thresholds

(rnA) for 3-s shocks using regression formulas
of Dalziel and Geddes.
W Geddes Dalziel
(kg) FIF r.F.lI.H. I.F.ll.H. F/H
20 343 137 124 102
50 558 218 185 213
70 668 259 215 286
100 808 311 251 397
Column headings indicate current pathway: F, H

indicate fore, hind limb; r, I indicate right, left.
Source: Data are based on regression formulas of
Dalziel (1968) (Eq. 6.5), and Geddes et al. (1973)
(Eq.6.7).
open-heart surgery a maximum 50-Hz leakage current of 60flA for a wire

placed in the right ventricle. That value, however, would not represent an
absolute lower limit on the VF threshold for open-heart exposure. Roy and
colleagues (1976) found the minimum VF current threshold to be inversely
related to the size of the contact electrode-a minimum value of 18.uA-rms
was observed for right-ventricle exposure of dog hearts to 60-Hz currents
through a contact electrode having an area of 0.224mm2• The threshold
voltage of the contact electrode relative to the surrounding tissue was,
however, relatively insensitive to electrode area-the minimum voltage
necessary to produce VF was found to be in the vicinity of l00mV-rms,
almost without regard to electrode size.
Scaling Arguments
As shown in Sect. 4.3, the electric field aligned with the long axis of an
excitable cell is responsible for the depolarizing force on the cell's mem-
brane. Current density is, however, a more commonly cited variable in
analyses of electric shock. For a given quantity of current injected through
the limbs, current density in the heart will generally diminish as the size of
the animal is increased. The direction of this effect corresponds to the body-
weight formulas mentioned above. On the other hand, fibrillation requires
not just excitation at some locus, but a critical mass of excited tissue. One
might argue that in the larger heart, the mass of the heart would tend to
counterbalance its reduced current density. If the fibrillation threshold were
related to the fraction of body current intercepted by the heart, as suggested
by Bridges (1985), then we would conclude that the fibrillation threshold
ought to be about the same for animals of nearly similar geometries but of
different sizes and weights.
Under the reentry theory for electrically introduced fibrillation (see
Chapter 5), below some critical size, a small heart ought to be more difficult
to fibrillate than a large heart. This could occur if the reentry paths were too
small to simultaneously support states that were excited, partially refrac-
tory, and resting. Experimental evidence indeed shows that the heart of a
small animal tends to revert spontaneously from a fibrillated to a normal
state (Zipes, 1975). This observation, however, says little about the thresh-
old of fibrillation.
Gross anatomical features can complicate the scaling of data from
animal to human. Most experimental data have been derived from
quadrupeds and applied to a bipedal human. The geometry of the heart and
thorax is considerably different between a quadruped and a human, and
it is likely that considerations other than body weight are important to
the scaling question (Lee, 1966). A further complication concerns the
position of the animal during experimentation, because the heart shifts
significantly within the thorax, depending on the position of the animal
(Bridges, 1985). Such shifts could cause the heart to alter its position with
Statistical Distribution of Thresholds 225
respect to a region of high current density when current is introduced

through the limbs.
Although there are uncertainties associated with body-weight scaling, it
is clear that interspecies differences do exist and that smaller animals gen-
erally have lower VF thresholds than larger animals. Less clear is the
appropriate scaling method within a single species, especially humans.
6.8 Statistical Distribution of Thresholds

Various experiments on both human and animals indicate the log-normal
distribution for fibrillating currents applied between the front and hind
limbs. An example is shown in Fig. 6.16 for fibrillating currents in dogs,
using combined data from the experiments of Ferris and Kiselev. The log-
normal distribution plotted as a straight line in the figure provides a much
better fit than the linear-normal distribution suggested by early investiga-
tors (Dalziel, 1968).
Table 6.11 summarizes distribution parameters extracted from the data
of several investigators who applied currents across the limbs of ex-
perimental animals. The experimental data were provided in terms of VF
thresholds for current durations ranging from 8ms to 5s and with several
different waveform types. The thresholds listed in Table 6.11 give the
observed 10 and 90 percentile ranks, normalized by the median value. The
± values indicate the standard deviation of the normalized percentile ranks
across the various experimental distributions. In all, 29 distributions were
represented in the cited reports. There was no recognizable sensitivity in
the normalized distributions with respect to the duration of current,
the waveform type, or the tested species. In view of this invariance, it is
reasonable to represent an overall statistic as listed at the bottom of the
table.
Table 6.12 provides a summary of log-normal rank values corresponding
to a distribution having a 10 percentile rank of 0.66, consistent with the
combined experiments of Table 6.11. The data are listed to the 0.5 percen-
tile rank-a value frequently used in safety analysis. The data probably do
not support projections below that value. The log-normal model indicates
that the 0.5 percentile rank lies below the median by the multiplicative
factor 0.43. This value is not substantially different from the projections of
Dalziel, who used a linear-normal model to fit experimental data below the
50 percentile rank.
Compared with the animal data discussed above, a much broader dis-
tribution has been reported for human patients undergoing open-heart
surgery for valve replacement. Watson and colleagues (1973) measured VF
thresholds in 56 patients exposed to 50-Hz alternating current for several
electrode arrangements, including 2.5-cm2 disk electrodes placed on the
epicardium, needle electrodes inserted into the myocardium, and bipolar
95
90
80
70
.:.!
c: 50
~
~
.p
c:
Q)
30
~
Q)
Q..
20
10
0.5
0.2~--~----~~--~~~~----~--~
20 100 200
Fibrillating current (rnA)
FIGURE 6.16. Statistical distribution of VF threshold for dogs; 3-s shocks applied
front-to-hind limb. (Combined data from Ferris et aI., 1936 and Kiselev, 1963, as
summarized by Dalziel, 1968).
pacing electrodes placed on the apex of the right ventricle. Figure 6.17
illustrates the distribution of fibrillating current. The original data of
Watson et al. have been replotted on coordinates on which a straight line
would conform to a log-normal distribution. The ratio of the 10 percentile
rank to the median VF current for Watson et al. 's data averaged 0.42 for the
three electrode arrangements; the 90 percentile rank was above the median
value by an average multiple of 2.54. These values indicate a significantly
broader statistical spread than with the animal data summarized in Table
6.11. It is difficult to determine whether the greater spread in human open-
Combined AC and DC Stimuli 227
TABLE 6.11. Experimental statistical distribution parameters for fibrillating currents

(fronUhind limb paths).
Current Normalized IF(mA)'
Species type Duration(s) 10% 90% Data sourceb
Sheep 60Hz 0.03 - 3.0 0.59 + 0.16 1.90 + 0.76 Ferris et at. (1936)
Dogs 60Hz 0.008 - 5.0 0.65 + 0.10 1.84 + 0.57 Kouwenhoven et all.
(1959)
Dogs 50Hz 3.0 0.67 1.57 Kiselev (1963)
Dogs 50Hz; 1.15( 0.70 + 0.08 1.48 + 0.17 Jacobsen et all. (1975)
pulsedc
Dogs 20Hz 0.1 - 2.0 0.64 + 0.06 1.59 + 0.18 Knickerbocker (1973)
Dogs DC 0.05 - 2.0 0.71 + 0.07 1.49 + 0.12 Knickerbocker (1973)
Overall data (n = 29) 0.66 + 0.10 1.69 + 0.47
a IF values are normalized by median.

bData of Ferris, Kouwenhoven, and Kiselev are summarized by Dalziel (1968).
C Duration was 150% of cardiac period.
heart measurements is a consequence of variations in the placement of the

electrodes with respect to the most sensitive excitable regions of the heart,
a reflection of the pathological state of the patient sample, or some other
mechanism not yet identified.
6.9 Combined AC and DC Stimuli

As noted in Sects. 5.10 and 6.3 to 6.5, cardiac sensitivity to AC and to
monophasic DC pulses can differ significantly. It follows that excitation and
fibrillation thresholds for AC can be modified when a DC component is
TABLE 6.12. Log-normal percen-

tiles for fibrillating currents (front/
hind limb paths).
Percentile rank (%) Normalized IF
99 2.14
95 1.67
90 1.51
75 1.24
50 1.00
25 0.80
10 0.66
5 0.60
1 0.47
0.5 0.43
IF has been normalized by median value.

90
Endocardial Intra-myocardial
80 electrodes
pacemaker
70 electrodes 26 gauge x 1 cm
~ 60
!!-
~
c: 50
~
40
.5l1
""§ 30
OJ
12
OJ
a. 20
Epicardial disc
10 electrodes
(1.8 cm dial
5
0.090.1 0.5 1 5 10
Fibrillating current (mA rms)
FIGURE 6.17. Distribution of 50-Hz fibrillating current in patients undergoing sur-

gery for valve replacement. Straight lines on this format represent log-normal
distributions. (Data from Watson et aI., 1973.)
50",-----.------.------.------.-----,
0-
I
0-
<l: 20
E
10
5 Zero current level
vvV'v rvv IV\- ~ ---

4 \
l' I I I I I {
100 75 50 25 0
Waveform balance (% cathodal)
FIGURE 6.18. Effects of waveform bias on stimulation by 1-kHz sine wave. Isolated
rabbit hearts, 1-cm2 active electrode on left ventricle. Horizontal axis indicates
degree of DC offset. Brackets indicate ± S.D. (Adapted from Roy et aI., 1985.)
Combined AC and DC Stimuli 229
included. The effects of DC bias are illustrated in Fig. 6.18 for a sinusoidal
stimulus at 1 kHz, with a duration of either 1 or 10 cycles (Roy et aI., 1985).
For these data, stimuli were provided to isolated rabbit hearts with a l-cm
electrode attached to the left ventricle and an indifferent electrode at a
distant location. The vertical axis in Fig. 6.18 indicates the excitation thresh-
old for cardiac pacing; the horizontal axis indicates the degree of DC bias as
a percentage of the peak of the AC component. At the extremes of the
horizontal axis, stimulus current was either fully cathodic or fully anodic;
in between are intermediate cases of DC bias, with a symmetric biphasic
wave represented at the center. Thresholds for symmetric sine waves are
significantly above the purely anodic or purely cathodic waveforms, reflect-
ing the cancellation effects of current reversal (see Sects. 6.4 and 6.5).
Thresholds for the monophasic currents are considerably below those for
the balanced biphasic current. The cathodic threshold is lower than the
anodic threshold by factors of 1.8 and 1.4 for the l-cycle and 10-cycle
stimuli, respectively.
Figure 6.19 illustrates cardiac excitation thresholds for various mono-
phasic waveforms applying to the previously described experiments of Roy
100~--~------~----~--~~----~----~~
50
20
10
2~----L-----~~----L---~------~----~~
0.1 0.2 0.5 2 5 10

Current duration (ms)
FIGURE 6.19. Strength-duration curves for stimulation by rectangular pulses and

DC biased sine waves of frequency 0.5, 1, 10, 50, 500, and 1,000 kHz. Rabbit hearts
(n = 5); 1-cm2 electrode at left ventricle. Brackets indicate range of mean thresholds
for all waveforms. (Adapted from Roy et aI., 1985.)
et al. In this case, the DC bias was such that the sinusoidal current was
unidirectional. Stimuli consisted of biased sine waves with frequencies of
0.5, 1, 10, 50, 500, and 1,000kHz and with durations ranging from 0.2 to
10ms, as indicated on the horizontal axis of the figure. An additional
stimulus waveform consisted of a rectangular pulse of the indicated dura-
tion. Thresholds are expressed on the vertical axis in units of transferred
charge. The brackets shown on the curves encompass the mean excitation
thresholds over the entire set of stimulus waveforms. When measured in
charge units, thresholds among the various waveforms all followed the
strength-duration curve of a rectangular monophasic pulse. Cathodic
thresholds in Fig. 6.19 are consistently lower than anodic thresholds by an
average ratio of 1.8 ± 0.2. This polarity sensitivity difference agrees with
data reported by Thalen et al. (1975) and Cranfield et al. (1975). Similar
polarity differences have also been seen with peripheral nerve stimulation
(Sect. 4.4).
Effects of waveform bias illustrated in Figs. 6.18 and 6.19 apply to excita-
tion thresholds with waveforms oflimited duration (up to 1Oms). It is much
more complex to determine VF thresholds for biased waveforms, particu-
larly when the stimulus duration is sufficiently prolonged that mUltiple
excitations may be produced. The experiments of Knickerbocker (1973) are
valuable in assessing VF thresholds for mixed waveforms of prolonged
duration. These experiments were carried out on anesthetized dogs, with
electrodes attached from the left foreleg to the right hind leg. The stimuli
consisted of DC or 20-Hz AC administered either singly or in combination;
durations ranged from 0.2 to 2s. Figure 6.20 gives a rough summary of the
mean VF thresholds. The mix of AC and DC components is plotted on
Cartesian coordinates, with the points on the axes corresponding to pure
DC (vertical axis) or pure AC (horizontal axis). The single-component
thresholds correspond to the data of Fig. 6.11.
A biased AC waveform will be purely monophasic when the DC compo-
nent is at least lilAc. where lAC is the rms AC current. This relationship is
represented by the diagonal line in Fig. 6.20, above which the waveforms
are purely monophasic and below which they are biphasic with varying
degrees of asymmetry. At all the durations studied, the pure AC thresholds
were lower than the pure DC thresholds. The threshold contours for the
mixed waveforms form patterns that depend not only on the excitatory
power of the stimuli but also on the associated excitation rate.
6.10 Electrodes and Current Density

Electrode Area
For a given current, the average current density beneath an electrode
increases inversely with the electrode area. At the same time, the smaller
Electrodes and Current Density 231
500
I
/
Monophasic /
waveforms I
400
/
I Biphasic
I waveforms
~ 300 I
....c: /
/
Q)
c:
o
c. /
E
o
~ 200
o
100
o 100 200 300 400

20 Hz component (mA rms)
FIGURE 6.20. VF threshold contours for combined 20-Hz AC and DC currents

of prolonged durations. Limb-to-limb electrodes with dogs. (Data from
Knickerbocker, 1973.)
electrode results in a reduced amount of tissue that may be brought to

excitation. A unified theory does not exist to predict the effect of electrode
area on VF thresholds, although there is considerable empirical data on this
issue (see Sect. 5.10).
Figure 6.21 summarizes VF thresholds for 60-Hz currents from electrodes
in contact with the heart (adapted from Roy, 1980). The individual points
are average and minimum thresholds from a variety of published data
applying to 60-Hz stimuli of at least 2s duration. Considering the diversity
of experimental conditions and tested species, the consistency of the data is
remarkable. The vertical axis of Fig. 6.21 expresses the ratio of threshold
current to electrode contact area. Threshold current density determined
this way drops as the electrode area is increased to about 5 cm2, after which
very little further drop is encountered. The average thresholds converge to
about 0.5 mA/cm2 for large electrodes, a value that is supported by data
from both human and dog heart experiments (Starmer and Whalen, 1973).
G3
N
1000 o
n
I'l
C\~J-'
8-
s;.
(")
E (/)
("J)
~ ...........- average values ::l
100 --().-. minimum ~.
~ •
·iii 0.. 0 3:
c:: ". q
Q)
"0 ~""'~ ....o
C '0. 0 • tTi
~
ct-. • CD
.... ......... "'0
::J 2t
(S.
U o -----0 0
10 -'0 ___________ 0 e:..
Q)
(/)
Ol
o 0 0----_____ 0- §'.
~
Q) o
S
I'l
~ ....
o·
::l
1
10-3 10-2 10- 1 10 102 103
electrode area (cm 2)
FIGURE 6.21. Average and minimum current density for VF thresholds with 60-Hz stimulating current of at
least 2-s duration. Data points apply to a variety of different experiments. (Adapted from Roy, 1980.)
Minimum current density thresholds were typically about a factor of 2

below the averages for the various experiments evaluated by Roy. Using
the log-normal model of Table 6.12, it is estimated that the "minimum"
values observed by Roy might represent 1 percentile values of a statistical
distribution. Additional current density thresholds (Ruiz et aI., 1985) are
consistent with Fig. 6.21.
Roy et aI. (1986) suggest the following electrode-size relationship for VF
thresholds based on experiments with dogs in which a disk electrode was
applied to the epicardial surface of the right ventricle:
(6.8)
where c = 0 for d ::5 2mm, C = 1.5 for 4::5 d::5 15.9mm, C = 2.0 for d ~
15.9mm, Iris the VF threshold, d is the electrode diameter area, and K is a
scaling constant. Roy's relationship states that, below a lower critical size
(2-mm diam. = 0.03-cm2 area), the VF threshold current is directly propor-
tional to electrode area (square of diameter). In between those critical sizes,
current thresholds are proportional to the 1.5 power of electrode diameter
(0.75 power of area). Analogous relationships were reported for excitation
thresholds (Lindermans and Van der Gon, 1978), in which threshold cur-
rent was found to be independent of electrode area for diameters below
0.4 mm and proportional to the 1.5 power of diameter between 0.8 mm and
1.8cm. Lindermans and colleagues justified their experimental results with
theoretical calculations of current density beneath a contact electrode and
the assumption of a /iminallength of 0.3 mm for myocardial excitation; i.e.,
a minimum size of depolarized tissue necessary to initiate excitation. That
liminal length is consistent with the value suggested in studies of individual
Purkinje fibers (Fozzard and Schoenberg, 1972).
The relationships between electrode area and threshold current density
will be different for excitation and fibrillation. With isolated guinea pig
hearts stimulated by DC pulses, Weirich and colleagues (1985) showed that
VF thresholds expressed in units of current are lowered if the electrode size
is increased, but the opposite effect is seen with thresholds of excitation (see
Fig. 5.20). Current thresholds of excitation go up with electrode size
because a minimum current density is required for stimulation. To attain
fibrillation, however, it is not sufficient merely to stimulate a few cells-
rather, a critical mass of heart muscle must be excited to cause the degree
of nonhomogeneity that is required for fibrillation.
The observations of Weirich et aI. are supported by Table 6.13, which lists
current and current density thresholds for excitation and ventricular fibril-
lation of dog hearts by 2-ms DC pulses (from Chen et aI. 1975). When
expressed in units of current magnitude, thresholds of excitation drop
steadily as the electrode size is made smaller, whereas fibrillation thresholds
do not differ greatly over the range of areas tested and actually rise at the
smallest tested area. The trends seen in Table 6.13 demonstrate that, as far
as excitation is concerned, the heart is more easily excited by a given
TABLE 6.13. Current density thresholds for stimulation of dog hearts by DC

pulses (t = 2ms).
Electrode area VF threshold EX. threshold VF/Ex.
(cm') (mA/cm') (rnA) (mA/cm') (rnA) ratio
0.005 5,000 25.0 14.0 om 357
0.08 275 22.0 2.0 0.16 140
0.28 104 29.1 1.8 0.50 58
Pulse duration = 2ms. In vivo stimulation of dog hearts (n = 20). Active electrode
implanted into myocardium.
Source: Data from Chen et al. (1975).
amount of current into a small electrode. With fibrillation, on the other

hand, there exists a trade-off between the excitation threshold and the
critical mass of excited tissue required to support fibrillation (Antoni, 1979).
Table 6.14 gives additional data applying to the pacing of human hearts by
2-ms pulses (from Furman et aI., 1967). The current densities listed may be
compared with the data for dog hearts given in Table 6.13. At an electrode
area of 0.28 cm2 , the thresholds in the experiments with dog hearts are given
as 1.8mA/cm2, and in the experiments with human hearts, as 3.0mA/cm2•
Considering the different experimental conditions and tested species, these
data are reasonably consistent. The data of Tables 6.13 and 6.14 provide
further support for the proposition that experimental thresholds obtained
from dogs provide conservative estimates for human applications.
Current Density
The foregoing discussion has dealt with current density thresholds inferred
from experiments in which an electrode was held in contact with cardiac
tissue. Current may also be introduced less directly, such as with electrodes
TABLE 6.14. Current density thresh-

olds for stimulation of human hearts
by DC pulses (t = 2ms).
Electrode area Ex. Threshold
(cm') (mA!cm') (rnA)
0.12 3.27 0.39
0.28 3.00 0.84
0.49 3.14 1.54
0.87 2.30 2.00
In vivo stimulation of human hearts.

Active catheter electrode inserted into
right ventricular apex.
Source: Data from Furman et al. (1967).
applied to the body at locations distant from the heart, or with current
induced by exposure to time-varying magnetic fields. In these cases, current
may be applied more or less uniformly to the whole heart rather than in a
concentrated locus beneath a contact electrode.
Stimulation thresholds for whole-heart exposure are traditionally evalu-
ated in terms of current density. Nevertheless, the electric field within the
biological medium may be a more relevant parameter, as suggested by
Starmer and Whalen (1973), who observed that VF thresholds are better
correlated with spatial potential differences than with current density. This
observation is consistent with theoretical and experimental studies with
nerve fibers (Sect. 4.4), which show that the relevant excitation parameter
is the electric field longitudinal to the fiber rather than the current density
per se. The electric field (E) and the current density (J) are simply related
by the conductivity (a) of the medium by J = Ea. Although current density
is often cited as an experimental parameter, conductivity is not always given
in published descriptions of experimental data. For that reason, the electric
field may be difficult to ascertain.
Current density is most commonly determined by dividing stimulus
current by the area of the electrode, as with the data in Fig. 6.21 and Tables
6.13 and 6.14. Current density determined this way will correctly give the
average value beneath the electrode, but may understate the maximum
density because current tends to concentrate at the edges of a contact
electrode (see Fig. 2.9).
Considering the possibility of nonuniform current distribution at the
interface between an electrode and biological tissue, the values plotted in
Fig. 6.21 probably provide conservative estimates of the maximum current
density. With this caveat in mind, the current density of 0.5 mA/cm2 shown
for the largest exlectrodes in Fig. 6.21 estimates the average current density
threshold for VF by whole-heart exposure to prolonged (>2s) 60-Hz
currents; 0.25 mA/cm is a more conservative estimate of a 1 percentile
experimental value. Table 6.4 indicates that excitation thresholds are about
40% of the 60-Hz fibrillation value for prolonged (t > 2s) 60-Hz and also
for monophasic pulse stimulation (t = 10ms). Applying this factor to the VF
current density results in an estimated average threshold of 0.2mA/cm2 for
excitation by either prolonged 60-Hz or single DC pulses and a 1 percentile
threshold of 0.1 mA!cm2. That minimum value is quite close to the value of
0.12 mA/cm2 calculated as a median threshold for excitation oflarge (20,um)
peripheral nerves (Table 4.2). The current densities cited above have been
obtained in experiments with healthy animals. It is not known whether
these values are representative of the human heart when it is in a pathologi-
cal state (see Sect. 6.8).
The current density required for defibrillation is much greater than that
required for excitation, or fibrillation. To achieve defibrillation, one must
ensure that a critical mass of cardiac tissue distant from a defibrillation
electrode is uniformly excited. To achieve defibrillation, it is estimated that
80% of the heart must be subject to a current density of at least 35 mA/cm2

(Sepulveda et aI., 1990).
Locus and Direction of Stimulus Current

Thresholds of fibrillation will vary with the location of electrodes, whether
they are contacting the heart or are applied externally. Table 6.15 (adapted
from Roy et aI., 1987) lists VF thresholds for a l-cm2electrode applied to
five different regions of a pig's heart. An indifferent flexible electrode (15 X
56cm) was wrapped around the back and sides ofthe animal. Seven animals
were tested using a 60-Hz stimulus of 5s duration. The data in the table are
listed in the order of increasing average current density; the stimulated
regions are identified in Fig. 6.22. The thresholds span a range of 2.2: 1 from
the least to the most sensitive area. The most sensitive area is the apex;
regions of the right ventricle have higher thresholds than those on the left
ventricle. At the heart apex, the average and minimum thresholds listed in
Table 6.15 compare favorably with the VF thresholds noted in the previous
section if electrode area is properly considered. The authors pointed out
that their results are consistent with the concept that a critical mass of
cardiac muscle is required for sustained fibrillation-the apex, having the
greatest muscle thickness, is associated with the lowest threshold. Excita-
tion via external electrodes would be expected to have the lowest thresh-
olds if the electrodes were oriented so as to provide the greatest exposure
to the apex of the heart.
Table 6.16 indicates the variation of VF thresholds with respect to the
pathway of 50/60-Hz currents delivered for long duration (t > 3s) through
electrodes placed at various locations on the body. The tabulated data,
derived from the references listed at the bottom of the table, indicate
relative thresholds with a single column. In column 1, for example, the value
If = 1.56 for a hand-to-hand path indicates a finding that 56% more current
was needed to induce VF than with a right hand-to-foot path having If = 1.0.
A listing of 1.0 in one column does not necessarily indicate the same VF
TABLE 6.15. VF thresholds for 60-Hz currents

applied through a 1-cm2 active electrode (data
listed in order of increasing average threshold).
VF threshold (rnA)
Electrode position n min avg rms
1 16 0.66 0.98 0.23
2 14 1.06 1.42 0.52
3 12 1.19 1.82 0.42
4 15 1.09 1.83 0.51
5 14 1.35 2.17 0.71
Source: Data from Roy et al. (1987).

Aorta
Left
Vena cava ----+.

Right
auricle
----1---
Right
ventricle 4--+-Left ventricle
Apex
FIGURE 6.22. Diagram of pig's heart showing position of 1-cm2 electrodes numbered
in order of increasing average VF thresholds. (From Roy et a1., 1987.)
threshold as a listing of 1.0 in another column. The If data in the first three
columns apply to experiments with animals (Ferris et ai., 1936; Geddes et
ai., 1973; Roy et ai., 1986); the values in the last column are criteria recom-
mended in Europe by the IEC for evaluation of electrical hazards (IEC,
1982).
The relative current thresholds listed in the table are a result of several
factors, including the following:
1. Current density. The current density in the vicinity of the heart will
vary with the placement of external electrodes. In experiments with dogs,
it has been reported that 7% of the applied current flows through the
heart with a forelimb-to-hind limb path, whereas only 3% is intercepted by
the heart for current applied across the forelimbs (Kouwenhoven et ai.,
1932).
2. Excitable fiber orientation. We have seen in Chapter 4 that current
aligned with the long axis of an excitable fiber is much more effective in
producing excitation than current flow in a transverse direction.
3. Exposure of sensitive region. As suggested by Table 6.15, we may
expect a 2: 1 difference in sensitivity for stimulation to different regions of
TABLE 6.16. Relative VF sensitivity with respect to path way of 50/60-Hz

current (large area, t;;:: 3s).
Electrode positions Relative IF
A. Hand-to-feet paths
Left hand-right foot (left fore-right hind) 1.0 1.0
Left hand-left foot (left fore-left hind) 1.13 1.0
Right hand-left foot (right fore-left hind) 1.0 1.25
B. Hand-to-hand paths
Right hand-left hand (right fore-left fore) 1.56 2.3 2.5
C. Foot-to-foot paths
Left foot-right foot (left hind-right hind) >50
D. Misc. hand paths
Right hand-back (right fore-back) 3.33
Left hand-back (left fore-back) 1.43
Right hand-chest (right fore-chest) 0.77
Left hand-chest (left fore-chest) 0.67
Hand-seat 1.43
E. Other paths
Left chest-right chest 1.04 1.4
Left foot-head 1.20
Chest-back 1.0
References: (1) Ferris et al. (1936); (2) Geddes et al. (1973); (3) Roy et al. (1986); (4)
IEC (1982).
the heart. Variation in the placement of external electrodes may provide

more or less exposure to sensitive regions of the heart.
4. Anisotropic impedance properties. The conductivity of muscle tissue is
greatest when the current is aligned with the direction of the muscle fibers
(Chilbert et aI., 1983). The conductivity parallel and perpendicular to the
fibers of ventricular muscle demonstrate a ratio of about 2: 1 (Ruch et aI.,
1963). Accordingly, the direction of current would affect its distribution
within the heart, and the mass of excited tissue.
5. Polarity of DC stimulation. We have seen in Chapter 4 that an excit-
able fiber will be depolarized most effectively at end structures that are
oriented toward the cathode of a current source.
While the factors mentioned above may all play a role in the heart's
sensitivity to electrode placement and the current path, it is not possible to
separate out their relative importance on the basis of available data. How-
ever, a number of generalizations can be made, as noted below.
Longitudinal versus Transverse Current

Current flow along the long axis of the body (longitudinal flow) is associ-
ated with significantly lower thresholds than with current flowing in an
orthogonal (transverse) direction. Hand-to-foot pathways are observed to

have significantly lower thresholds than hand-to-hand paths-the ratio of
thresholds for the two orientations is about 2: 1. Table 6.10 lists average
thresholds for longitudinal paths, attributed to subjects of different body
weights.
Foot-to-Foot Paths
Available data indicate that current applied from one foot to another is
very unlikely to result in VF. Ferris et ai. (1936) found that current applied
from foot to foot failed to produce VF at a current level that was 50 times
greater than the VF threshold for hand-to-foot contacts. The relatively high
thresholds for foot-to-foot contact results because relatively little current
reaches the heart when it is introduced through the legs or hind limbs.
Chest Electrodes
When one of the electrodes is placed on the chest, the VF threshold de-
pends critically on its precise placement with respect to anatomical features
of the heart. The lowest threshold for a chest electrode is obtained when it
is placed directly over the apex of the heart (Geddes et aI., 1973)-a finding
consistent with open-heart tests (Roy et aI., 1987). With large-area
(300cm2) electrodes on the chest, VF thresholds have been found to be 1.4
times lower when the current path was from the chest to the back of the
animal than with a transverse (side-to side) orientation (Roy et aI., 1986).
Thresholds of stimulation with a large area chest electrode were measured
typically between 40 and 70mA with a minimum value of 20mA in tests of
humans (Zoll et aI., 1985). The 60-Hz fibrillation thresholds for 200-mm2
chest electrodes on dogs averaged 68mA (Roy et aI., 1986).
DC Stimulation
As with AC stimulation, a longitudinal orientation of DC is associated with
significantly lower thresholds than with transverse current flow. Fatal acci-
dents in Europe involving DC exposure have been reported to occur only
with a longitudinal current path (Biegelmeier, 1987). Additionally, the
polarity of the DC electrodes significantly affects VF thresholds, with posi-
tive foot electrodes being the most hazardous. In early experiments (Ferris
et aI., 1936), it was found that the VF threshold was about 36% lower when
the feet were at a positive potential relative to a hand electrode as com-
pared with the opposite polarity. Biegelmeier (1987) reports that VF
thresholds with positive foot and negative hand electrodes are about one-
half the thresholds pertaining to the opposite direction of current flow.
7
Sensory Responses to Electrical
Stimulation
7.1 Introduction
Sensory sensitivity to electrical stimulation depends on a host of factors
associated with the stimulus waveform, its method of delivery, and subjec-
tive variables. In most situations involving electrical safety or acceptability,
current is applied to the body by cutaneous electrodes. There are also
practical applications in which electric current may be applied subcutane-
ously or induced internally by external electromagnetic fields. Although
the emphasis in this chapter is on electrocutaneous stimulation, many of the
principles discussed may be applied to other modes of stimulation. The
reader is directed to Chapter 9 for additional discussion of peripheral nerve
stimulation by time-varying magnetic field effects or by induced shock
within intense electric field environments. In addition to sensory effects
described in this chapter, stimulation by electric current and electromag-
netic fields can also elicit visual and auditory sensations. These will be
treated in Sect. 9.8.
7.2 Mechanisms of Electrical Transduction

Current of a fraction of a microampere can be detected when the finger is
gently drawn across a surface charged with small AC potentials (Grimnes,
1983b). Such levels are roughly 100 times less than commonly tested electri-
cal thresholds. Detection of such small current results from electrome-
chanical forces arising from electrostatic compression across the stratum
corneum (the outermost layer of dead skin cells). As analyzed by Grimnes,
the electrostatic force K is
(7.1)
where A is the contact area, e is the dielectric constant of the corneum, d is

its thickness, and v is the instantaneous voltage. The compression of the

240
Mechanisms of Electrical Transduction 241
corneum would not normally be sensed. But when tlJe skin is moved along
the charged surface, there is a vibratory frictional force on the finger that is
maximized on each half-cycle of the alternating voltage. This vibrational
force stimulates mechanoreceptors and is responsible for the detection
of microampere currents. Grimnes estimates that the minimum voltage
contributing to a detectable vibration is about 1.5V at 50Hz.
The detection of microampere currents through mechanical vibration is,
for most purposes, of passing interest, although it may be important for a
researcher to know about it when designing perception tests. Of greater
significance is the mode of detection when the current level is raised to
roughly 0.1 mA or above. At that point, perception can be initiated by the
electrical excitation of neural structures, according to the mechanisms
discussed in Chapters 3 and 4.
Exactly what is excited with electro cutaneous stimulation, and what is
the specific site of initiation? At the lowest levels of stimulation, it is likely
that peripheral structures are involved, because these are closest to the
surface electrode. Among fiber classes, the larger-diameter myelinated
fibers have the lowest electrical thresholds, and circumstantial evidence
presented in this chapter points to the involvement of one or another class
of mechanoreceptor. The precise site of cutaneous electrical stimulation is
unknown; whether stimulation is initiated at the axon proper, at the site of
the generator potential of sensory receptors, or along free nerve endings
has not been demonstrated. Some evidence, however, exists, as noted in
Chapter 4, that the site of initiation is near neural end structures, including
receptors or free nerve endings. Electrocutaneous perception is a local
phenomenon; subjects typically report sensation occurring locally at the
electrode site rather than remotely as might be supposed if the excitation
occurred on the axons of deeper-lying nerves. It is only when the current is
raised substantially above the perception level that distributed sensations
are felt.
If the current is raised sufficiently above the threshold of perception,
excitation of unmyelinated nociceptors becomes possible. Because of their
higher electrical thresholds and generally deeper sites, these structures are
not likely to be involved at perception threshold levels. At still higher
current levels (some tens of milliamperes for long-duration stimuli),
thermal detection due to tissue heating becomes possible. Neuroelectric
thresholds may exceed thermal thresholds if the waveform of the electric
current is inefficient for electrical stimulation, such as with sinusoidal
currents of very high frequency (> 105 Hz).
The electrical stimulus is nonspecific as to the receptor class that might be
stimulated. Nevertheless, there is some degree of selectivity on the basis of
electrical properties-thresholds tend to be lower as the neuron is closer to
the corneum, as the fiber diameter is increased, and as its orientation is
aligned with the internally generated electric field (i.e., direction of current
flow). At suprathreshold levels, an array of different fiber types will be
242 7. Sensory Responses to Electrical Stimulation
stimulated, including deeper-lying structures such as small-diameter

nociceptors and motor neurons.
One might question whether the electrical stimulus would feel like some
natural modality corresponding to the receptor type that is being excited.
Research, subjects have preferentially selected punctuate pressure descrip-
tors from a list of 10 choices for capacitor discharge stimulation applied
cutaneously to the forearm, as shown in Fig. 7.1 (Reilly and Larkin, 1985a).
Similar results have been obtained for stimulation by 300-ms square-wave
pulses: At levels just above perception, subjects used descriptors of distrib-
uted pressure; but as the stimulus level was increased, they used punctuate
pressure descriptors (Tashiro and Higashiyama, 1981).
Electrical stimulation by pulse trains evokes sensations closer to natural
stimuli when the nerve is stimulated directly by a percutaneous microelec-
trode. In the experiments ofVallbo and colleagues (1984), a micro electrode
was inserted above the elbow into the median nerve, which has a receptive
field in the hand. When stimulated by a train of pulses, subjects reported
localized sensations in the hand akin to steady pressure, transient pressure,
or vibration, depending on the particular fiber being stimulated and the
electrical stimulation parameters. The authors reported that stimulation of
a single nerve fiber could elicit perceptible sensations for some classes of
fibers.
"Needle prick 0 Touch

6400 pf series • Carpet spark • Pressure
200 pf series
40
o L--L__ L-~~~-L~~~~____~____L--L__L-L-~~
0.2 0.3 0.4 0.6 0.8 1.0 2.0 3.0 5.0

Voltage (kV)
FIGURE 7.1. Qualitative ratings by subjects receiving suprathreshold capacitor dis-

charges. (From Reilly and Larkin, 1984.)
Mechanisms of Electrical Transduction 243
1.0
0.8
c
0
.~ t = 100 ms
0.6
Qi
"0
15 0.08
>-
:E:
:0 0.4
III
~
c..
0.2
0.0
0.1 0.2 0.4 0.6 0.8 1.0 2.0
Stimulus current (rnA)
FIGURE 7.2. Psychometric functions for detection of square-wave pulses applied

cutaneously; O.5-cm2 saline-treated electrode on forearm. (Data from Rollman,
1974.)
Some understanding of the neural pathways for electrical stimulation can

be inferred from experiments in which A or C fibers are selectively blocked
in human subjects (Torebjork and Hallin, 1973). When mainly A-fiber
response is present, weak electric shocks are felt as tactile sensations, and
strong shocks are perceived as a short, sharp blow without prolonged pain.
When A-fiber response is blocked, there is an impaired discrimination of
weak stimuli, and strong stimuli evoke painful sensations because of C-fiber
activity.
The electrocutaneous stimulus is a very unnatural one in terms of its
spatial distribution of stimulation, lack of receptor specificity, abnormal
neural recruitment properties, abnormal temporal patterns of excited
action potentials, and concurrent recruitment of sensory and motor neu-
rons. These factors all contribute to a sensory quality that is unlike any for
which nature has equipped us and probably contribute to the highly
aversive quality that most people attribute to electrical stimulation only
somewhat above their perception.
The relatively small dynamic range of electrical stimulation is reflected in
extremely steep psychometric functions (i.e., the functional relationship
between the stimulus level and the probability of detection). Figure 7.2
illustrates psychometric functions applying to a single individual detecting

cutaneously applied square-wave pulses of various durations (Rollman,
1974). The coefficient of variation (standard deviation to mean ratio) was
typically 0.08, a value that is significantly smaller than that observed for
other sensory modalities. Because of the steep psychometric functions, a
change of stimulus magnitude of only 20% typically elevates the detection
probability from 10% to 90% for each curve shown in Fig. 7.2. Figure 7.2
also demonstrates that thresholds are inversely related to the duration of
simulation, as discussed below.
7.3 Perception of Transient Monophasic Currents

Strength-Duration Relationships
Sensory sensitivity varies greatly with stimulus parameters, including wave-
shape, duration, repetition pattern, polarity, and whether the stimulus
is monophasic or biphasic. The relationship between the intensity of a
monophasic current and the threshold of reaction has been the most sys-
tematically studied aspect of waveform dependency. The S-D relationships
for neural excitation have a long history of exploration, beginning with the
early studies of Weiss (1901) and Lapicque (1907). In Chapter 4, the early
formulations are described in terms of an exponential [Eq. (4.15)] or hyper-
bolic [Eq. (4.17)] relationship. Although their mathematical expressions are
different, the two formulas express similar results (Fig. 4.3).
Perception thresholds converge to a minimum current (Io) as the current
duration is made very long and to a minimum charge value (Qo) as it
is made very short. The minimum current has traditionally been called
rheobase. Whether the exponential or hyperbolic formulations are used, the
S-D curve may be described in terms of two parameters: the equivalent
S-D time constant (ie ) and the minimum (rheobase) threshold. The time
constant is a measure of the temporal responsiveness of the excitable
system. In both exponential and parabolic formulations, ie may be obtained
by the ratio Qr//o [Eq. (4.20)].
The S-D relationship applies to both perception and suprathreshold
values. Figure 7.3 illustrates neurosensory and neuromotor S-D curves
from the experiments of Alon and colleagues (1983). The upper set of
curves represents thresholds as the peak current of a monophasic square-
wave stimulus. The lower set shows stimulus charge (product of amplitude
and duration). The broken line indicates the exponential expressions [Eqs.
(4.6) and (4.9)], adjusted such that the values Qo and 10 correspond to the
experimental perception threshold values. With this adjustment, the experi-
mental and analytic curves agree remarkably well.
Table 7.1 summarizes S-D data from various sensory stimulation experi-
ments. The column headed ie refers to the S-D time constant and the
1000
Linear
model
\ ...........
,......
~ 100
.s
'E
~
u
::J Tolerance
u Pain
(5
.s::.
...
Ul
Q)
.s::.
I-
10
Motor
(a) Curent thresholds
100
Tolerance
Pain
U 10 "
..::; ..~~....."
./
Q)
e'
co
.s::. Perception
u
:g
0
.s::.
Ul
~ /,,'
.s::. ......./.
I-
..............
...••.....~
Linear
model
(b) Charge thresholds
0.1
1 10 100 1000 10000
Stimulation duration (Ils)
FIGURE 7.3. Strength-duration curves for sensory and motor reactions to square-
wave pulses; forearm stimulation, 4-cm2 electrode. (Data from Alon et aI., 1983.)
245
TABLE 7.1. Strength-duration data for nerve stimulation.
T, 1., tv
Q" +>-
Subject N Stimulus locus Electrode Reaction (us) (mA) (,lie) Reference 0\
A. Human subjects
Human 6 Forearm 4-cm 2 carbon sponge Perception 138 3.5 0.48 Alon et al. (1983) ;-J
Human 6 Motor 127 8.5 1.08 Alon et al. (1983) [fJ
Human 6 Pain 70 24.5 1.72 Alon et al. (1983) ~
i:l
Human 6 Tolerance 142 33.8 4.80 Alon et al. (1983) f/>
0
Human 4 2-mmdiam. Pain 760 0.5 0.38 Notermans (1966)
'<
"'
Human 3 Finger (Cap. discharge) 1.1-mm diam. Perception 200-900 0.12 Larkin and Reilly (1984)
~
Human Ulnar nerve. forearm 0.5-cm' elect. paste Perception 480 0.25 0.12 Rollman (1975) ~
Human Finger pad OA-cm 2 Perception 217 0.69 0.15 Hahn (1958) "0
'"
Human Forehead; forearm O.lmm' 0
Perception 189 1.8 0.34 Girvin et al. (1982) i:l
Human 6 Ulnar nerve AP <260 2.0 <0.52 Heckmann (1972) ~
'"
Human 6 Ulnar nerve Muscle twitch 275 3.2 0.88 Heckmann (1972) '"
Human Denervated muscle Muscle twitch 8333 6.0 50 Heckmann (1972) 8"
Human Enervated muscle Muscle twitch 30-700 Harris (1971) tTI
(D
Human Denervated muscle Muscle twitch >4300 Harris (1971)
~
Human 2 Forearm 3.9-cm' elect. paste Muscle twitch 140 9.9 1.39 Crago et al. (1974) ",.
Human 12 Foot 8mm dia. Perception 295 1.1 0.32 Friedli & Meyer (1984)
"'
Human 8 Finger 8mm dia. Perception 383 0.5 0.19 Friedli & Meyer (1984)
r=..
[fJ
....
B. Animal. in-vitro nerve preparation S·
Rat. cat 12 Tibialis ant. musc. lOmm, intra. muse. Muscle twitch 85 5.2 0.44 Crago et al. (1974) E-
Rat. cat 12 Peroneal nerve 0.021 Crago et al. (1974) ~
AP 55 0.38 ....
Rabbit 2 Tibial nerve O.l-mm diam., in vivo AP 80-100 Reilly et al. (1985) o·
i:l
(4-13,um diam.) (subcutaneous)
Cat Ap fiber In virto AP 30 Li and Bak (1976)
Cat A, fiber In vitro AP 650 Li and Bak (1974)
Toad Myelinated nerve In vitro AP 148 Tasaki and Sato (1951)
Dog 4 Back Magnetic coil Muscle twitch 148 Bourland et al. (1990)
Dog 4 Chest 2.5cm dia. Inspiration 245 49.4 Voorhees et al. (1992)
C. Theoretical model
SENN model Myelinated nerve Point elect. AP 92-128 Reilly and Bauer (1987)
SENNmodel Myelinated nerve Uniform field AP 120 Reilly and Bauer (1987)
* Information not available.

-Information not applicable for table summary.
Perception of Transient Monophasic Currents 247
columns headed 10 and Qo to the minimum current and charge thresholds. In

most cases, the tabulated value of Te has been determined by Qrllo. In a few
cases, where the minimum thresholds were not available, Te was determined
by chronaxy/O.693 [see Eq. (4.19b)] or by a simple curve fit of Eq. (4.6). The
table entries are subdivided according to categories of human perception,
animal or nerve preparations and a theoretical model for myelinated nerve.
The "human" category includes both neurosensory and neuromuscular
responses. The values of Qo and 10 depend on a number of factors such as
electrode size, the locus of stimulation, the reaction threshold being tested,
and whether the current is supplied transcutaneously or subcutaneously.
These factors are discussed in detail in later sections of this chapter. The
ratio Qallo is much less sensitive to these factors.
Empirical time constants for neuromuscular reactions do not differ sig-
nificantly from neurosensory values, although the absolute threshold for
motor responses is typically higher. Although motor neurons typically have
larger diameters than sensory neurons, and hence would be expected to
have lower electrical thresholds, the fact that they are situated in deeper
strata below the skin accounts for their higher thresholds. When the muscle
is denervated, motor stimulation takes place by direct excitation of muscle
fibers rather than motor neurons (see Sect. 8.3). In this case, motor thresh-
olds rise significantly, and the time constants increase dramatically
(Bauwens, 1971; Harris, 1971). The fact that the parameters of the S-D
curves vary with conditions of neuropathy (Friedli and Meyer, 1984) sug-
gests a potential diagnostic tool.
Although sensory time constants encompass a range similar to that noted
for neural stimulation (Parts Band C of Table 7.1), there is a greater
representation toward the larger values with sensory data. The geometric
mean of Te values given in Table 7.1 for human perception is about 270flS;
when the data of Parts Band C are included, the geometric mean is closer
to 170flS. The myelinated nerve SENN model indicates a range of Te de-
pending on the distance between a point electrode and the nerve fiber. For
uniform E-field excitation, the model indicates Te = 121flS; in comparison,
a value of 140flS was obtained under experimental conditions closely ap-
proximating uniform field stimulation, in which the internal E-field was
induced by magnetic field exposure over a large area of the torso (Schaefer
et aI., 1991).
Experimental S-D time constants are not simply a property of the excit-
able tissue, but depend on the method of stimulation as well. For instance,
time constants have been found to increase monotonically with electrode
size in nerve and neuromuscular preparations (Pfeiffer, 1968; Davis, 1923)
as well as in cardiac tissue (see Sect. 6.2); this dependency is consistent with
theoretical arguments, although the effect is expected to be much stronger
in muscle than in nerve tissue (Jack et aI., 1983).
For the data described in Table 7.1, absolute perception thresholds for
short-duration stimuli and small electrodes converge to a constant charge
value of 0.1,uC, and for long-duration stimuli to a constant amplitude value

of about 0.25 rnA. These values must be qualified by a variety of factors
affecting sensitivity such as body locus, electrode size, skin temperature,
tactile masking, electrode contact, and stimulus waveform features. Addi-
tionally, individual differences in sensitivity will result in a distribution of
threshold values.
Capacitor Discharges l
Exposure to transient electric shock is a common occurrence. We have all
experienced shocks when we walked across a carpet on a dry day and then
touched a grounded object. In such cases, our body acts as a capacitor that
stores electric charges at levels of several thousand volts. Then, when we
come sufficiently close to a grounded object, the stored charge is suddenly
discharged at some discrete body location through a spark that may be felt,
seen, and heard. The peak current of a carpet spark can be very large-
typically more than an ampere-a level that could be lethal if sustained.
Fortunately, the event is very brief, in the submicrosecond range. As a
result, the shock is well below lethal intensity, but nonetheless can be
annoying to many people.
Capacitor discharges may be used beneficially in biomedical applications.
One research application takes advantage of the fact that a capacitor dis-
charge has a sudden current onset, but a gradual decay (Accornero et aI,
1977). This produces a stimulus that is effective on its leading edge, while
avoiding the possibility of excitation on the current break as with a square-
wave stimulus.
Various cutaneous sensations are possible with capacitor discharges.
Near threshold, the sensations commonly resemble touch or mild pinprick.
Above threshold, the pinprick may be followed by a burning sensation akin
to the delayed pain related to C-fiber activity. Bishop (1943) reported other
sense qualities as well and suggested that a punctuate "spark" from a
capacitor could be used selectively on the skin to excite discrete neural
channels. The capacitor discharge stimulus is monophasic (current flows in
only one direction) and can be made brief in relation to the depolarization
process in neural tissue. In these two respects, capacitor discharges re-
semble the short-duration, constant-current pulses often used in sensory
and electrophysiological research.
In two other respects, capacitor discharge stimuli may differ from the
current pulses commonly in use. First, the discharge waveform can have
both a very high initial voltage and a very high current peak, even at
threshold levels of stimulation. Furthermore, while constant-current stimu-
1Portions of this section have been adapted from Larkin and Reilly (1984) and
Reilly and Larkin (1984).
lation requires an electrode in good contact with the skin, a high-voltage

capacitor discharge can occur without contact if the electrode is simply
brought close enough to sustain an electric arc between it and the skin.
These characteristics are also present in electrostatic "carpet sparks" of the
sort familiar to all who live in dry, upholstered environments (Chakravarti
& Pontrelli, 1976).
The capacitor discharge stimulus is typically associated with a high-peak
current density at a relatively high voltage. The electrical properties of the
skin are much less significant with high-voltage capacitor discharge stimuli
than with traditional constant-voltage or constant-current devices because
the skin impedance breaks down in response to high-voltage and high-
current densities of a capacitor discharge (see Sect. 2.5).
When a capacitance (C) discharges through a resistance (R), both current
and voltage have simple exponential waveforms given by:
I(t) = Ioe- t/ T
(7.2)
V(t) = 10(1 - e- t/ T
) (7.3)
where Vo is the initial voltage on the capacitor, 10 = VoiR, and i is the
discharge time constant given by the product RC. The amount of stored
charge that may be released in the stimulus is
Q = evo (7.4)
Although Eqs. (7.2) and (7.3) represent an ideal case, discharges to biologi-
cal materials, including human skin, exhibit additional complexities such as
nonlinear impedance (see Chapter 2). Despite these complexities, the ca-
pacitor discharge waveform is typically very close to Eqs. (7.2) and (7.3), so
that an exponential approximation is sufficient. While the time course of the
discharge current may be expected to follow the exponential form [Eq.
(7.2)], discharges to intact skin are highly dependent on the initial voltage as
well as the capacitance, as indicated by Figs. 2.29 and 2.35.
We have systematically explored the sensory and biophysical aspects of
capacitor stimuli. As an example of this work, Fig. 7.4 illustrates mean
perception thresholds for anodic capacitor discharges (Reilly and Larkin,
1987). Thresholds are shown for delivery of the stimulus by: (a) the finger
tapping an energized electrode; (b) discharge to a large (1.27-cm-diameter)
circular electrode held against the fingertip; (c) discharge to a small (0.11-
cm-diameter) electrode held against the fingertip; and (d) discharge to a
needle piercing the corneum of the forearm. The curve shapes and their
relative displacements are based on intensive study of six expert subjects
(Reilly and Larkin, 1985a), but their absolute positions have been slightly
adjusted along the vertical axis to reflect measurements with a sample of
124 subjects (see Sect. 7.10). In Fig. 7.4, threshold charge at 100pF is 0.25.uC
for procedure (a); 0.19.uC for procedure (b); O.l1.uC for procedure (c); and
0.07.uC for procedure (d).
4.0 r---.....-----,r---r-..--.----..--r---r--r.....
(a) Tapped electrode

(tap force = 20 dB)
Large contact electrode
1.0 (dia. = 1.27 cm)
(c) Small electrode

>
~
(dia. = 0.11 cm)
QJ
...'"
0
Ol
>
"0
0
..r::.
'"...QJ 0.1
..r::.
f-
(d) Subcutaneous
needle electrode
0.01 '--_~_---'-_.L..-J.......L-=--_....L.-_-J,.._'--L......I
10 2 103 104
Discharge capacitance (pF)
FIGURE 7.4. Threshold sensitivity contours for four methods of stimulation using
capacitor discharges of positive polarity. (Measured data slightly adjusted to con-
form to large-sample results-see text.) (From Reilly and Larkin, 1987.)
Two of the curves in Fig. 7.4 follow straight lines representing constant
charge; that is, CV = constant. The other two curves depart significantly
from a constant-charge contour at the larger capacitances. Besides differ-
ences in curve shapes, there are significant vertical displacements among
the curves. The contour shapes of Fig. 7.4 can be understood with reference
to S-D relationships for capacitor discharges, as seen in theoretical models
[Fig. 4.14 and Eq. (4.14)], as well as experimental perception data (Wessale
et aI., 1992). Figure 7.5 illustrates S-D data from the author's experiments
for capacitor discharge stimuli (Larkin and Reilly, 1984), plotted in terms of
normalized charge units against the measured discharge time constant, i.
The vertical axis has been normalized by the minimum threshold charge
corresponding to the smallest values of i (obtained with a lOOpF capacitor
discharge). The curve in Fig. 7.5 is an analytic S-D expression according to
Eq. (4.14), with the parameter ie = O.6ms. The figure demonstrates that, for
monophasic current pulses that are brief relative to i" thresholds converge
to a minimum charge. This minimum charge is not sensitive to the fine

structure of the stimulus waveform as long as the stimulus remains
monophasic. For biphasic (oscillating) waveforms, the waveshape can be
critical (see Sects. 4.6, 7.5).
It is possible to explain the contour shapes of Fig. 7.4 in the light of S-D
relationships. For the range of parameters represented in the figure, stimu-
lus time constants range from about O.l,us to more than 1,OOO,us (see Fig.
2.35). For procedures (b) and (d), discharge time constants were in all cases
below 3,us, a value much smaller than the time constants typical of excitable
membranes. These brief time constants result from the relatively low im-
pedance at the electrode/skin interface associated with the capacitive dis-
charge. In curve (b), the low impedance results from the large contact area
of the electrode; in curve (d), the electrode is small, but it bypasses the high-
impedance corneal layer of skin. In contrast, discharge impedance associ-
ated with curves (a) and (c) becomes very large as the stimulus voltage is
reduced. At small capacitance values, the time constants associated with
all four curves in Fig. 7.4 are sufficiently short to be within the charge-
dependent region of the S-D curve. In this region, the vertical displace-
ments of the threshold curves can be explained by other factors discussed
below.
Capacitor discharge thresholds developed by Swiss and Austrian groups
form the basis of acceptability criteria published by the IEC (1987). The
10
8 I
0 A ,
a
(; 6 o I
oj
~
ell
.c
o 0' I
0 4 ')'-
"0
o/..
Q)
(;
Vi A--...
"0
Q) A / It> Linear model
.~ /
'l-. 't e =0.6ms
Cii 2
E 0/-
<5 /
z A
,.. " .10 -
--
_..... ./ ....
.
10 50 100 500 1000 4000
Stimulus time constant (j.J.s)
FIGURE 7.5. Strength-duration data-perception of cutaneous capacitor discharges

for various procedures. (Thresholds normalized by minimum perception charge.)
(Adapted from Larkin and Reilly, 1984 in Perception and Psychophysics, vol. 36,
pp. 68-78, reprinted by permission of Psychonomic Society, Inc.)
10 1
U
3
Q)
e>
<0
..c:
()
10 0
8
6
4
10. 1
10 0 10 1 10 2
Voltage (V)
(a) 80 X 150 mm electrodes, adults

(b) 5 mm rod electrodes, adults
(c) As in (a), children age 4-14 yrs.
(d) As in (b), children age 4-14 yrs.
(e) 80 X 100 mm electrodes, adults
(I) 4 mm dia. wire electrodes
FIGURE 7.6. Perception thresholds for capacitor discharges. Curves a-d from Swiss
measurement; curves e and f from Austrian measurements. (From Biegelmeier,
1986.)
European measurements are shown in Fig. 7.6, as summarized by

Biegelmeier (1986). Curves a and c apply to Swiss experiments in which
subjects grasped with both hands cylinders of 80-mm diameter and length
100mm. Capacitor discharges were applied to the cylinder electrodes.
Curve a is for adults, and curve c is for children. Curves band d, for adults
Suprathreshold Responses 253
and children, respectively, apply to a 5-mm-diameter rod electrode placed

on the fingertip. For the Austrian data, curve e applies to an 80-mm cylinder
grasped in the hands.
There are notable similarities in the contact electrode thresholds of Fig.
7.4 (curves band c) and European data if one allows for differences in
plotting formats. In both sets of data, thresholds are defined by constant
charge at small capacitances and then approach constant voltage at large
capacitances. Also, in both sets of data, thresholds for large and small
electrodes cross over when plotted against capacitance: at small capaci-
tances, thresholds are lower for the smaller electrodes; at larger capaci-
tances, the reverse is true. The curve shapes, and their relative
displacements, can be understood if we account for the effects of discharge
time constant and electrode size (see Sect. 7.6).
7.4 Suprathreshold Responses

Magnitude Scaling
One method of characterizing subjective intensity is through magnitude-
scaling experiments. For instance, a subject may be given a reference stimu-
lus that is assigned some arbitrary numerical value and asked to rate other
stimuli as a proportion or multiple of the reference. The absolute numbers
themselves are meaningless, but their rate of growth with respect to the
stimulus intensity reveals the subject's internal sensory interpretation.
In many cases, experimental data on SUbjective magnitude (M) versus
stimulus intensity (1) is fitted to a simple power function of the form
(Stevens, 1975).
M=alfJ (7.5)
where a is a magnitude-scaling constant, and f3 indicates the rate of growth
of sensory magnitude. Alternatively, the relationship is sometimes ex-
pressed in a form requiring three experimental parameters (Kaczmarek et
aI.,1992):
(7.6)
where 10 is the threshold-level stimulus intensity. The advantages of the two
forms have been debated over decades in a broad literature. The simpler
form of Eq. (7.5) is preferred by many for electrical stimulation; it has been
noted to give similar residual variance as compared with the more complex
form of Eq. (7.6) (Reilly et aI., 1982) and is argued to more closely convey
the subjective impression of observers (Rollman, 1974). At any rate, the
form of the two equations is similar for stimulus intensities sufficiently
above the perception threshold. A variety of values for the exponent f3 has
been reported for electrical stimulation (Tashiro and Higashyama, 1972); to
1000~----r-----~--r-~~-----r----~---r--~
3.5 2.7 2.3
/
/
/
/
/
/
/
10 100
Physical intensity (arbitrary units)
FIGURE 7.7. Psychophysical power function slopes for electrical, thermal, and
mechanical stimulation. Solid lines represent spark discharges at negative (-) or
positive (+) polarity. (Dashed lines from Stevens et ai., 1958, 1974.) The relative
positions of the functions are arbitrary on these coordinates. Slope values are given
at upper right for each function. (From Reilly et ai., 1982.)
some extent, these variations may be traced to features of the stimulus

waveform or to its locus or method of application.
Slopes of psychophysical power functions for electrical stimulation and
two other modes of natural stimulation are illustrated in Fig. 7.7 (Reillyet
aI., 1982). The numerical values assigned to each curve represent the expo-
nent f3 from Eq. (7.5). Values of f3 for capacitor discharge stimulation are
significantly greater than that for heat or pressure, but not as great as with
60-Hz stimulation. Stimulation of the finger results in uniformly higher
slopes (larger exponents) than for stimulation on the arm or leg. Negative
polarity discharges produce lower exponents than positive-polarity dis-
charges; on the average, the difference is 18%.
Categorical Scaling
Suprathreshold reactions may also be tested using categorical ratings in
which subjects choose from a list of affective (how unpleasant) or intensive
(how strong) descriptors. Although it might be thought that affective and
intensive scales would be unfailingly linked, this is not always the case.
Indeed, with the administration of narcotic analgesia, it is possible to reduce
significantly the intensive aspects of painful electrical stimulation without
affecting its unpleasantness (Gracely et aI., 1979).
Figure 7.8 provides an example of suprathreshold measurements includ-
ing both subjective magnitude scaling and intensive descriptors (Reilly and
Larkin, 1984). In these experiments, subjects were remarkably consistent in
their use of the two scales even though they were derived in separate
procedures. As a result, it is possible to display both on a single graph.
These data show that the growth of sensation magnitude is much greater
than that of stimulus magnitude. When fitted by a power function, per-
ceived magnitude grows at about the 2.5 power of stimulus magnitude for
stimulation of the finger, the 1.6 power for the arm, and the 1.4 power for
the leg. This range of exponents corresponds well with values reported in
past studies using pulsed electrocutaneous stimulation (Sternbach and
100
Very strong
Strong
Moderate
Q)
10
'0
::J
.~ Weak
Ol
<0
E Very weak
~
0
en
c:
Q)
en
800 pF
On finger On leg On arm
+ Polarity • .•
- Polarity 0 A 0
0.1
0.1 0.2 0.4 0.6 1.0 0.2 0.4 0.6
Discharge voltage (kV)
FIGURE 7.8. Growth of sensory magnitude for capacitor dischargers to three stimu-
lation sites on the body. Vertical coordinate shows numerical magnitude judgments
and ranges of adjectival rating categories. Composite data for eight subjects. (From
Reilly and Larkin, 1984.)
Tursky, 1965; Rollman, 1974; Sachs et aI., 1980; Higashiyama and Rollman,
1991).
The faster growth of sensation magnitude for the fingertip may be a
consequence of its small volume relative to the arm or leg. Because of the
volume constraint, current density becomes uniform along the finger
beyond the stimulation point. This appears to result in a more spatially
extensive sensory excitation: at suprathreshold levels, subjects report ex-
tended sensations along the finger.
In Fig. 7.8, sensation magnitude tends to progress along a reduced slope
at the larger stimulus levels. A similar description of a two-limbed power
curve has been noted by others (Rosner & Goff, 1967). The reason for the
deceleration may be explained by a general theory of interaction between
excitatory and inhibitory neural processes, in which inhibition increases
faster with stimulus level than does excitation (Atkinson, 1982). It is equally
plausible, however, that the change in slope reflects a shift among popula-
tions of excitable neurons. The group of polymodal nociceptive fibers, for
example, would respond only when their relatively high excitation thresh-
olds are reached. Other groups of neurons may respond at various thresh-
olds of the electrical stimulus. Because electrical stimulation has no
specialized transduction mechanism, it is likely that the psychophysical
function reflects a mixture of neural populations. The finding that both A
and C fibers can participate in the process of pain transduction (Meyer and
Campbell, 1981; Campbell et aI., 1979) tends to support this hypothesis.
Stimulation levels corresponding to specific affective or intensive adjec-
tives may be conveniently reported as multiples of perception thresholds.
Suprathreshold multiples, if properly determined, are relatively consistent,
despite wide variations in absolute thresholds due to factors such as
electrode size or intersubjective sensitivity variations. Table 7.2 lists
suprathreshold responses as multiples of the mean perception threshold as
reported by investigators who used single monophasic pulse stimuli. In the
data of Larkin and Reilly, for example, "tolerance thresholds" were deter-
mined by presenting the subjects with pairs of stimuli in an ascending
sequence. The tolerance limit was reached when subjects indicated an
unwillingness to accept a second stimulus or to proceed to the next higher
level. Tolerance limits determined in this manner are highly dependent on
the context of the experimental procedure.
AC or Repetitive Stimuli
A brief electrical stimulus will elicit a single action potential (AP) on
excited sensory neurons. The growth of sensory magnitude with increasing
strength of a brief stimulus can be accounted for by the corresponding
increase in recruitment of sensory neurons, including less-sensitive nocice-
ptive fibers residing in deep subcutaneous strata. An additional factor,
however, is involved in the growth of sensation for continued AC or repeti-
TABLE 7.2. Average suprathreshold multipliers for single monophasic pulses.

Body Multiple above perception
locus N Electrode Stimulus· Unpleasant Pain Tolerance Reference
Finger 8 0.8-mm CD,800pF 2.3 3.5 7.1 Reilly and
diam. Larkin
(1984)
Forearm 8 0.8-mm CD,800pF 3.5 5.5 11.0 Reilly and
diam. Larkin
(1984)
Leg 8 0.8-mm CD,800pF 3.1 9.1 Reilly and
diam. Larkin
(1984)
Finger 124 Tapped CD,200pF 2.6 Larkin and
electrode Reilly
(1986)
Triceps 6 4-cm2 carbon Pulse, 4.3 8.9 Alon et al.
sponge 5-1O,us (1983)
Triceps 6 4-cm2 carbon Pulse, 7.4 10.5 Alon et al.
sponge 20-1000,us (1983)
·CD, capacitor discharge stimulus.
tive pulse stimulation: sensory magnitude will also grow because of the
higher CNS responses to repetitive APs (see Chapter 3). As the magnitude
of an alternating current or pulse train is increased above the threshold
level, there is a corresponding increase in the rate of AP production. As a
result, continued AC or repetitively pulsed stimulation may be judged to be
more intense than a single brief electrical stimulus of equal magnitude.
Table 7.3 lists suprathreshold multiples for AC or repetitive stimuli based
on reported experimental data. Although it is clear that AC or repetitive
stimuli produce a dynamic range from perception to pain that is much less
than with single pulsed stimuli (Table 7.2), there is much more variance in
the multiples for the reported data. The pain multiples of Budinger and of
Bourland are small in comparison to the other experiments. In these experi-
ments, stimulation was induced by exposure of the torso to bursts of
sinusoidal magnetic fields, and excitation was typically felt at the buttocks,
lower back, or flank. It is not clear whether the relatively small multiples
applying to these experiments were the result of the relatively extended
spatial distribution of excitation or to some other unique feature associated
with magnetic stimulation. The multiples of Higgins appear to be particu-
larly great. In this case, stimulation was provided by a small concentric
electrode arrangement on the forearm. The relatively large multiples in
Higgins' experiments might be explained by the restricted depth and lateral
extent of excitation associated with the particular electrode arrangement.
From these observations, one might suspect that painful sensations would
be enhanced as the spatial extent of electrical excitation is made greater,
although experimental verification of this conjecture is lacking.
TABLE 7.3. Average suprathreshold multipliers for AC or repetitive pulse

stimulation.
Body Multiple above perception
locus N Electrode Stimulus" Unpleasant Pain Tolerance Reference
Fingertip 6 Tapped 60-Hz AC 1.8 2.4 3.5 Reilly and
electrode Larkin (1987)
Forearm 40 l-cm diam. Pulse train, 2.0 5.7 Rollman and
100Hz Harris (1987)
Fingertip 2 * 100-10 kHz 2.9 4.0 Hawkes and
AC Warm (1960)
Fingertip 367 Tapped 10-kHzAC 2.3" Chatterjee et al.
electrode (1986)
Forearm 12 Concentric 60-Hz AC 4.2 6.6 11.8 Higgins et al.
ring (1971)
Fingertip 12 Saline bath 60-Hz AC 1.3 Currence et al.
(1987)
Torso 2 Magnetic 600-1950Hz 1.3 Budinger et al.
field AC (1994)
Torso 52 Magnetic 128 pulses 1.6 b 2.0 Bourland et al.
field (1997)
*Information sketchy or not available.

"Chatterjee threshold described as "pain," but may be functionally equivalent to "tolerance".
bBourland category of "uncomfortable" attributed to "pain".
Note: N = number of subjects tested.
Figure 7.9 illustrates sensory power functions for stimulation by O.l-ms

pulses, delivered at a 60-Hz rate (from Rollman, 1974). The parameter n
indicates the number of pulses in the train; f3 indicates the sensory magni-
tude power law, according to Eq. (7.5). It is seen that the power exponent
grows with increasing n. More striking is the increase in sensory magnitude
with increasing n for pulse trains at the same stimulus intensity. At 3.0-mA
current level, for example, perceived sensory magnitude increases sixfold as
the number of pulses increases from 1 to 30.
Dynamic Range of Electrical Stimulation

The data in Table 7.2 and 7.3 show that the dynamic range for electro-
cutaneous stimulation is very small. For stimulation by single pulses, the
ratio of pain to perception threshold is only 3.5 to 7.4, depending on the
locus of stimulation; for AC or repetitive stimulation, the dynamic range is
even less. Contrast the measured dynamic range of electrical stimulation
with that for hearing or pressure sensitivity, both of which have dynamic
ranges of about 100,000 to 1. We can thus appreciate the importance of
perception threshold measurements in electrical sensitivity studies. If we
can measure the perception threshold accurately, we know that a relatively
small increase will result in a strong perceptual effect.
Supra threshold Responses 259
The ordering of tolerance values on the finger, arm, and leg has been
found to be consistent among subjects who may differ widely in absolute
threshold levels (Reilly et aI., 1982). Overall, the fingertip is more sensitive
than the leg by nearly a factor of 2, and the arm takes a midway position.
This ordering contrasts with measurements taken at the perception thresh-
old (see Sect. 7.7), in which electrical stimulation is more easily detected on
the forearm than on the fingertip. One factor that may account for this
difference is that current flow through the finger is more volume limited
than in the arm or leg. Consequently, at points distant from the introduction
of current, current density may be higher in the finger than at equivalent
distances on the arm or leg. Subjects experience supra threshold stimulation
100
n = 30 3
50
"0
(])
30
::l
.t::
c:
Ol
<1l 20
E
(])
·u>(])
E 10
::l
(J)
(])
.~
iii
a:;
a: 5
2 3 5 10 20 30 50
Current (rnA)
FIGURE 7.9. Sensory power functions for train of pulses (width = 0.1 ms), delivered
at 60-Hz rate; n indicates number of pulses in train; f3 indicates power function
exponent. (Adapted from Rollman, 1974, in Conference on Cutaneous Commu-
nication Systems and Devices, pp. 38-51, reprinted by permission of Psycho nomic
Society, Inc.)
as "radiating" through the finger and hand. This indicates that current
density remains relatively high over a broad region and that more than one
population of afferent neurons may be involved. On the leg or arm, in
contrast, the sensation appears to be more limited spatially.
There is a wide variation among individual tolerance limits. Although
part of this variation may reflect individual differences in sensitivity, it is
clear that "tolerance" is not simply a sensory limitation. Subjects with high
tolerance limits may have a detached attitude about the possibility of injury,
whereas those with low limits may be more cautious or fearful. This varia-
tion does not seem tied to an individual's experience or familiarity with
electrical shocks. Indeed, the author has encountered subjects who had no
experience, but were willing to accept levels substantially higher than those
set by the highly experienced investigators. The acceptance of a large
stimulus does not necessarily imply a lack of cutaneous sensitivity. Individu-
als who tend to have very nearly the same thresholds and give numerical
magnitude estimates in the same range may have markedly different
tolerance levels.
In view of the preceding comments, the tolerance limits in Tables 7.2 and
7.3 cannot be taken as absolute limits on an individual's capacity for painful
stimulation. The concept of tolerance has empirical application only as a
relative, context-dependent measure. Implied demand, or social expecta-
tions, influence a person's willingness to cooperate in a scientific endeavor.
For some individuals, neither pain nor the potential for skin injury deters
their voluntary exposure to electric shock if they believe that there is some
benefit to be gained.
7.S Stimulus Waveform Factors2

Sect. 7.3 showed that neural excitation thresholds are markedly sensitive to
the duration of a monophasic stimulus current. Other features of the stimu-
lus waveform, such as biphasic properties and pulse repetition patterns, can
be equally important. A general theoretical framework for evaluating such
effects can be provided by appropriate neural excitation models, such as the
SENN model discussed in Chapter 4.
Biphasic Stimuli
As noted in Sect. 4.6, the current reversal of a biphasic pulse can suppress
a developing AP that was elicited by the initial phase. To compensate for
the reversal, a biphasic pulse must present a higher initial current. When
thresholds are measured in terms of the initial current, the biphasic pulse
2Portions of this section have been adapted from Reilly (1989).

Stimulus Waveform Factors 261
TABLE 7.4. Relative thresholds of biphasic stimuli:

comparison of SENN model and experiment.
Waveform parameters Threshold ratio (Q/Q2)
tp (Q/Q2) (!Is) 6 (Q/Q2) (fts) Model Experiment
20,20 10,50 1.20 1.19
50,50 10,50 1.10 1.12
50,20 50,50 1.17 1.23
Notes: tp = phase duration; 6 = interpulse delay.

Source: Experimental data from Biitikofer and Lawrence
(1978). SENN model as in Chapter 4.
may therefore have a higher threshold than a monophasic pulse. According

to Fig. 4.17, the threshold elevation depends on the phase duration and the
time delay for current reversal. Thresholds are elevated more when the
pulse is short and when the current reversal immediately follows the initial
pulse.
Experimental results with biphasic stimuli generally confirm the modeled
threshold relationships of Fig. 4.17. One study tested suprathreshold sen-
sory sensitivity of the hand to biphasic pulses having durations of 20 to
SOl's, and with delays from 10 to 50,us (Biitikoffer and Lawrence, 1978).
Although it is difficult to compare absolute thresholds with model results,
we can compare the relative sensitivity for different waveforms. Table 7.4
lists the relative thresholds of biphasic pulse doublets having different
phase durations (tp ) and interpulse delays (0). The experimental and model
results agree within a few percentage points. Other experimental data
confirm that a biphasic stimulus has a reduced efficacy for neuromuscular
stimulation as compared with a single monophasic pulse of the same phase
duration (Gorman and Mortimer, 1983; van den Honert and Mortimer,
1979a).
The finding that thresholds are elevated for a single biphasic pulse should
not be interpreted as implying that thresholds for all oscillatory stimuli are
necessarily elevated above monophasic stimuli of the same phase duration.
Indeed, if a biphasic stimulus is repeated as an oscillatory waveform, thresh-
olds may decrease with successive oscillations to the point that they eventu-
ally reach the single pulse monophasic threshold (see Sect. 4.6).
Repetitive Stimuli
We have seen in Chapter 4 that the threshold for exciting a single AP by a
sequence of pulses may be lower than the threshold for a single stimulus
pulse. Repetitive stimuli can also enhance sensory (and motor) response if
multiple APs are generated. In either case, there is an integration effect of
the multiple pulses. In the first case, the integration takes place at the
membrane level. In the second case, response enhancement takes place at
the CNS level for neurosensory effects and at the muscle level for neuro-
muscular effects (see Chapters 3 and 8).
Multiple pulse threshold effects predicted by a neuroelectric model are
illustrated for two pulses in the lower section (M < 1) of Fig. 4.17. The
effects are most pronounced for short pulses and short interpulse delays.
The neuroelectric model also predicts that thresholds consistently fall as the
pulse number is increased until a minimum plateau is reached. Figure 4.23
suggests a temporal summation process that effectively sums repetitive
pulses, with an integration time that is roughly four times the S-D time
constant. At a delay of 500.us, the neuroelectric model shows no measurable
integration effect. Similar results with neuroelectric models have been
reported by others (Biitikofer and Lawrence, 1979).
Tasaki and Sato (1951) experimented with myelinated toad nerve to
examine the shift in threshold of a stimulus pulse caused by a previous
conditioning pulse of either the same or the opposite polarity. The duration
of the two pulses was lO.us, and the interpulse delay varied from 0 to 200.us.
Results from these experiments are shown in Table 7.5; tabulated values
express the threshold of a pulse doublet as a multiple of a single-pulse
threshold. Also shown are theoretical predictions taken from Fig. 4.17; the
experimental data and theoretical prediction agree quite well. Even for
pulse delays as long as 5 ms, cutaneous perception threshold reductions on
the order of 7 to 8% have been reported for continuous pulse trains with
pulse durations from 100 to 400.us (Hahn, 1958).
The thresholds of perception as well as supra threshold reactions both
decline with increasing number of pulses in a stimulus train. Figure 7.10
illustrates data of Gibson (1968) obtained with monophasic pulses individu-
ally having durations of 0.5 ms and delivered at a rate of 100Hz. The vertical
axis indicates the threshold normalized by the value obtained with n= l.
The plotted curves represent averages of thresholds obtained for stimula-
tion at two different groups of body loci, which appeared to have different
functional dependencies with pulse number. Group I applies to the
TABLE 7.5. Threshold multiples for monophasic and

biphasic pulse doublets (tp = lO.us).
Theory Experiment
Delay (.us) Monophasic Biphasic Monophasic Biphasic
0 0.54 1.99 0.60 1.40
20 0.65 1.21 0.76 1.36
50 0.72 1.07 0.70 1.30
100 0.80 1.02 0.82 1.15
200 0.91 1.00 0.95 1.04
Source: Theoretical data derived from SENN model (see

Chapter 4). Tabulated values are the threshold of a two-pulse
doublet expressed as a ratio of the single-pulse threshold.
Experimental data derived from Tasaki and Sato (1951).
,
\,
\\
0.8
\\\\
Perception, group I
'0
:g
Ul
~
£ 0.6
.~
'lii
Qj
a:
Perception, group II
.•......,
"\. ""'. Pain, group I
0.4
' .......
". Pain, group II
0.2 '--_ _--'-_-'---'-____---'--'-........-'--_ _--'-_-'-_'---"----~_'_......

1 10 100
Number of stimulus pulses
FIGURE 7.10. Relative thresholds versus number of stimulus pulses at 100-Hz rate;
pulse width = 0.5 ms. Designations I and II refer to averages for two different
groups of body loci. (Data extracted from Gibson, 1968.)
fingertip, lip, and ear; group II applies to the forearm and back. The coef-
ficient of variation of the data represented in the figure was typically 0.1.
Notermans (1966) related the threshold of pain to the number of pulses,
each having a width of 5ms and delivered at a rate of 100Hz. His results,
when plotted on a format similar to that in Fig. 7.10, nearly coincide with
the pain data of Gibson.
We have seen, in connection with Fig. 4.22, that a theoretical neuron
model effectively sums pulsed stimuli over a duration roughly four times
the S-D time constant. The data of both Gibson and Notermans demon-
strate thresholds that continue to decline with increasing n, out to the
maximum duration of the pulse train (about 200ms). This observation
indicates a temporal integration time constant of roughly 50ms. It is difficult
to justify a time constant that long solely from the properties of a single
neuron, whose S-D time constants are typically a fraction of a millisecond
(see Table 7.1). The precise mechanism of temporal integration in electrical
perception remains poorly understood.
Besides lowering thresholds relative to a single pulse, multiple pulses that

are individually near or above threshold can generate a train of APs. The
consequences can be elevated sensory magnitudes for afferent stimulation
or enhanced muscle contraction for efferent stimulation. Sensory magni-
tude and unpleasantness in both A and C fibers are related to several
factors, including the AP frequency and burst duration (Campbell et aI.,
1979; Gybels et aI., 1979). Figure 7.11 illustrates the effects of sensory
enhancement because of multiple pulses that are individually above thresh-
old. The stimulus was designed to study 60-Hz electric field induction
effects, in which a train of capacitor discharges is produced on individual
half-cycles of a 60-Hz waveform (i.e., 120-Hz repetition rate) (Reilly and
Larkin, 1987). Additional details concerning this figure are given in Sect.
Pain
20
Unpleasant
10
8
6
Neutral
Ql
"0
:J
.<::
c
4 ,
Cl
'" " "

""
E
c
""
0
';; 2 7
'"C
VI
Ql
Cf)
1.0
0.8
0.4~ ____- L_ _ _ _ _ _ ~ _ _ _ _~~_ _ _ _~_ _ _ _ _ _~
1 2 4 8 16 32
Number of stimulus one-half cycles
FIGURE 7.11. Effect of the number of stimulus cycles on sensation magnitude for
60-Hz AC electrical field-inducted stimuli, discharge capacitance = 100pF. (From
Reilly and Larkin, 1987.)
9.4. The horizontal axis gives the number of stimuli, and the vertical axis
gives the perceived magnitude (an arbitrary, but consistent, scale assigned
by subjects). Also shown are regions of subjects' qualitative descriptors.
In this experiment, sensory magnitude grows at about the 0.8 power of
the number of stimuli. Notice that a stimulus that is perceptible, but
qualitatively "neutral," can become highly unpleasant when presented as a
repetitive train.
Rollman (1974) found that sensory magnitude scaled approximately as
the 0.5 power of the number of pulses up to N = 30 (pulse width = 0.1 ms,
repetition interval = 16ms; see Fig. 7.9). Experiments with biphasic pulse
trains (Sachs et al., 1980) found that sensory magnitude scaled with N to the
0.8 power; these results were likely a combination of membrane integration
and multiple AP effects, because the experimental repetition interval
(50,us) was much too small to support an AP on each pUlse.
Sinusoidal Stimuli
The sinusoidal stimulus is a special case biphasic waveform, which has been
discussed in Chapter 4 and also in Sect. 7.4. We have seen that the current
reversal of a biphasic waveform can reverse the development of membrane
excitation caused by the initial phase of the stimulus and elevate excitation
thresholds. The degree of desensitization caused by the biphasic current
reversal is increased as the duration of the stimulus is reduced. This
phenomenon partially accounts for the high-frequency upturn noted in
strength-frequency (S-F) curves for neural excitation (Fig. 4.21). The S-F
curve reaches a minimum plateau as the frequency is reduced. If the
frequency is reduced further, the threshold begins to rise again as the time
rate of change of the sinusoidal stimulus becomes small.
The overall shape of the frequency sensitivity curve has been developed
in Chapter 4 using a neuroelectric model for myelinated nerve. A math-
ematical fit to the thresholds determined from the model expressed by Eq.
(4.35) is based on an analogy to the exponential S-D expression for
monophasic stimulation. The expression includes terms to account for the
threshold rise at high and low frequencies. By analogy with the S-D expres-
sion, a high-frequency parameter Ie specifies the frequency at which the
threshold upturn occurs; a similar term fo applies to the low-frequency
upturn. The low-frequency term is constrained to a maximum value to
account for the fact that stimulation is possible with direct current. A
maximum value of KL ::5 4.6 may be inferred from the ratio of the percep-
tion threshold for continuous DC to that for 60-Hz AC (Dalziel and
Mansfield, 1950b).
Although we can suggest a functional form of an S-F curve from theoreti-
cal principles, the parameters of this functional relationship need to be
determined experimentally. Table 7.6 summarizes experimental S-F data
determined from neurosensory and neuromuscular responses to sinusoidal
TABLE 7.6. Empirical strength frequency curve parameters.

Locus of 10 t. 10
Species stimuli N Electrode Response (mA) (Hz) (Hz) Ref.
Human Hand 115 Hand-held wire Perception 1.1 500 <60 a
(m)
Human Arm 2 0.27 cm2; elect. Perception 0.11 1,000 50 b
paste
(concentric)
Human Thorax 8 1.0cm2 Perception 0.2 100 <20 c
Human Neck 8 25cm2 Perception 0.5 150 <20 c
Human Fingertip 2 * Perception 0.4 500 <100 d
Tolerance 1.0 400 <100 d
Pain 1.8 460 <100 d
Human Hand Large-area grip Tetanus 15.0 600 8 a
Frog Sciatic * In vitro,
nerve elect. spac. AP 1,500 -30 e
=2mm
elect. spac. AP 533 -30 e
= 25mm
Rat Gast. tib. 8 Subcutaneous Muscle 200 <100 f
musc. twitch
Toad Mye. * In vitro AP 200 <100 g
nerve
Frog Mye. * In vitro AP 500 28 h
nerve
* Information incomplete or not available.

References: (a) Dalziel (1954); (b) Anderson and Munson (1951); (c) Geddes et al. (1969);
(d) Hawkes and Warms (1960); (e) Hill et al. (1937); (f) LaCourse et al. (1985); (g) Tasaki and
Sato (1951); (h) Wyss (1963).
stimulation. The table lists experimental estimates of Ie and 10 from Eq.

(4.35). The table also lists the minimum threshold for electrocutaneous
stimulation. For proper comparison of the threshold, one should take into
account the electrode size (see Sect. 7.6).
The values of Ie in Table 7.6 encompass a rather wide range, with a
median (and also geometric mean) of 500 Hz. Part of the variation might be
due to the electrode size and its configuration. In the experiments of Hill et
al. (1937),/e changed from 1,500 to 533Hz when a bipolar electrode spacing
was changed from 2 to 25 mm. For the electrocutaneous sensory data
reported in Table 7.6, there appears to be a tendency toward large Ie values
with smaller electrodes, but there are insufficient data to establish a clear
relationship. The direction of the effect of electrode size on Ie is consistent
with the corresponding effect on the S-D time constant, T" which increases
monotonically with electrode size in neural stimulation (Pfeiffer, 1968;
Davis, 1923) and also with cardiac stimulation (see Sect. 6.3).
The form of the S-F curve suggests that the perception threshold current
ought to increase indefinitely as its frequency is increased. As a practical
matter, however, if the electrical threshold is raised high enough, sensory
thresholds will be dominated by thermal perception. Because heating

effects are largely independent of frequency, perception thresholds cease to
rise at that point. Experimental data with continuous sinusoidal stimuli
show that, at a frequency of about 100kHz, perception thresholds reach a
maximum plateau as a result of thermal perception (Chatterjee et aI., 1986;
Dalziel and Mansfield, 1950b) as seen in Fig. 7.12. However, for pulsed
stimuli, the frequency above which thermal perception will dominate elec-
trical excitation depends greatly on the fraction of on-time (duty factor), as
the heating capacity of electrical current (i.e., its rms value) is proportional
to the square root of the duty factor (see Sect. 11.3). Consequently, one can
extend the frequency where thermal perception dominates electrical
perception by using pulsed stimuli of low duty factor.
Thermal perception of contact current was tested by Rogers (1981) in
the frequency range 2 to 20 MHz. Fifty adult subjects touched an elec-
trode energized by a radio frequency source; the return path was through
a ground plane on which the subjects stood. Perception and dis-
comfort thresholds rose with increasing frequency. The median perception
Males
Females
10 5 10 6
Frequency (Hz)
FIGURE 7.12. Perception and tolerance thresholds for AC stimulation of high

frequency. Finger contact on 25-mm2 plate. (Adapted from Chatterjee et aI., 1986,
© 1986 IEEE.)
thresholds were 45, 50, 80, 86, and 97mA at frequencies of 2,5,10, 15, and
20MHz, respectively, for stimulation on the back of the finger. Discomfort
thresholds at those frequencies were 150, 173, 183, 190, and 206mA. The
ratio of discomfort to pain ranged from 3.3 at 2 MHz, to 2.1 at 20 MHz; these
ratios are not unlike those for electrical stimulation at much lower frequen-
cies (Table 7.3). Fingertip thresholds were approximately a factor of 2
higher than those on the back of the finger. With a grasping contact, current
of 500mA or more could be conducted for short periods without
discomfort.
The reason for the gradually rising thresholds with frequency was not
addressed by Rogers but might be explained by the impedance properties
of the skin. As the frequency is increased, capacitive coupling of the skin
would increasingly bypass resistive elements (see Sect. 2.2). Accordingly,
PR energy would tend to be dissipated in deeper tissue rather than in a
concentrated locus within the skin as the frequency is increased.
Figure 7.13 illustrates S-F curves from several of the references listed in
Table 7.6. Also drawn is a theoretical curve defined by Eq. (4.35), using the
empirical constants Ie = 500 Hz and 10 = 30 Hz. In general, the analytic
expression provides a good fit to the functional form of the S-F curves,
provided that the constants Ie and 10 are properly selected.
30r-------------------------------------~
····· ..··0· .. ·· .. Hill et al; 1936 (large elect.)

.........•....... Hill et al; 1936 (small elect.)
·········0······· Wyss, 1963
.........•....... Tasaki & Sato, 1951
.........* ....... LaCourse, 1985
......... -&-••••••• Anderson & Munson, 1951
Model
10 100 10000
Frequency (Hz)
FIGURE 7.13. Strength-frequency curves for perception of sinusoidal currents ap-

plied cutaneously. Symbols represent experimental data from several investigators.
Solid curve is analytic expression with t. = 500 Hz and fo = 30 Hz.
The analytic S-F expression provides a good fit to experimental data in

most cases. However, that particular representation may not apply univer-
sally. In patients with diabetic neuropathy, frequency sensitivity reported
by Katims et al. (1987) deviates significantly from the analytic form. In
Katims' patients, thresholds were typically elevated above the values noted
for normal individuals. It was especially notable that perception thresholds
on the hands and feet of the diabetic patients were significantly smaller
at 2,000Hz than at 250Hz; the direction of this frequency sensitivity is
opposite to that found in normal people. The electrophysiological basis for
this finding is presently unknown.
The S-F expression of Eq. (4.35) has been verified for human perception
thresholds up to about 100kHz. However, one would expect the relation-
ship to apply to perception thresholds at much greater frequencies. As
noted in Sect. 4.6, neuromuscular stimulation has been observed to approxi-
mately follow frequency-proportional thresholds up to 1 MHz. Also, one
could infer experimental correspondence up to several MHz based on the
fact that the strength-duration law can be verified for capacitor discharges
as brief as O.1.us.
Polarity Effects
Measurements taken under a wide variety of experimental conditions dem-
onstrate lower thresholds for monophasic stimuli of negative polarity than
with positive polarity. In our experiments perception thresholds for nega-
tive polarity stimuli averaged 23 % lower than those for positive polarity
(Reilly and Larkin, 1983). We found similar polarity differences (about
30%) when the discharges were applied to corneum-piercing needles,
demonstrating that polarity-sensitive mechanisms do not depend on the
properties of the corneum.
Similar polarity sensitivity has been found by others (Higashiyama and
Rollman, 1991). Gibson (1968), for example, reported cathodic perception
thresholds to be approximately 50 to 75 % of those for anodic pulses. Girvin
and colleagues (1982) tested perception thresholds for monophasic pulses
having durations ranging from 63 to 1,000.us. At the shorter durations, the
ratio of cathodic to anodic thresholds was 0.77; at the longer durations,
the ratio was 0.63. Polarity sensitivity similar to that found for sensory
excitation has also been noted for simulation of cardiac tissue (refer to
Chapter 6).
We have seen in Chapter 4 that polarity sensitivity is predicted from
theoretical considerations of neural excitation. With anodic stimulation, the
distribution of current influx and efflux along the axon dictates that the
nerve may be excited by current from an external electrode of either polar-
ity, but that a cathodic electrode should be significantly more effective (Fig.
4.12). A model for excitation near the terminal structures of nerves also
demonstrates that cathodic thresholds are lower than anodic thresholds and
that the magnitude of the polarity effect is similar to that noted in sensory
experiments (Fig. 4.13).
Biphasic stimuli may also exhibit a sensitivity to the polarity of the initial
phase, but the polarity effects are generally less than with single
monophasic pulses. The neuroelectric model described in Chapter 4, for
example, shows that with short-duration biphasic pulse doublets, the
threshold of excitation is lower if the initial phase is cathodic rather than
anodic.
7.6 Electrodes and Current Density

In Chapter 4, we showed that the relevant force contributing to electrical
excitation is the electric field within the biological medium. The field has the
maximum effect when it is aligned along the long axis of an excitable cell
(neuron, nerve ending, or muscle fiber). The electric field, E, is related to
the current density, J, by E = J/a, where a is the conductivity of the
medium. It can be rather difficult to determine the spatial distribution of E
or J in a typical experimental situation involving electrodes within or con-
tacting biological materials. Frequently, one has only a gross estimate of the
average current density at the surface of the stimulating electrode.
Cutaneous Electrodes
The relationship between electrode area and perception sensitivity was
determined for capacitor discharges to contacts of different sizes (Reilly
and Larkin, 1984; Reilly et aI., 1983). To avoid artifacts of electrode place-
ment, the precise point of contact was varied from one trial to the next,
within a perimeter defined by the largest electrode. A low capacitance
(100pF) was used in these tests so that stimulus discharges would have time
constants within the charge-dependent region of the S-D relationship
regardless of the electrode size. The electrode was either dry or treated
with electrode paste.
Results for untreated skin (Fig. 7.14) show that electrode size is a critical
parameter only for diameters greater than about 1 mm. Below this point,
sensitivity is nearly constant. We hypothesize that for dry skin, current is
conducted through discrete channels beneath a contact electrode, so that
effective current density depends on the number and size of these channels
and not simply on the electrode size. This hypothesis is consistent with the
observation that current concentration is not homogeneous in the corneal
layer of skin (Mueller et aI., 1953; Panescu et aI., 1993). As discussed in Sect.
2.2, tests with electroplating on the skin surface suggest a density of about
one channel per square millimeter (Saunders, 1974). This estimate corre-
sponds very well with the plateau in Fig. 7.14. For electrodes smaller than
1 mm2 , a single channel of excitation may be produced in dry skin. For larger
CD (a) Paste-treated
e>
<0
fingertip
.c
o
32 0.1
o
.c
Ul (b) Fingertip
~
.c
~
(c) Forearm
100 pF
+ polarity
0.01 L------'_--L---"---'-..L-_'-----'----'--'---L.-_.L-----''--'--'-'
0.1 1.0 10 100
Electrode diameter (mm)
FIGURE 7.14. Effect of electrode contact size on perception of discharge from 100-
pF capacitor. Curve (a) applies to paste-treated skin; the others apply to dry skin.
(From Reilly and Larkin, 1985a.)
electrodes, the discharge current may pass through the dry epidermis in
more than one place. If so, current density would be constant for any
electrode smaller than about 1 mm2 and would decrease only when the
electrode is made to cover at least two current channels.
If current density is a critical parameter for human sensitivity, then
thresholds eventually must rise as electrode area is increased. This effect is
expected whether the current is uniformly distributed or is concentrated in
small current channels. For the data shown in Fig. 7.14, thresholds above
I-mm diameter increase as the one-third power of area on the forearm
and leg and as the one-sixth power on the fingertip. Slopes falling between
the one-sixth and one-third powers are evident in the data of previous
investigators who used 60-Hz stimulation, as shown in Fig. 7.15 (Jackson
and Riess, 1934; Forbes and Bernstein, 1935; Nethken and Bulot, 1967;
Dalziel, 1954).
Although cutaneous sensory thresholds rise as the electrode size is
increased, the rate of rise is much less than would be expected if thresholds
were simply inversely proportional to electrode area. The lack of area
proportionally is most likely related to the nonuniform conduction of cur-
rent beneath a cutaneous electrode. While electrode paste treatment or
skin hydration would be expected to reduce this effect, the current distribu-
tion beneath even a treated electrode is not uniformly distributed (refer to
Sect. 2.2). Furthermore, the presence of conduction "hot spots" in treated
skin cannot be ruled out.
When the skin is treated with electrode paste, the area relationship is
markedly altered. In Fig. 7.14, for example, thresholds on the paste-treated
o Jackson & Reiss, 1934 (finger)

• Forbes & Bernstein, 1935 (finger)
• Dalziel, 1954 (finger, hand)
D. Nethken, Bulot, 1967 (upper arm)
Contact area (mm 2)
, ",I , , , I , , .I , ",I
0.1 10 100
Contact diameter (mm)
FIGURE 7.15. Sensory thresholds for 60-Hz stimulation versus area of contact.
Data of Jackson and Reiss apply to pain threshold; other data apply to perception
threshold.
fingertip are elevated above the dry-skin thresholds by a factor of about 2

for small-diameter contacts. As the electrode size is increased, the thresh-
olds of dry and paste-treated skin eventually cross each other. The data of
Forbes and Bernstein shown in Fig. 7.15 also demonstrate that paste treat-
ment elevates thresholds by a factor of 2 above that for a small, dry elec-
trode and that the thresholds of dry and treated electrodes eventually
converge as the electrode size is increased. The elevation of thresholds for
hydrated skin can be explained by the fact that treatment would tend to
result in a more even distribution of current beneath the contact electrode,
thereby reducing (but not eliminating) current-density "hot-spots" and
electrode-edge current enhancement. Furthermore, current is more

readily conducted beyond the edges of the electrode if the skin is hydrated.
The crossover of thresholds at larger contact areas is more difficult to
explain.
The role of skin impedance breakdown in electrical sensation should not
be discounted. It has been noted that, with rising applied voltage, the
sensation of electric shock on the dry, untreated skin is accompanied by a
dramatic reduction in impedance (Nute, 1985; Gibson, 1968; Mason and
Mackay, 1976; Mueller et aI., 1953; Paneseu et aI., 1993). Such a decrease is
likely to be caused by skin breakdown in discrete locations, leading to a
sudden increase in current density in those locations.
Based on the preceding, we can postulate thresholds for brief mono-
phasic stimuli delivered to the finger or hand as given in Table 7.7.
The smallest listed "perception" threshold of O.l1.uC applies to a O.01-cm2
contact electrode; that value has been determined from the capacitor dis-
charge measurements of Reilly and Larkin (1987), as indicated by curve (c)
at 100-pF capacitance in Fig. 7.4. The corresponding value on the forearm
has been taken from the body-location sensitivity data, as discussed in Sect.
7.7. Values at other contact areas have been calculated by assuming that
thresholds rise as the one-sixth power on the fingertip or hand and as the
one-third power on the forearm, as discussed previously. The thresholds for
unpleasantness, pain, and tolerance have been determined from Table 7.2.
One postulate used to derive Table 7.7 is that the area power law deter-
mined for perception currents also applies to the supra threshold categories
of annoyance, pain, and tolerance. A second postulate is that the
suprathreshold multiples of Table 7.2 are invariant with respect to electrode
size. The perception threshold for the finger/hand in Table 7.7 compares
favorably with the European data using capacitive discharges (Fig. 7.6).
Note that in the Swiss large-electrode data, the contact area is probably on
the order of 100cm2•
TABLE 7.7. Calculated sensory thresholds for single monophasic stimuli to finger
or hand (threshold in p,C for short-duration currents).
Electrode Finger, hand threshold CuC) Forearm threshold (flC) area
area (cm2) Percept. Annoy. Pain. To!. Percept. Annoy. Pain. To!.
0.01 0.11 0.25 0.39 0.78 0.09 0.31 0.48 0.97
0.1 0.16 0.37 0.56 1.14 0.19 0.66 1.04 2.08
1.0 0.23 0.53 0.81 1.63 0.41 1.43 2.24 4.49
10.0 0.34 0.78 1.19 2.41 0.88 3.08 4.84 9.68
100.0 0.51 1.17 1.79 3.62 1.89 6.62 10.40 20.79
Notes: Median threshold for adults. Contact to finger or hand. Brief-duration «20flS)
currents. Area power law = 116 for finger/hand; 113 for arm. Suprathreshold multiples from
Table 7.2.
Current Density Considerations

In many instances, researchers report current density thresholds for periph-
eral nerve stimulation. Because of the aforementioned difficulties, thresh-
olds calculated from the average current density beneath a cutaneous
electrode may be unrealistically low. Current density thresholds have been
calculated from a theoretical myelinated nerve model (Sect. 4.4). For fiber
diameters of 5,10, and 20,um, the current density thresholds were calculated
as 0.492, 0.246, and 0.123mNcm2, respectively, for medium conductivity of
0.2S/m (Table 4.2). In Chapter 9, these theoretical thresholds are compared
with experimental values applying to stimulation by externally applied
magnetic fields or by in-vitro nerve stimulation by uniform electric fields. In
these experiments, difficulties due to nonuniform electrode conduction are
not present. If allowances are made for variations in stimulation param-
eters, the experimental values are found to be within the range predicted by
the theoretical model. Accordingly, the value of about 0.lmA/cm2 appears
to represent a reasonable estimate of the current density threshold for
stimulation of the most sensitive peripheral nerves.
7.7 Body Location Sensitivity

It is well known that different parts of the human body have different
sensitivities to tactile and thermal stimulation. Figure 7.16 displays the
rank -orders of body-part sensitivity for three measures of tactile sensitivity:
pressure, two-point discrimination, and point localization (Weinstein,
1968). The rank-ordering applies to a mixed population of male and female
SUbjects. Females were found to have consistently lower pressure thresh-
olds on all body parts than males, as illustrated in Fig. 7.17; the vertical scale
expresses the logarithm of the sensitivity relative to 0.1 mg. The gender
difference is appreciable, reaching 10: 1, for example, on the back, forearm,
and leg. No such difference appears in Weinstein's data for spatial discrimi-
nation. We shall return to a discussion of male/female differences in sensi-
tivity in Sect. 7.10, where we show that females tend to be more sensitive to
electrical stimulation as well, and we advance a hypothesis for these differ-
ences. Relative pressure sensitivity in absolute units varies in Fig. 7.17 by
nearly 10:1 from the most sensitive regions (the face) to the least sensitive
(the foot).
Electrical sensitivity also varies considerably across different body parts,
as demonstrated in our experiments (Reilly et aI., 1982; Reilly and Larkin,
1984). In these experiments, sensitivity thresholds were measured for five
individuals at each of seven body loci: forehead, cheek (both right and left),
tip of third finger, thenar eminence (portion of the palm at the base of the
thumb), underside of forearm, back of the midcalf, and a point 3cm above
the ankle bone. The stimuli were discharges from a 200-pF capacitor. To
Body Location Sensitivity 275
Pressure Two-~oint Point localization

Middle finger Index finger
rn"~;;~k~~d
Upper lip Index finger Upper lip
Thumb Nose
___ Ring finger Middle finger
Belly Little finger Ring finger
Back
~~"iP Little finger
Ch~l
Shoulder Thumb
Little finger Nose Cheek
Ring finger Palm Hallux
Upper arm Hallux
[Fo~;h~ad P;'j
Forehead
---]
Middle finger
Forearm Sole Sole
Thumb Belly Belly
Breast Breast Shoulder
Thj
Index finger Forearm Forearm
Shoulder Calf
Palm Back Breast
~If Upper arm
SOI~ Back
Hallux Calf Thigh
] Items grouped within a bracket not significantly different.

---- All upper items significantly different from all lower items.
FIGURE 7.16. Rank-order of tactile sensitivity across various body parts, for three
measures of sensitivity. (From Weinstein, in D. R. Kenshalo, ed., The Skin Senses
1968. Courtesy of Charles C Thomas, Springfield, 11.)
mmlmlze the influence of concomitant tactile stimulation, a 1.6-mm-

diameter contact electrode was used, flush mounted in a large plastic
holder. At each body locus, the electrode contacted a slightly different spot
to avoid an average measurement that might have been unduly influenced
by spots of unusually high or low sensitivity.
Figure 7.18 illustrates the results, ordered in decreasing sensitivity from
left to right. Results have been normalized in relation to the fingertip
threshold for each subject. The fingertip was chosen for this purpose be-
cause it showed smaller variability than the other points, both among and
within subjects. Electrical sensitivity across body loci is seen to vary by
about 4: 1, with points on the face being the most sensitive, the lower leg the
least sensitive, and the finger about in midrange.
4
Males
---I--- - ....
.... ...
,,
~ 3 :,
, '-, , Females
... ...
... ", -- , I
,,
.....
-I---
I
ci , -,
g ,,
,,
,
....... -
.... ... I
I
I
I
:2 2 ,, , I
I
'-,
.EVl ,
,,
I
I
~ ,, ,,, .'
,,,
,,,
-=c: ,,, ,,,
,,
,
IV
,,, , ,,,
,,, ,, ,
I ::
CP
~
,, ,
,, ,,, ,, :~ , ,,
,, ,, .I"
J
x D I • ..,.
.2.9!
~c51 'Iii
0
i:
.S!'
F-
~
eZ j ~l
Vl:::l
~§'
Q.
:N ....
i:
i!=-::-I: I: I:
, Fi~~
0
FIGURE 7.17. Pressure sensitivity for males and females (Adapted from Weinstein,
in D. R. Kenshalo, ed., The Skin Senses, 1968. Courtesy of Charles C Thomas,
Springfield II.)
a.
--eCP
4.0 0
•• EMP}
WDL •
VTF Individual subjects
•• 8•
~
Cl
2.0
•
•
LGH 0__
,
:/
LBK
.9 0 Geometric Means
CP
>
~
~ 1.0 o • ...___0 0
.
:g
0
..c:
0.8 ~--j •
0
... ~/.
II)
CP 0.6
..c:
I-
fi •
0.4 •
0.2 •
Cheek Forehead Forearm Fingertip Thenar Calf Ankle
eminence
Body locus
FIGURE 7.18. Threshold data for seven body loci, normalized to the threshold for the
fingertip, flush-tip contact electrode, 200-pF capacitance. (From Reilly and Larkin,
1984.)
276
Skin Temperature 277
Five of the seven locations tested in this experiment were included in

Weinstein's study of cutaneous tactile sensitivity; a sixth, the thenar emi-
nence, can be compared with Weinstein's measurement on the palm. By
comparing Figs. 7.17 and 7.18, it can be seen that the rank ordering of
pressure and that of electrical sensitivity are similar. On the basis of these
results, it is reasonable to expect that electrical sensitivity at other points of
stimulation will correlate highly with tactile pressure sensitivity.
Although the ankle is about one-half as sensitive to electrical stimulation
as the fingertip, this difference should not be construed to mean that the
lower extremities are in all conditions protected from accidental exposure
that cannot be felt on the fingertip. There may be situations in which it is
natural to use some force in touching an electrical source with the fingertip,
but not with the ankle. In such cases, the force of the finger tip can mask an
electrical sensation (see Sect. 7.9).
We observe that the further the site of stimulation is from the brain, the
higher the threshold. Whether this observation implies some connection
with the synchronous timing of AP volleys, or differences in receptor type
or density at various body locations, can only be guessed. Lacking sufficient
histological and experimental data, any explanation for electrical sensitivity
differences at different body locations remains speCUlative.
7.8 Skin Temperature 3

It is not surprising that human sensitivity to electrical stimulation depends
on skin temperature, as the responsiveness of the underlying neural fibers
depends on their temperature. But there is a major difficulty in extrapolat-
ing from the temperature-dependent behavior of a single neuron to the
temperature-dependent behavior of human skin sensitivity: The skin, to-
gether with its blood supply, is a very efficient thermoregulator that protects
underlying tissues from the stress of changing temperatures at the body
surface. It is important, therefore, to measure human sensitivity under
controlled conditions of ambient temperature. Larkin and Reilly (1986)
performed experiments for this purpose, using an air-immersion technique
that mimics exposure to environmental heat and cold. The experiments
tested both perception thresholds and supra threshold responses for cap-
acitor discharges on the fingertip and on the forearm. An electrode-
tapping procedure was used for the fingertip, and a contact electrode was
used for the forearm; the stimulated arm was placed inside an insulated,
temperature-controlled chamber.
Perception threshold measurements, illustrated in Fig. 7.19, demonstrate
that the effects of skin temperature are similar on the fingertip and forearm.
3Portions at this section have been adapted from Larkin and Reilly (1986) and from
Reilly and Larkin (1985a).
0.7
~ Cooling Heating~
'C
0.6 2.0 .~
~ ~
~0.5 l"0
~o 1.5 :g
:g
o ~
"£i
~
0.4 -=~
.c ::J
t- en
en
~
a.
a.
0.3 1.0 t:
(JJ
Cl
c:
• Fingertip
u::
o Forearm
A Stevens (1980)
Skin temperature (OC)
FIGURE 7.19. Effect of skin temperature on cutaneous thresholds. Solid lines with
circles show detectability of electrical pulse on fingertip and forearm. Broken line
shows punctuate-pressure sensitivity on fingertip. (From Stevens, 1980, Fig. 1.)
Reference lines of constant slope are shown for QIO = 1.3 and 2.0. (From Larkin and
Reilly, 1986.)
Warming the skin affects sensitivity very little in either case, but there is a
drastic effect with cooling. The somewhat higher rate of climb for the
fingertip at low temperatures is consistent with its relatively lower priority
in thermoregulation compared with the forearm. Figure 7.19 also shows
normalized punctuate-pressure thresholds adapted from a study of tem-
perature effects on pressure thresholds (Stevens et aI., 1977; Stevens, 1980).
The temperature dependence of pressure and electrical thresholds is simi-
lar: Pressure sensitivity degrades if the skin is cooled, but shows very little
dependence on warming, up to 43°C.
In a suprathreshold procedure, subjects adjusted a stimulus on the cooled
or heated right hand to match the perceived intensity of a capacitor dis-
charge stimulus delivered to the left hand that was maintained at room
temperature. The stimuli presented to the left hand ranged from just per-
ceptible levels to the threshold of pain. Results of the sensation-matching
procedure are given in Fig. 7.20. The horizontal axis plots the charge deliv-
Skin Temperature 279
ered to the left fingertip. The vertical axis shows the stimulus needed on
thermally treated skin to achieve an equivalent level of sensation. The most
striking aspect of these data is that the matching charge falls along straight
lines parallel to the diagonal. Thus, the change in sensitivity is uniform over
the range of stimuli tested and can be represented by a multiplicative
constant for each temperature. The multipliers for the matching experiment
are listed on Fig. 7.20. They indicate that, for example, charge must be
raised by about 60% to reach the same sensation magnitude on the skin at
lOoe, compared with skin at normal temperature (30°C). These multipliers
fall very close to similar numbers derived from the threshold data in
Fig. 7.19. Thus, skin temperature affects perception sensitivity and
suprathreshold sensitivity nearly identically.
The fact that heating or cooling affects electrical sensitivity differently is
not surprising in view of the body's more efficient thermoregulation against
heat than against cold. The tolerance of pain to skin temperature occurs
around 45°C, at which point tissue damage begins to occur with prolonged
1.2 " . - - - -.....---.--~--r--r--.....___,,__,

+ polarity
1.0 1.25 cm 2 electrode
0.8
0.6
0.5
0.4
0.3 Multipliers:
40° 0.98
30° 1.00
20° 1.16
10° 1.60
0.2
0.2 0.3 0.4 0.5 0.6 0.8 1.0
Charge (Ilc) applied to skin at 30° C
FIGURE 7.20. Effect of heating and cooling the skin on suprathreshold sensation
levels. Measurements for capacitor discharges at 200pF to the fingertip. (From
Larkin and Reilly, 1986.)
heating (Chapter 10; Fig. 11.12), and nerve conduction is irreversibly

blocked (Klump and Zimmerman, 1980). On the other hand, the body can
withstand cooling well below normal temperatures. Below 30 DC, the skin
loses heat faster than it can be restored metabolically (Tregear, 1966). It can
therefore be expected that neural responsiveness would be affected by
cooling of the skin much more so than with heating.
The temperature dependence of neural activity is frequently reported in
terms of a QlO coefficient, which is the ratio of a given index of activity at
one temperature to that when the temperature is shifted by lODC. For a
variety of indices of neural responsiveness, QlO coefficients for isolated
neural preparations typically range from about 1.3 to 3.0 (Larkin and Reilly,
1986). The data of Fig. 7.19 indicate a QlO value of 1.3 for cooling on the
arm. On the fingertip, a QlO value around 1.3 applies to temperatures down
to about 14 DC, below which a value around 2.0 is indicated.
7.9 Tactile Masking

The mechanical stimulation of touch can have a masking effect on electrical
sensation. Most of us learn, for example, that a carpet spark is less bother-
some if we grasp the grounded object firmly, rather than tentatively. The
attenuating effects of contact force are well known among electricians and
others who are called upon to touch "hot wires." The explanation of this
phenomenon is not fully understood, but is likely to include the fact that
electrical and mechanical stimulation excite the same neural pathways in
the skin. At low voltages, people commonly report that capacitive dis-
charges feel very much like tactile pressure, mild pricks with a needle, or
pinpoint touches. Furthermore, we have seen that both tactile perception
and electrical perception share several features in common, including
similar rank-ordering with respect to body-part sensitivity, temperature
sensitivity, and male/female differential sensitivity.
We studied the relationship between tactile force and electrical threshold
by measuring perception thresholds when subjects tapped an energized
electrode with calibrated force (Reilly and Larkin, 1983). Subjects were
trained to tap with a force within 2dB of a target position on an accelerom-
eter scale. The scale ranged from OdB (the lightest force for repeated
tapping) to 50 dB. Repeated tapping at a force of 60 dB was painful. Steps of
10 dB on this scale represent approximately equal intervals of subjective
sensation, as well as of physical intensity. Thus, 50dB is a "strong" tap,
40dB is "firm," 30dB is "moderate," 20dB is "light," and lOdB is "very
light." Tap force was calibrated in a mechanical arrangement, in which a
force of 20dB was registered by dropping a steel ball with ll,uJ of kinetic
energy (Reilly et aI., 1982).
Figure 7.21 illustrates results from an experiment in which capacitance
was held fixed while tap force was changed randomly from one trial to
Tactile Masking 281
4.0 + Polarity
2.0
:;-
e.-
O>
OJ
(1) 1.0
0:::
0
> 0.8
:g
0
.c
...
<Il
0>
.c
0.6
I-
0.4 3200 pF
~ Average for 4 electrode types

0.2L-~ ______ ~ ______ ~ ______ ____- J_ _ _ _ _ _
~ - L_ _~
o 10 20 30 40 50
Tap force (dB re minimum force)
FIGURE 7.21. Effect of tap force on detection threshold for single-spark discharges
into the fingertip. Each symbol represents an individual subject. Except where
noted, a I-mm raised-tip electrode was used. (From Reilly and Larkin, 1985a.)
another. The middle set of curves represents results for four subjects using
a tip electrode (i-mm-diameter tip, elevated 0.5 mm above an insulating
holder) at 600pF. The threshold voltage at the highest tap force, 50dB, is
approximately double the voltage at the minimum tap force. A second set
of experiments was performed with capacitor discharges at 200 and
3,200pF. Results for one individual are plotted in the top and bottom of Fig.
7.21. The data represent trials using a variety of electrodes of different
shapes and sizes. Two general conclusions were made from these experi-
ments. First, the electrode size and shape have almost no effect in the
tapping procedure. Second, the masking effect of tap force is greater for
electrical stimuli with shorter durations. The masking of electrical thresh-
olds by touch can be large and should not be discounted in any assessment
of human sensitivity. Accordingly, tap force should be monitored or con-
trolled in any test involving active touching.
Rollman (1974) tested the masking effects of a conditioning pressure

stimulus on electrical thresholds and the reverse (electrical masking of
pressure stimulus). The degree of masking depended on the time sequence
of two stimuli. For coincident pressure and electrical stimuli matched to the
same subjective intensity (determined by cross-modal matching to an au-
dible tone), Rollman reported a 1- to 3-dB threshold shift in the electrical
stimulus if the mechanical stimulus is the conditioning event; for the reverse
situation (detection of pressure, masking by electrical), the pressure thresh-
old was shifted by 12dB.
A concurrent pressure stimulus has been observed to raise thresholds of
painful60-Hz electrical stimulation (Higgins et aI., 1971). This observation
was offered as evidence supporting Melzack and Wall's (1965) so-called
"gate theory" in which stimulation of myelinated fibers is said to inhibit
pain perception from nearby unmyelinated nociceptors.
7.10 Individual Differences in Electrical Sensitivity4

It is not uncommon for one person's detection threshold to be as much as
twice that of another person, although the measurement procedures are in
all respects identical. These individual differences persist from day to day
and from week to week; they are not temporary artifacts of the measuring
procedure. In spite of these differences, electrical sensitivity data exhibit a
high degree of parametric invariance, in the sense that curves of sensitivity
(for example, psychometric functions or equal-sensation contours) com-
monly have the same shape regardless of who is picked as the test subject.
Curves for different test subjects can frequently be related to One another
by a simple multiple. Consequently, basic studies of sensitivity need not
require a large number of participants. However, none can serve as a basis
for estimating the distribution of sensitivity variation in the general popula-
tion. For that purpose, a large sample of test participants is needed.
Potential Sources of Variability

Variations in sensory measurements fall into three general categories: trial-
to-trial variations, long-term sensitivity drifts, and subject-to-subject differ-
ences. Trial-to-trial variations in thresholds are observed when an
individual is tested repeatedly in a single session. For capacitor discharge
stimuli, these variations depend on the method of applying the stimulus
(Larkin and Reilly, 1984). For stimuli delivered by actively tapping an
energized electrode with the fingertip, perception thresholds exhibit a coef-
ficient of variation (ratio of standard deviation to mean) of about 0.12; for
4Portions of this section have been adapted from Larkin and Reilly (1984).
Individual Differences in Electrical Sensitivity 283
capacitor discharges delivered to an electrode held in contact with the

skin, the observed coefficient of variation is about 0.05. In the contact-
electrode case, the coefficient of variation is comparable to that observed in
other studies using constant-current pulses (Rollman, 1974). In the tapped
electrode case, the electrical sensation can be masked by tactile stimulation,
raising the threshold of detectability and increasing its variance. Gradual
threshold drifts reflect a second source of variability: Even highly practiced
subjects do not give precisely the same results from day to day. We exam-
ined individual sensitivity drifts in a longitudinal study in which four
subjects were tested once weekly over a period of 8 weeks (Reilly et aI.,
1984). Treating the session means for each subject as a statistical variable,
the coefficient of variation across sessions averaged 0.11 for the four sub-
jects. These long-term drifts cannot be separated from intersubjective
comparisons without enormous experimental effort. Such an effort
would not be repaid by any substantial increase in the precision with which
two individuals can be compared, because the drifts appear to account
for only a small fraction of the total intersubjective variance. Individual
differences in sensitivity are a much greater source of variability in
sensory measurements. These differences persist over time and greatly
outweigh the fluctuations observed from trial to trial or from one week to
the next.
Elevated sensory thresholds exist in individuals with neuropathological
conditions (Friedli and Meyer, 1984; Katims et aI., 1986; Notermans, 1967;
Pruna et aI., 1989), as well as psychopathological conditions (Hare, 1968).
Neuromotor excitation threshold shifts also exist in individuals with
neuropathy (Parry, 1971; Harris, 1971). Significant variations from normal
populations have been observed in the thresholds of electrocutaneous
perception in uremic (Katims et aI., 1991) and HIV-positive individuals
(Taylor et aI., 1992). Among individuals who appear to be normal, wide
variations in electrical sensory thresholds are also found. The basis for these
variations is not well understood.
Distribution of Sensitivity
We studied electrical sensitivity of 124 subjects (Larkin et aI., 1986). Rather
than make an extensive study of people chosen completely at random-a
practical impossibility in any case-the study focused on specific subgroups
whose sensitivity might be expected to be either high or low. Subgroups
consisted of college students (25 women and 24 men), 25 female office
workers, and 50 men engaged in skilled trades (electricians, carpenters,
plumbers, and sheet-metal workers). The sampling strategy allowed direct
comparisons between men and women and between occupational groups.
Among the groups, there was a wide distribution of ages and body sizes, so
that a statistical estimate could be made of their relationship to electrical
sensitivity. A variety of physical attributes was quantified; namely, age,
height, weight, temperature of fingertip and forearm, circumference of

fingertip and forearm, index of habitual physical activity, and index of prior
shock experience.
Statistical distributions of sensitivity were determined separately for the
subgroups. Geometric means are summarized in Table 7.8. One indication
is that there are sensitivity differences related to sex and occupation: males
appear to have higher thresholds than females, and maintenance men have
higher thresholds than college men. As shown below, the apparent group
differences appear to be artifacts of a body-size dependency.
Table 7.8 also provides the geometric means of ratios among individuals'
sensitivity values. The column labeled BfA gives the ratios of perception for
finger versus arm stimulation. This ratio not only reflects the greater sensi-
tivity of the arm relative to the finger (see Sect. 7.7), it also arises because
of tactile masking and electrode size effects. The effect of tactile masking is
to elevate thresholds for fingertip stimulation relative to a contact electrode
condition (see Sect. 7.9). The combined effects of these factors result in an
average fingertip-to-arm threshold ratio (BfA) of 2.9. The column labeled
CIB indicates the multiple above the perception threshold needed to pro-
duce the rating "definitely annoying." These "perception-to-annoyance"
ratios agree well with results from other experiments designed specifically
to study suprathreshold reactions (Sect. 7.4). The column labeled CID
indicates the relative sensitivity of 200- versus 6,400-pF discharges from a
TABLE 7.8. Geometric averages and ratios of thresholds from the large sample
study.
Perception thresholds Annoyance levels
(A) (B) (C) (D)
Arm-contact Fingertip Fingertip Fingertip
electrode tapping tapping tapping Ratios
Group N 200pF 200pF 200pF 6,400pF BIA ClB CID
Female 25 0.083 0.252 0.576 1.91 3.0 2.3 9.7
students
Male 24 0.107 0.277 0.664 2.14 2.6 2.4 9.9
students
Female 25 0.074 0.266 0.648 1.95 3.6 2.4 10.7
office
workers
Male 50 0.113 0.314 0.924 2.94 2.8 2.9 10.0
maintenance
workers
Total 124 0.097 0.284 0.734 2.33 2.9 2.6 10.1
sample
Notes: Data given in #C of charge for positive-polarity stimulation; values of capacitance are
listed at heads of columns.
0.98
Ul
::J
"S 0.95 (b)
E
(a) I fingertip /
oel
.~
Q) 0.90 Perception Perception 0/

-=S2 forearm I ,/
0.80
OJ
c J tI
J'
'6
c
8. 0.60 o /
,
Ul o
....
Q)
I
Ul
(ij 0.40 F~ M
~
F/
f J
::J
If 0/
0,'
"C
:~ (c)
"C
/ Annoyance
, ..
.!: 0.20
I
"0
k F/ fingertip
.~ 0.10 J ""
of /0 •
1:
o
g. 0.05 I I /o
a: o •
0.02 '--_.L.---L.--'---'---'-.L....L-'--_--'----'_ _.L.---L.---'---'--'-...J
0.2 2.0 4.0 6.0 8.0
Stimulus voltage (kV)
FIGURE 7.22. Cumulative probability that a capacitive discharge (+polarity, 200pF)

exceeds perception or annoyance thresholds; samples of 74 men (M) and 50 women
(F). Discharges on the fingertip via a tapped electrode; discharges on the arm via a
O.5-cm-diameter contact electrode. (From Larkin et aI., 1986.)
tapped electrode: On the average, 3.2 times as much charge is required at

6,400pF to produce the same annoyance rating obtained at 200pF. This
capacitance effect stems from differences in discharge time constants at the
different capacitance values (see Sect. 7.3).
Threshold distributions are well represented by the log-normal distribu-
tion, both for subgroups and for pooled data. Figure 7.22 gives an example
of sensitivity distributions for 200-pF discharges, separated into male and
female subgroups. Voltage levels are indicated on the horizontal axis. The
vertical axis plots the relative number of individuals reporting perception
[curves (a) and (b)] or annoyance [curves (c)]. The horizontal axis is loga-
rithmic, and the vertical axis is subdivided according to the standard
Gaussian probability distribution. The data are nearly linear on these plots,
indicating the log-normal distribution. The straight lines through the data
are minimum chi-square fits to the log-normal model. This depiction omits
some outliers and therefore slightly underrepresents the total variability in
the sample. The voltage ratio spanning the middle 90% of the fingertip
perception thresholds (that is, from the 5th to the 95th percentile) is ap-
proximately 2: 1. The ratio is closer to 3: 1 in the case of forearm thresholds
and the annoyance ratings.
As explained below, electrical thresholds are correlated with measures of

body size. Consequently, subgroups that are correlated with body size will
exhibit a threshold distribution that is less dispersed than a more heteroge-
neous group, as illustrated in Table 7.9 by the data of Larkin and Reilly. For
comparison, statistical parameters from other experimenters are also
shown. In the first subgroup of the table, adult males and females are
combined in presumed equal numbers. The second subgroup, applies
to either males or females. Although these subgroups are associated
with significantly different thresholds, the distributions are virtually indis-
tinguishable when their thresholds are normalized with respect to their
respective medians (see Fig. 7.22). The third subgroup, assumed to be of
homogeneous body weight, has weight-adjusted thresholds in accordance
with a regression formula presented below.
For analysis of electrical acceptability, we would like to know the distri-
bution of thresholds across a population of exposed individuals. For ex-
TABLE 7.9. Statistical distributions of sensory thresholds.

Normalized
threshold
Body locus Subgroup N Stimulus Measurement 10% 90% Ref.
Forearm M+F 124 200-pF cap. Perception 0.54 1.67 a
discharge
Fingertip M+F 124 200-pF cap. Perception 0.71 1.43 a
discharge
Fingertip M+F 124 200-pF cap. Annoyance 0.70 1.81 a
discharge
Forearm MorF 74,50 200-pF cap. Perception 0.61 1.52 a
discharge
Fingertip MorF 74,50 200-pF cap. Perception 0.78 1.32 a
discharge
Fingertip MorF 74,50 200-pF cap. Annoyance 0.74 1.55 a
discharge
Forearm Weight- 124 200-pF cap. Perception 0.73 1.36 a
adjusted discharge
Fingertip Weight- 124 200-pF cap. Perception 0.79 1.26 a
adjusted discharge
Forearm M+F 40 1-ms pulses Perception 0.34 1.69 b
@10pps
Forearm M+F 40 1-ms pulses Pain 0.43 1.57 b
@ 10pps
Hand (grip) M 115 Continuous Perception 0.63 1.50 c
DC
Hand (grip) M 115 60-Hz AC Perception 0.73 1.27 c
Thresholds normalized by median value.

References: (a) Larkin and Reilly (1984); (b) Rollman and Harris (1987); (c) Dalziel and
Mansfield (1950b).
TABLE 7.10. Normalized perception threshold distribu-

tion for log-normal model of adult population.
Percentile M+F Weight adjusted
rank (%) Forearm Finger Forearm Finger
99 3.11 1.91 1.78 1.51
95 2.24 1.58 1.50 1.33
90 1.85 1.41 1.36 1.25
75 1.40 1.21 1.18 1.12
50 1.00 1.00 1.00 1.00
25 0.72 0.83 0.85 0.89
10 0.54 0.71 0.73 0.80
5 0.45 0.63 0.67 0.75
1 0.32 0.52 0.56 0.66
0.5 0.29 0.49 0.53 0.63
Note: Data express multiples of median threshold.
ample, one may wish to select a criterion of exposure that protects some
percentile rank of sensitive individuals. Table 7.10 indicates thresholds for
a representative log-normal model, based on the author's measurements.
The data listed under (M + F) apply to a group of adults, with males and
females in equal proportion. The "weight-adjusted" data apply to calcu-
lated thresholds, using body-weight regression formula. The data in
this table have been normalized by the median values of the respective
subgroups.
Correlates of Sensitivity
The cumulative distributions plotted in Fig. 7.22 show an unmistakable
difference between the two sexes. For a homogeneous group of students,
for example, the forearm thresholds for males is 1.3 times that for females
(column A of Table 7.8). The direction of this difference is not surprising in
view of the research literature on sex differences in cutaneous sensitivity.
Women reportedly are more sensitive than men to pressure (see Sect. 7.7),
to vibratory stimulation (Verillo, 1979), and to pain (Notermans and
Tophoff, 1975). Women are also reported to be more sensitive to a variety
of electrical stimuli, including pulse trains (Rollman and Harris, 1987),
60-Hz currents (Dalziel, 1972), and sinusoidal currents of frequency from
104 to 107 Hz (Chatterjee et aI., 1986). The reported ratio of the thresholds
for males versus females is typically consistent with our findings (about
1.3:1).
To my knowledge, no physiological hypothesis has ever been advanced to
account for these differences, although there have been suggestions that
cultural and social factors may account for them-especially in the case of
pain tolerance. Our data do not support a cultural explanation, since detec-
tion thresholds were determined in a manner that effectively eliminates all
sources of bias in response tendencies, and the ratio between annoyance

level and threshold level is remarkably constant across the two sexes (see
the column labeled CIB in Table 7.8).
The data in Table 7.8 also suggest that higher electrical thresholds might
be associated with occupational groups involved in strenuous labor. There
is a common belief that those who lead more strenuous lives acquire a more
stoic or tolerant attitude toward physical discomfort and pain. The physical
demands of daily life may contribute to this effect if skin sensitivity is
altered by injury to the epidermis or by the development of hard callused
layers. Whether pain tolerance is culturally acquired, as many have sug-
gested (Zborowski, 1952; Sternbach and Tursky, 1965; Craig and Weiss,
1971; Wolff and Langley, 1975; Clark and Clark, 1980), or reflects an under-
lying physiological adaptation, is not at issue in the present study. We might
have expected college students and office workers to differ greatly from
electricians, carpenters, and sheet-metal workers, who are more frequently
exposed to strenuous physical labor and harsh weather. If the threshold and
annoyance levels of the maintenance workers are compared with those of
the other groups (Table 7.8), there is a difference in the expected direction.
However, the difference is nearly as large for forearm sensitivity, where
calluses do not develop, as it is for the fingertip. This observation suggests
that skin thickness alone does not account for the result.
Much the same conclusion applies in the case of physical exercise and the
subjects' reported levels of previous exposure to electric shock. Neither of
these factors is well correlated with sensitivity; they do not explain why
maintenance workers have higher thresholds and annoyance levels than the
other subgroups.
It might seem that a factor related to sex and occupation could partially
account for individual sensitivity differences. But multiple regression analy-
sis shows that the apparent sex and occupation differences appear to be
artifacts of an underlying body-size dependency: Woman in our sample
group tended to be smaller than the men, and maintenance workers were
larger than college men (Larkin and Reilly, 1986). When these size differ-
ences were taken into account, no statistically significant differences in
sensitivity remained. A regression equation based on body weight was
determined:
(7.7)
where V is the predicted median perception or annoyance voltage, W is the
subject's weight in kilograms, and a and b are empirical parameters. Best-
fit values for a and b that apply to discharges to the fingertip, from a 200-pF
capacitor were, for perception and annoyance, respectively, a = 971,1904;
and b = 0.0053, 0.0093. Statistical variations about the median indicated by
Eq. (7.7) follow the log-normal distribution, with a 90th percentile value at
a factor of approximately 1.5 times the median and a 10th percentile at
approximately 0.7 times the median.
Equation (7.7) provided a better overall fit (greatest r) to the data than
the other two-parameter functions. However, from theoretical consider-
ations (Schmidt-Nielsen, 1984), one might expect a power function of body
weight. In this case we can represent the data by
V= cW (7.8)
where c and d are empirical parameters. The least-squares fit of Eq. (7.7)
was only marginally better than that of Eq. (7.8). The estimated parameter
values that apply to discharges to the fingertip, from a tapped 200-pF
capacitor, were for perception and annoyance, respectively: c = 204, 121;
and d = 0.386,0.678. With 200-pF discharges to a 0.5-cm-diameter electrode
held against the arm, perception parameter values were c = 13.8 and d =
0.702.
Table 7.11 illustrates weight-dependent fingertip thresholds for dis-
charges of 200pF based on Eqs. (7.7) and (7.8). The two formulations
closely agree over the weight range ofthe test subjects (41 to 128kg). Note
that the dynamic range of sensitivity, represented by the perception-to-
annoyance multiple, increases from low to high body weights.
Downward extrapolation of Eqs. (7.7) and (7.8) yields rather different
estimates of sensitivity at the lower body weights. A representative weight
for children (23kg) is included in Table 7.11. In this case, the power func-
tion gives a value 16% lower for perception and 26% lower for annoyance,
compared with the exponential function. Without further testing, it is not
possible to make a secure choice between these alternatives nor to be
confident that extrapolation to children is valid.
The regression formulas presented above used body weight as a matter of
convenience. Other relationships with body size (e.g., forearm diameter)
yield equally valid correlates of sensitivity, since various measures of body
size are strongly correlated with one another. Based on these findings, it
TABLE 7.11. Calculated thresholds for various body-

weight classes, discharge to the fingertip from a 200-pF
capacitor, tapped electrode delivery.
Annoyance
Weight Perception Annoyance multiple
(kg) (lb) (7.7) (7.8) (7.7) (7.8) (7.7) (7.8)
23 50 0.22 0.18 0.47 0.35 2.15 1.87
45 100 0.25 0.24 0.58 0.55 2.45 2.29
68 150 0.28 0.28 0.72 0.73 2.57 2.58
91 200 0.31 0.31 0.88 0.88 2.81 2.81
113 250 0.35 0.34 1.09 1.03 3.11 3.00
Thresholds are expressed in flC of charge. Corresponding

voltages are calculated according to Eq. (7.7) or (7.8) as
indicated in column head.
Source: From Reilly and Larkin (1987).
would be interesting to examine whether male/female differences in tactile

pressure sensitivity might also reflect an underlying body-size effect. The
fact that pressure sensitivity of women's breasts is more closely related to
the inverse of breast size, rather than to the overall size of the women
(Weinstein, 1963), suggests that the size of the stimulated body part is
relevant.
An intriguing question remains: why is electrical sensitivity related to
body size? Conceivably, the relationship might be explainable if cutaneous
receptor density were inversely related to the size of the individual. How-
ever, until further histological and experimental evidence becomes avail-
able, the etiology of the body-size role in electrical sensitivity can only be
guessed.
7.11 Startle Reactions5

Muscle movement that is initiated through reflex activity involves both
sensory and motor responses (see Sect. 3.6). Reflex activity operates out-
side conscious control to regulate many vital body functions. When this
activity is in response to a surprise sensory insult, the reaction is frequently
called a "startle reaction." A startle reaction to electrical stimulation could
potentially pose a hazard to an individual who was engaged in certain
inherently dangerous activities, such as climbing a ladder, handling hot
liquids, or handling power cutting tools.
It is very difficult to design a meaningful experiment to define startle
reactions. One difficulty arises from the fact that it is necessary to introduce
the element of surprise in the context of informed consent on the part of
experimental subjects. Expectations in an experimental setting are likely to
introduce a bias in the measured reaction threshold or the extent of the
reaction.
Published data on startle reactions to electrical stimulation were ob-
tained from experiments carried out at the Underwriters Laboratories
(UL) in connection with the development of ANSI standards on leakage
current (Smoot and Stevenson, 1968a and 1968b; Stevenson, 1969). These
studies were designed to help define levels of current that could pose a
potential hazard because of a startle reaction. Men and women volunteers
were asked to transfer rice from one container to another using a small
metal cup held in the fingers. Random levels of current were applied at
unexpected times to various parts of the hand, wrist, or lower arm, including
the small cup itself. The return electrode was through the other arm im-
mersed in salt water. Subjects' reactions were given numerical ratings ac-
cording to the distance that the hand moved during a startle reaction, the
speed of movement, and the amount of rice spilled. A panel consisting of
5 This section has been adapted from Reilly (1978b).

Electrical Stimulation of Domestic Animals 291
approximately 240 people viewed the volunteers' reactions by videotape

and made judgments as to what reaction might pose a hazard in the home,
such as because of spillage of hot grease or falling from a ladder. Most panel
members were engineers and scientists concerned with electrical safety,
although 11 homemakers were also represented on the panel. A borderline
hazardous reaction was typically judged as one corresponding to between
3 and 12 in. of motion at a moderately fast rate, with some rice spillage
(usually less than 10%).
Initial tests established that women were somewhat more sensitive than
men. As a result, succeeding tests were carried out with women to establish
a conservative estimate of hazardous thresholds. Based on 20 women
tested, the average reaction judged as borderline hazardous was elicited at
a conducted current6 of approximately 2.2 mA when applied to the lower
arm and 3.2mA when applied to the cup held between the fingers. These
average reaction currents may be compared with the median perception
levels for women of 0.73mA for gripped conductors, 0.59mA for pinched
contacts (Thompson, 1933), and 0.24mA for lightly touched contacts
(Dalziel, 1954). Thus, it appears that average reaction currents are at least
a factor of 3 above average perception currents. Such a threshold multiple
would place the startle threshold stimulus in the painful affective category,
as can be seen from the data of Table 7.3.
Based on the UL reaction tests, a limit of 0.5 mA was adopted as the
ANSI standard for leakage current from portable appliances (see Chapter
11). We would like to determine the probability that the specified leakage
might cause a reaction fitting into the "borderline hazardous" category
defined by the UL observers. Lacking the statistical distribution of reaction
thresholds, we assume that the distribution of reaction thresholds has the
same coefficient of variation (standard deviation to mean ratio) as do
perception thresholds. The coefficient of variation of the 60-Hz perception
threshold for touched or gripped contacts is about 0.37 (Dalziel, 1954).
Applying that value to a normal distribution for reaction currents along
with the mean reaction values mentioned above, it is estimated that an
unexpected current of 0.5 mA applied to a woman's lower arm will have a
2% probability of causing a reaction fitting into the UL "borderline hazard-
ous" category. If the contact is pinched between the thumb and fingers, the
probability is 1 %.
7.12 Electrical Stimulation of Domestic Animals

Farmers are becoming increasingly aware of the phenomenon of stray
voltage, a term used to describe unintended electrical potentials between
contact points that may be accessed by a human or farm animal. Stray
6Because of body impedance, conducted current was on the average 0.72 times the
available short-circuit current.
voltage of sufficient magnitude can cause unpleasant shocks to farm ani-

mals, which can ultimately lead to animal handling problems. Much of the
attention to stray voltage problems has focused on dairy cows, for which
potentials of only a few volts at 60 Hz can be unpleasant. With consider-
ation of such low potentials, it is not surprising that stray voltage can be
a real problem on some farms. Consequently, much research has been
devoted to understanding the conditions under which dairy cows may
be disturbed by low-level electrical stimuli. A useful summary of some of
this research has been given by Lefcourt (1991).
Researchers typically apply 60-Hz currents to electrodes contacting
various body surfaces of the animal. Locations of contact electrodes may
vary considerably in given situations. The electrical impedance model of
Fig. 2.41 can be used to determine median cow impedance for a variety
of exposure conditions.
It is not possible for a cow to directly tell us when she can perceive a
current as can humans. Instead, one must infer perception from the
animal's behavior. Often an experimenter tests the minimum current that
elicits some reaction. If the experiment is carefully designed, the minimum
reaction current can approach the animal's threshold of perception.
Reactions to 60-Hz Stimulation

Table 7.12 summarizes experimental data on cow perception of 50- or 60-
Hz current (Reilly, 1994). The upper section of the table lists minimum
reaction thresholds (i.e., the just-noticeable response of the animal, such as
lifting a foot, vocalization, sudden movement, or opening the mouth). Ani-
mals tested by Norrell, Currence, and Gustafson were trained to react to
break the current. It is possible that these trained animals may have re-
sponded to just-perceptible current. The other reaction thresholds may lie
somewhat above the minimum perception thresholds.
The lower part of Table 7.12 applies to avoidance reactions. In these
tests, cows were given the opportunity to receive a food treat after 25 to 30
presses of a lever with the muzzle, each press resulting in an electric shock.
The listed current is the minimum value that suppressed the plate-pressing
activity.
In Table 7.12, median minimal behavior thresholds for muzzle contact
range from 2.5 to 4.8mA with an average of 3.9mA; for contact by other
body regions, generally greater thresholds are observed. For the 10% most
sensitive cows, reported thresholds are in the range 1.0 to 3.2mA. It is clear
from the avoidance threshold data that the minimum listed currents are
not particularly aversive. Norell's cows, for instance, were initially not
willing to endure 30 shocks of 1.53 mA to the muzzle to receive a food
treat. But after only two sessions, the same cows were willing to accept 30
shocks of 3.63mA. In this context, current less than 3.63mA to the muzzle
is seen as only a mild deterrent. Whittleson's cows were willing to endure
TABLE 7.12. Bovine Electrical Thresholds: 60/50-Hz Current.

Threshold (rnA)
@ given
percentile
N(a) Current Path Reaction 10% 50%(b) Ref. Notes
Minimal Behavior Thresholds
7H mouth/4 hooves Open mouth to break 1.2 2.7 [a] c
current
2 front/2 rear Lift foot to break 2.4 3.7
hooves current
6H mouth/4 hooves Open mouth to break 1.1 4.2 [b] c
current
2 front/2 rear Raise hoof to break 1.0 3.3 f
hooves current
shoulder/4 Raise hoof to break 2.3 9.0 g
hooves current
18H,3G 1 fron tl2 rear Any (twitch, lift foot, 2.5 3.5 [c] d
hooves hump back, sudden
movement)
5H rt front knee/rt Mild reaction (flinch, 2.5 [d] e
rear hock vocalization)
distinct reaction 3.5
(startle)
120 nose c1ip/4 Mild reaction (twitch, 3.2 4.8 [f]
hooves leg move)
A voidance Thresholds
6H muzzle/4 feet Suppress 30 plate [a] h
presses
session 1 1.53
session 2 2.85
session 3 3.63
7J rt front teatl4 Suppress 25 plate 7.1 [e] h, i
hooves presses
4 teats/4 hooves Suppress 25 plate 5.6
presses
rump/4 hooves Suppress 25 plate 6.1
presses
chest/4 hooves Suppress 25 plate 4.0
presses
Notes:
(a) Number of subjects; H = Holstein; G = Guernsey; J = Jersey.
(b) 10 and 50 percentiles given where available; otherwise, mean substituted for 50%.
(c) Experimental percentile adjusted for null response.
(d) Current duration t = 0.17, 18.
(e) EKG electrodes + electrode gel on shaved leg.
(f) Shoulder electrode = 60 cm2 with EKG gel.
(g) 50% estimated by extrapolation.
(h) 30 (or 25) plate presses required for food treat-each plate press causes shock.
(i) 50-Hz current used.
(D Transient thresholds adjusted for continuous 60Hz, rms.
References:
[a] Norell et al. (1983); [b] Gustafson et al. (1985); [c] Currence et al. (1990)
[d] Lefcourt (1982); [e] Whittleson et al. (1975); [f] Reinemann et al. (1997).
even greater current levels to other body parts for the privilege of a food
treat.
Reactions to Transient Stimulation

Despite the fact that most cow research has used 60-Hz stimulus currents, a
detailed examination often reveals a highly complex voltage or current
waveform consisting of a distorted 60-Hz sinusoidal variation, on which is
superimposed short-duration transients. These transients may be produced
when on- or off-farm equipment is switched on or off. Transients may be
brief in duration, but can exceed the amplitude of the more persistent 60-Hz
background and display a wide variety of waveform characteristics with
various patterns of oscillation and repetition.
A number of researchers have tested bovine reactions to transient or
intermittent currents (Gustafaon, 1988; Currence et aI., 1990, Reinemann
et aI., 1994; Reinemann et aI., 1996; Reinemann et aI., 1998). In the experi-
ments of Reinemann and colleagues, Holstein cows received electrical
stimuli through nonpiercing ball-end clips attached to the cow's nose; the
return path was through the four feet. Electrical thresholds were evinced
104
Monophasic
Sinusoid Biphasic Sinusoid
103
~
.s ~D
c:
...~
::::J
0
"'C 102
/-
"0
..r:: Monophasic
Ul
~ Square
-=
~
I'll
Q)
a.. 10
Stimulus phase duration (I1S)
FIGURE 7.23. Strength-duration data for reaction thresholds of dairy cows. Curve
applies to single sinusoidal cycle; points apply to monophasic stimuli. (From
Reinemann et aI., 1996).
.0
.........
=
...,.n.·······
.' .' .'
.' .'
.... c···· multi cylce
0··········
10 10 3
Stimulus frequency (Hz)
FIGURE 7.24. Strength-frequency data for reaction thresholds of dairy cows. Curves
apply to single or multiple cycle sinusoidal currents. (From Reinemann et aI., 1996.)
through changes in behavior (leg motion, facial twitch, or tail switch) or

through animal movement recorded by a load cell within the floor of the
stall. Electrical transients consisted of monophasic square wave stimuli or a
single cycle of a sinusoidal current.
Figure 7.23 shows average strength-duration data for 120 cows for a
monophasic square wave (single point), a monophasic half-cycle sine wave
(single point), and single biphasic cycle of a sine (curve). The horizontal
axis gives the "phase duration," which is the duration of a single mono-
phasic phase of the stimulus, as indicated in Fig. 4.16. Using Eq. (4.20)
along with the charge threshold of the monophasic square wave and the
rheobase current (at t = lOms), we infer from these data a strength-dura-
tion time constant Te = 200,us. This value is similar to that found in human
sensory experiments (Table 7.1). The experimental data agree well with the
theoretical SENN model (Fig. 4.16), except that the time scale of the
experimental curve is shifted to a longer strength-duration time constant-
an observation that is often found with human sensory data.
Figure 7.24 shows biphasic thresholds on a strength-frequency format.
Thresholds are shown for a single cycle of a sine wave and for a continuous
sine wave. These data are similar to the theoretical SENN model data
shown in Fig. 4.21, except that the experimental curve is shifted to lower
frequency values as compared with the theoretical model. This frequency
shift is a consequence of the temporal shift in the strength-duration curve as
noted above.
Statistical Distribution of Bovine Reaction Thresholds

Reaction thresholds were obtained for 120 cows with stimulation by a
single-cycle 60-Hz current (Reinemann et aI., 1998). The cumulative
frequency distribution is shown in Fig. 7.25. On this plotting format, the
log-normal distribution would appear as a straight line. The data
conform well to the log-normal distribution between the 5- and 99-percen-
tile ranks. It would require a larger sample to determine whether conform-
ance to this statistical model exists over a wider percentile range. The range
in cow thresholds from the 90- to 10-percentile ranks encompasses a ratio of
2.37. Similar results have been obtained in tests of human perception of
transient currents, for which a log-normal model also applies, with the
range from 90- to lO-percentile thresholds encompassing a ratio of 2.0 for
stimulation of the fingertip and 3.4 for stimulation of the forearm (Table
7.10).
In comparing our results with other published data, note that Figure 7.25
represents peak currents for transient stimulation, whereas most other
researchers report rms values for continuous 60-Hz current. The median
reaction threshold in Figure 7.25 is 9.0mA-peak. If we adjust that value to
an rms metric (multiplier = 0.707) and further adjust from single cycle to
continuous stimulation as indicated in Fig. 7.24 (multiplier = 0.75),
we would infer a median threshold of 4.8mA-rms for continuous 60-Hz
stimulation.
Magnetic Field Stimulation of Cows

One might ask: If stray voltage of only a few volts can cause animal behav-
ior problems, might not low-level electromagnetic fields on the farmstead
also cause reactions in farm animals and result in behavior problems?
Indeed, public agencies including Public Service Commissions, governmen-
tal groups, and the courts have confronted such questions because of law
suits brought on behalf of farmers (primarily dairy) who have alleged
disturbing perception of extremely low-frequency fields by their farm ani-
mals. To address these concerns, a theoretical study was conducted on the
conditions necessary to excite peripheral nerves of cows via time varying
magnetic fields (Reilly, 1995). Using models of magnetic excitation that are
developed in Chapter 9, this study showed that stimulation of peripheral
nerves of a 1300-lb cow would require a magnetic flux density of 54mT at a
frequency of 60 Hz over the entire body of the animal. Such a field is several
99.99
99.9
99
95
90
Cl
c 80
'6
c 70
0
Cl.
(J)
~ 50
'E
II)
e
II)
30
a. 20
10
5
.1
.01
10 30
Threshold current (rnA - pk)
FIGURE 7.25. Cumulative frequency distribution for reaction threshold of 120 diary
cows for single-cycle, 60-Hz transient current. Threshold current refers to the peak
of the stimulus waveform where an animal responded.
orders of magnitude greater than what can be encountered in the home or

on the farmstead.
Cow magnetic thresholds can be approximately scaled to other animals
using the body weight adjustment of Eq. (9.22). In the cow example pre-
sented here, the body weight is 1300 lb. For a 180-lb human, for instance,
the body weight scaling factor relative to a cow is 1.93. Using that factor as
a multiple of the cow threshold, we arrive at an extrapolated threshold
of 104 mT for the human; that estimate does not differ greatly from the
value obtained with a more exact analysis based on the human body shape
(Fig. 9.17).
7.13 Visual and Auditory Effects

Other sensory responses to electrical stimulation include visual and audi-
tory effects, which have been studied using both contact electrodes and
electromagnetic exposure. These effects, which are treated in detail in Sect.
9.8, will be mentioned in brief here.
It is possible to stimulate auditory channels intentionally for application
to cochlear prostheses or unintentionally in the case of accidental exposure
to electric current or electromagnetic fields.
Auditory effects from pulsed microwave stimuli involve the absorption of
microwave energy within the brain when exposed by an electromagnetic
field (Fig. 9.22). The absorbed energy launches an acoustic wave through
thermoelastic expansion of brain tissue, and the acoustic wave is conveyed
to the inner ear through bone conduction within the skull.
Auditory sensations can be elicited with audio-frequency currents
through electrodes placed on the head (Fig. 9.21). Auditory thresholds
obtained in this manner are near neuromuscular thresholds. A plausible
explanation for these auditory effects is that excitation of muscle units
convey vibratory stimuli to the inner ear through bone conduction.
Visual sensations called phosphenes can be produced by magnetic fields
or electric currents applied to the head (Fig. 9.20). The rheobase E-field
threshold for phosphenes is approximately a factor of 100 below that for
action potential excitation; S-D time constants for phosphenes are approxi-
mately 100 times greater than corresponding values for nerve excitation
and 10 times greater than those for excitation of muscle tissue (Sect. 9.8).
The responsible mechanism is thought to involve the polarization of
neurosynaptic processes through electric fields within the retina (Sect. 3.7).
Electrophosphenes are of interest in the development of acceptability
criteria for electrical exposure because the underlying neurosynaptic inter-
actions are also found within the central nervous system (Sect. 11.2).
8
Skeletal Muscle Response to
Electrical Stimulation
JAMES D. SWEENEY
8.1 Introduction
In the eighteenth and nineteenth centuries, many seminal investigations
into the basic properties of electricity were electrophysiological in nature
and focused on neuromuscular phenomena. In Galvani's commentary
(1791), a frog nerve-muscle preparation was used to propose a theory of
"animal electricity" in which muscle fibers were thought to store charge.
Volta (1800) later demonstrated that the bimetallic electrodes used by
Galvani were a source of electrical stimulation current rather than a path-
way for discharge of animal electricity. In modern society, electrical stimu-
lation of skeletal muscle may occur unintentionally as one possible effect of
an accidental electric shock or intentionally in medical devices for artifi-
cially induced muscle exercise or control.
This chapter presents the fundamental principles and effects of electrical
stimulation of skeletal muscle. We first introduce the principles of skeletal
muscle electrical stimulation in general and then focus on electrically in-
duced skeletal muscle responses that may occur during intentional stimula-
tion or electric shock. To achieve an understanding of the skeletal muscle
responses that may be elicited by any form of electrical stimulation, it is
necessary to consider the basic structure and function of muscle and the
nerves that innervate muscle.
8.2 Neuromuscular Structure and Function

Skeletal Muscle Innervation
The myelinated nerve fibers that innervate skeletal muscle fibers derive
from a-motor neurons (see Fig. 3.1), whose cell bodies lie within the ventral
horn of the spinal cord. These efferent axons (approximately 9 to 20.um in
diameter) leave the spinal cord through spinal-nerve ventral roots and
project out to muscles through peripheral nerve trunks. Skeletal muscles

300 8. Skeletal Muscle Response to Electrical Stimulation
may contain thousands of individual muscle fibers. However, a one-to-one

relation does not generally exist between a-motor neurons and muscle
fibers. Rather, as an a-motor neuron fiber nears and enters a skeletal
muscle, it branches into a number of collateral axons, each of which inner-
vates a single skeletal muscle fiber. An a-motor neuron and the skeletal
muscle fibers it innervates are known as a motor unit. These skeletal muscle
fibers, which upon contracting contribute to total muscle force, are known
as extrafusal muscle fibers. Intrafusal muscle fibers produce only a minute
fraction of total muscle force output. They are innervated by efferent
myelinated y-motor neuron fibers (averaging 5,um in diameter) and affer-
ent sensory myelinated Ia and II fibers (averaging about 17 and 8,um in
diameter, respectively). The sensory receptors within skeletal muscle, the
muscle spindles, incorporate a small number of intrafusal fibers. These
receptors sense muscle length and the rate of change of muscle length and
are responsible for the well-known stretch reflex of skeletal muscle-that is,
when a skeletal muscle is stretched, excitation of muscle spindles elicits
reflex activation of the muscle. The second type of skeletal muscle recep-
tors, the Golgi tendon organs, lie in series with muscle fibers as they enter
tendons. These receptors are innervated by afferent myelinated Ib nerve
fibers (averaging 16,um in diameter) and report muscle tension information
to the central nervous system.
Peripheral Nerve Trunk Structure

Peripheral nerve trunks in general carry motor, sensory, and sympathetic
nervous system information between the central nervous system and the
periphery (or vice versa). Axons are not simply "packed" into nerve trunks.
Peripheral nerve trunks usually contain several fascicles (i.e., bundles, also
called "funiculi"), each of which contains many axons as well as endoneurial
connective tissue (see Fig. 8.1). A thin perineurial connective tissue sheath
encloses each fascicle. The surface of the nerve trunk is covered by
epineurial connective tissue. The fascicular structure of a nerve trunk is not
constant throughout its course. On the contrary, regrouping and redistribu-
tion of fascicles (and the axons they contain) occur throughout the nerve
trunk. For example, in the far periphery, some fascicles tend to exit from
main nerve trunks and provide innervation to one or two skeletal muscles.
They therefore must, at this point, contain all of the appropriate muscular
efferent and afferent axons. As peripheral nerve trunks near the spinal
roots, however, axons that innervate particular muscles must regroup into a
multisegmental arrangement of fascicles. Sunderland (1978) provides an
excellent review of this and related material. An implication of this aspect
of trunk structure is that more central electrical stimulation of nerve trunks
tends to elicit more diffuse muscular responses. Electrical stimulation
of nerve trunks or branches in the far periphery can give quite specific
responses.
Neuromuscular Structure and Function 301
FIGURE 8.1. Microanatomy of a peripheral nerve trunk and its components. (a)
Fascicles surrounded by a multilaminated perineurium (p) are embedded in a loose
connective tissue, the epineurium (epi). The outer layers of the epineurium are
condensed into a sheath. (b) and (c) illustrate the appearance of unmyelinated
axons and myelinated nerve fibers, respectively. Schw, Schwann cell; my, myelin
sheath; ax, axon; nR, node of Ranvier (components not to scale). (From Lundborg,
1988, p. 33.)
Muscle Action Potentials and

Excitation-Contraction Coupling
Action potentials of a-motor neurons are generated by the cell body given
sufficient dendritic input. Each action potential then propagates along a
myelinated nerve fiber until it reaches a synapse at the neuromuscular
junction (or motor end plate) (see Fig. 3.1). At the neuromuscular junction,
neurotransmitter (acetylcholine or ACH) is released into the synaptic cleft
in response to the presynaptic action potential. The postsynaptic response
on the muscle fiber membrane (sarcolemma) is, in general, a muscle action
potential (MAP), which propagates away bidirectionally given that the
motor end plate tends to be located toward the middle of the fiber. Muscle
action potentials penetrate each fiber through the transverse-tubule
(T-tubule) system. Propagation of an MAP into the T-tubule network of a

fiber elicits release of calcium ions by the sarcoplasmic reticulum system
into the proximity of the myofilaments. These calcium ions, upon binding
with troponin C, initiate muscle fiber contraction (see below). This process,
whereby muscle action potentials lead to contraction, is known as
excitation-contraction coupling.
Skeletal Muscle Force Production

The skeletal muscle filaments (i.e., myofilaments) of Fig. 3.23 consist of
relatively thick myosin filaments and thinner actin filaments (see Fig. 5.5).
Each myosin filament contains many individual myosin molecules, each of
which possesses a "head" and a "tail." The myosin tails are wrapped to-
gether to form a myosin filament, with the myosin heads protruding out-
ward. (Each head and the "arm" of myosin attaching the head to the myosin
filament are known as a cross-bridge.) Each myosin head possesses strong
ATPase activity. That is, ATP is readily bound and cleaved (into ADP and
phosphate). An actin filament consists of a double-helix F-actin molecule as
well as tropomyosin strands, which in resting muscle act to cover "active"
sites on the actin, and troponin complexes, which are attached to the
tropomyosin molecules.
During muscle contraction, the excitation-contraction coupling process
described above results in the release of calcium ions near the myofila-
ments. Calcium binding to troponin is thought to elicit movement of tro-
pomyosin so that actin "active" sites are uncovered. Myosin heads, upon
binding ATP, produce ADP and phosphate (which remain bound to the
head) and achieve a conformational state wherein the heads protrude to-
ward surrounding actin filaments. When actin "active" sites become ex-
posed, myosin heads attach to these sites. This binding is thought to elicit a
new conformational change in the myosin cross-bridge, which "pulls" the
actin relative to the myosin-providing the force of muscle contraction
(Murphy et aI., 1996). The myosin head then releases from the actin, frees
the bound ADP and phosphate, and can bind a new molecule of ATP. At
this point, the cycle can repeat if sufficient free calcium and ATP are
present at the myofilaments. A calcium pump in the sarcoplasmic reticulum
acts continuously to decrease the calcium concentration near the myofibrils.
(Thus, in response to a single action potential, a brief surge in myofibrillar
calcium concentration, and therefore a brief contractile response, occurs.)
To prevent rapid depletion of ATP concentrations within a contracting
muscle, ADP molecules must be rephosphorylated. Some energy for this
process is immediately available from phosphocreatine. However, anaero-
bic and aerobic metabolic processes must provide the energy for contrac-
tions that last longer than a few seconds. Glycolysis of stored carbohydrates
(such as glycogen) is a relatively fast, but metabolically inefficient, anaero-
bic process by which new A TP can be formed. Oxidative ( aerobic) meta-
bolic processing of carbohydrates, fats, or proteins can provide slower,

lower, but steadier levels of energy for ATP production.
The contractile output of a single motor unit receiving a single action
potential input is a muscle twitch (actually, such a twitch represents a
summation of responses from the individual muscle fibers within the motor
unit). As we have seen in Chapter 3, two fundamental physiological means
exist by which the nervous system can modulate force production within a
skeletal muscle: temporal summation (also known as rate modulation) and
recruitment. In temporal summation (Fig. 3.24), the rate of action potential
input to the fiber is controlled so that twitch summation results in a desired
force level. Note that this summation process is in general nonlinear and
can be quite time dependent. Recruitment in this context refers to the
activation of greater or lesser numbers of motor units within the skeletal
muscle (i.e., a-motor neuron fibers).
Differentiation of Skeletal Muscle Fiber Types

The muscle fibers of motor units that innervate a given skeletal muscle can
be categorized with respect to physiological, ultrastructural, and metabolic
characteristics. Table 8.1 presents a highly simplified, but useful, differen-
tiation among fiber types based on three common classification schemes
(see Burke, 1981, for a detailed review). In general, muscle fiber and motor
unit properties lie within an overlapping continuum-and classification
schemes serve mainly as aids in relating structure-function properties of
the muscle. The classification method of Brooke and Kaiser (1970)
places muscle fibers into three categories, I, IIA, and IIB (based on his-
tochemical measurements of ATPase reactivities). Peter et al. (1972) differ-
entiated fiber types based on a combination of physiological and metabolic
properties (SO, slow oxidative; FOG, fast oxidative-glycolytic; FG, fast
glycolytic). The method of Burke et al. (1973) differentiates motor units on
the basis of their twitch speed and fatigue resistance (S, slow; FR, fast-
resistant; FF, fast-fatigable). An "intermediate" type of fast muscle fiber
TABLE 8.1. Correlation of skeletal muscle major fiber type physiological, ultra-
structural, and metabolic properties.
Fiber type lIB, FG IIAB IIA,FOG I, SO
Corresponding motor unit type FF FI FR S
Myofibrillar ATPase, pH 9.4 High High High Low
Glycogen High High High Low
Phosphorylase High High High Low
Neutral fat Low Medium High
Capillary supply Sparse Rich Rich
Aerobic metabolism capacity Low Medium Medium-High High
Anaerobic metabolism capacity High High High Low
Source: Adapted from Burke, 1981, p. 349 with permission from Oxford University Press.
also exists in some muscles (type IIAB or FI, fast intermediate resistance to
fatigue).
Slow-twitch skeletal muscle fibers are optimized for maintenance of rela-
tively low force levels, at relatively low speeds, but for prolonged periods of
time. Fast-twitch fibers are optimized for quick production of relatively
high force levels for briefer time intervals. We see from Table 8.1 that the
physiological, ultrastructural, and metabolic properties of the major skel-
etal muscle fiber types reflect a trade-off between the ability to produce
force quickly and powerfully, or slowly and steadily. Type I fibers generate
force relatively slowly (myofibrillar ATPase activity at high pH is relatively
low), yet can maintain force well through a relatively high aerobic meta-
bolic capacity (high neutral fat content, rich capillary supply, etc.). In the
other extreme case, type lIB fibers can generate force quickly (myofibrillar
ATPase activity is relatively high) via the use of anaerobic metabolic
mechanisms (high glycogen storage, high phosphorylase). However, musc1!~
force fatigue occurs relatively rapidly in these fibers. Type IIAB and IIA
fibers represent varying degrees of compromise between the design
requirements of force generation, speed/power, and fatigue resistance.
Recruitment and Firing Patterns

In normal physiological use, motor unit recruitment and firing patterns also
appear to be optimized for performing desired motor tasks (see Burke,
1981,1986). In 1965, Henneman and colleagues advanced a theory known
as the "size principle" of motor neuron recruitment (Henneman et al.,
1965a, 1965b), in which it was proposed that small a-motor neurons inner-
vating slow motor units are recruited at lower "functional thresholds" than
large a-motor neurons, which innervate fast motor units. The term "func-
tional threshold" is analogous to "excitability level" in this context and
essentially reflects the level of synaptic input necessary to excite the motor
neuron (Burke, 1981).
The logical implication of this theory is that, in "mixed" skeletal muscles
containing a distribution of muscle fiber types, low-force tasks requiring
fatigue resistance (such as force generation to maintain posture) are met by
recruiting slower motor units. Only when greater and/or faster force levels
are desired are the more readily fatigued fast motor units brought to bear.
Although this theory has, to a great extent, stood the test of time and
further experimentation (see Henneman and Mendell, 1981; Burke, 1981,
1986), it is clear that some exceptions do occur in normal physiological use
and that motor neuron size alone does not completely determine functional
excitability. For example, in rapid, powerful "ballistic" movements, rela-
tively synchronous activation of both slow and fast motor units, and even
preferential activation of fast units, can occur.
In general, a-motor neurons innervating slow-twitch motor units tend to
have average firing rates that are lower than those that innervate fast-twitch
units (see Burke, 1968, 1981). This seems sensible in that twitches generated
by slow motor units start to fuse at approximately 5 to 10Hz and reach a
tetanic state by 25 to 30Hz, while fast motor units may require 80 to 100Hz
to reach total fusion (see Burke, 1981; Wuerker et aI., 1965). In typical
physiological use, however, motor units rarely exhibit sustained firing fre-
quencies that even approach these tetanic fusion limits. For example, in the
human extensor digitorum communis muscle, Monster and Chan (1977)
observed that virtually all motor units began firing at about 8Hz regardless
of functional threshold (i.e., motor unit type) and increased their firing rates
to a maximum of 16 to 24Hz at maximal voluntary force levels.
Burke (1986) has suggested that recruitment patterns that fit the size
principle probably meet the normal functional demands placed on muscles
well while maintaining an acceptable metabolic cost. In instances where
recruitment deviates from the "orderly" pattern dictated by the size prin-
ciple, metabolic cost is sacrificed to achieve less frequently needed fast
and/or powerful motor demands.
Fatigue in Normal Physiological Use

Muscle fatigue is often defined as an inability to generate the required or
expected force (Edwards, 1981). Bigland-Ritchie et aI. (1986) have pointed
out that neuromuscular fatigue might more generally be defined as "any
reduction in the maximum force generating capacity, regardless of what
type of work is being done." It should also be emphasized that fatigue
should be regarded as a normal protective measure for survival rather than
as a deficit or abnormal phenomenon. If fatigue mechanisms did not exist,
it would be possible to exercise muscles to the point where A TP levels were
completely exhausted and muscle rigor (i.e., irreversible binding of
cross-bridges to actin) occurred.
We have seen that voluntary skeletal muscle contractions depend upon a
sequence of events involving the motor neuron (dendritic input, cell body
excitability, action potential propagation, motor end-plate presynaptic
transmission), excitation-contraction coupling (motor end-plate postsynap-
tic transmission, muscle action potential propagation along the sarcolemma
and into the T-tubules, calcium release from the sarcoplasmic reticulum),
and the actual contractile machinery (ATP dependent cross-bridge cycling,
ATP rephosphorylation through metabolic processes). Fatigue may be in-
duced due to impairment of any of these processes (Bigland-Ritchie et aI.,
1986; Fitts, 1994). It is generally accepted that both short-lasting and long-
lasting components exist in physiological fatigue. Quickly reversible effects
include sarcoplasmic electrolyte shifts and pH changes that may impair
MAP propagation through the T-tubule system, sarcoplasmic reticulum
calcium ion release, and/or force development at the myofilaments. Recent
evidence implicates a particularly important role in the fatigue process for
alterations in intracellular calcium exchange. Such changes may occur
secondary to reductions in the rates of calcium uptake and release via the
sarcoplasmic reticulum (see Williams and Klug, 1995). Force development
is obviously also dependent upon metabolic processes and products. At the
motor unit level, a clear correlation between aerobic (oxidative) metabolic
capacity and resistance to fatigue exists. As mentioned above, the fatigue-
resistant fiber types S (SO) and FR (FOG) can be associated with the high
aerobic capacity I and IIA categories; while easily fatigued type FF (FG)
units are associated with type IIB low aerobic capacity (high anaerobic
capacity). Longer-lasting fatigue may be due to depletion of metabolic
energy stores (such as glycogen) or excitation-contraction coupling impair-
ment (Le., "low frequency fatigue" (Edwards et aI., 1977».
In normal physiological use, motor unit firing patterns appear to be
adaptable to match fatigue levels. Bigland-Ritchie et ai. (1983) have
reported that, in a series of brief maximal voluntary contractions of the
human pollicis muscle, the average firing rate of more than 200 motor units
(from five subjects) dropped from 29.8 Hz (±6.4) to 18.8Hz (±4.6) between
30 and 60s after contraction onset, and to 14.3Hz (±4.4) after 60 to 90s. In
the same study, twitch contraction times before and immediately after a 60s
maximal voluntary contraction were not significantly different. However,
twitch relaxation times were prolonged by approximately 50%. This has the
effect of lowering the firing frequency needed to achieve fused tetanic
contractions. Thus, firing frequency decreases during fatigue may in part
occur because higher rates are unnecessary to achieve fused force output.
Dynamic Nature of Skeletal Muscle Fiber Types

A great deal of evidence exists that the skeletal muscle fiber "type" of an
individual motor unit is not fixed, but is instead a dynamic quantity that
depends on the neural input to and motor usage of the unit (see reviews of
Pette and Vrobova, 1985; Burke, 1981). Prolonged endurance training, or
constant electric stimulation at low frequencies (10Hz), can result in a
transformation of fiber type populations within mixed skeletal muscles.
This transformation is in the direction of faster to slower twitch and
increased aerobic metabolism capacity. Increases in fatty acid metabolic
capacity, fiber cross-sectional areas, whole-muscle fatigue resistance, and
capillary density can be elicited within weeks to months after training onset.
At moderate levels of exercise demand, oxidative capacity increases are
most dramatic as the result of apparent conversion of type FF fibers to type
FR. With constant electrical stimulation, actual conversion of type II (fast)
myosin to type I (slow) eventually can be brought about in addition to the
metabolic and ultrastructural changes already noted. Conversely, extended
disuse of skeletal muscle tends to result in fiber atrophy, decreases in tetanic
force output per fiber, and aerobic capacity loss in fatigue-resistant fibers.
The properties of muscle therefore appear to be optimized for normal
use, as determined by overall activity level, and to some extent activity and
Fundamental Principles of Skeletal Muscle Electrical Stimulation 307
force production patterns. Fast-twitch fibers, which normally see relatively

low levels of phasic activity, appear capable of adapting to new high, tonic
levels of activity. Conversely, muscle fiber inactivity appears to result in an
opposite shift to metabolic profiles more appropriate for infrequent, phasic
levels of activation.
8.3 Fundamental Principles of Skeletal Muscle

Strength-Duration (S-D) Relations for
Neuromuscular Excitation
The most basic requirement for electrical stimulation of a skeletal muscle
fiber is that an action potential(s) must be elicited on the a-motor neuron
innervating the fiber or within the membrane of the muscle fiber itself. As
has been discussed in Sect. 3.5, the electric excitability of myelinated nerve
is generally much greater than that of skeletal muscle (see Parry, 1971;
Walthard and Tchicaloff, 1971; Mortimer, 1981). Figure 8.2 depicts S-D
curves (see Sect. 4.2) for transcutaneous stimulation of human skeletal
muscles by regulated-voltage waveforms. We see that, in this example,
direct electrical stimulation of denervated muscle requires substantially
higher rheobase (i.e., the stimulus strength needed for large pulsewidths)
and chronaxy (i.e., the stimulus pulsewidth at twice rheobase) values than
for normal stimulation, where a-motor neuron axons can be activated.
Strength-duration curves for cat skeletal muscle stimulation with an
implanted intramuscular electrode (see Sect. 8.4) and regulated-current
monophasic stimuli are shown in Fig. 8.3. In these experiments on cat
skeletal muscle (Mortimer, 1981), evoked isometric muscle response was
held constant at a small fraction of total possible force. Once again, we see
that the curve for direct muscle stimulation (in this case, following admin-
istration of the motor end-plate blocking-agent curare) shifts up and to the
right in comparison to the curve for neural stimulation. Crago et al. (1974)
conducted similar experiments on intramuscular stimulation of curarized
and noncurarized cat skeletal muscle. They found that, over a range of
stimulus pulse durations and amplitudes, direct (curarized) muscle stimula-
tion accounted for only 3 to 7% of isometric muscle force evoked by neural
(uncurarized) stimulation with identical parameters.
An S-D curve time constant Te can be determined for both neural and
direct muscle stimulation by performing a mathematical fit of experimental
data to Lapicque's form of the S-D curve as expressed by Eq. (4.17). Such
data indicate Te for denervated skeletal muscle of 4.3 to 57.6ms, and for
intact muscle of 20 to 700,us (Oester and Licht, 1971). In that just-
suprathreshold intact muscle stimulation is accomplished indirectly through
excitation of a-motor neuron axons, we would expect that the S-D time
120
Normal Nerve Denervated Muscle
100 Muscle
Undergoing \ ..
~
"0
Q)
80 ·
Denervation
• ""0
..,
:E ", ..........
c.
«
"3
E
Q)
III
60
..................»~::~.......--"-..
0..
40
...........
" ........... ..
20
0
0.01 0.1 10 100
Pulse Duration (ms)
FIGURE 8.2. Strength-duration curves for transcutaneous stimulation of human

Peroneus Longus skeletal muscles by regulated voltage waveforms. The left-hand
curve is typical of normally innervated muscle. The middle curve was obtained
following a left lateral popliteal nerve lesion when the Peroneus Longus was under-
going denervation atrophy. The right-hand curve, measured six weeks after the
middle record, is typical of denervated muscle. (Adapted from Sunderland, 1978,
p. 259.) (Reprinted from S. Sunderland, Nerves and Nerve Injuries, Churchill
Livingstone, 1978, p. 259.)
constant values for intact muscle and large myelinated nerve stimulation
would be similar. Indeed, for mammalian peripheral myelinated nerve
fibers, experimentally obtained values from 29#s to over 800#s can be
determined (Ranck, 1975; Li and Bak, 1976).
We have seen in Chapter 4 that 1'e is a property not just of the excitable
membrane, but also of the method of excitation. In general, Te becomes
greater as the stimulus current crossing the cell membrane becomes more
gradually distributed along the longitudinal dimension. The absolute mini-
mum values of Te for neural and direct muscle monophasic rectangular
stimulation are the actual membrane time constants (Tm) for a-motor neu-
ron nodes of Ranvier and skeletal muscle sarcolemma.1 Based on data
obtained from rabbit peripheral myelinated axons by Chiu and colleagues
(Chiu et aI., 1979; Chiu and Ritchie, 1981), the author has estimated that
I This situation occurs, for example, when a regulated current stimulus is applied
directly across a small patch of excitable membrane. The S-D time constant 0e then
becomes equal to the membrane time constant Om [i.e., Eq. (4.4) simplifies to Eq.
(4.5)].
40
30
«
g
Q)
"0
.a
=a 20 Curarized Muscle
E
<
Q)
rJ)
"3
a.
10
o
0.01 0.1 10 100
Pulse Duration (ms)
FIGURE 8.3. Strength-duration curves for cat skeletal muscle stimulation with an
intramuscular electrode. In these experiments, evoked muscle response was held
constant at a small fraction of total possible muscle force. The left-hand curve
represents data for direct nerve stimulation. The right-hand curve represents data
for direct muscle fiber stimulation (following administration of curare). (Adapted
from Mortimer, 1981, p. 170 with permission from Oxford University Press.)
node of Ranvier Tm values would lie in the range of 10 to 20.us. Albuquerque

and Thesleff (1968) found that, for innervated rat skeletal muscles, sar-
colemmal Tm was approximately 1.7ms. While interspecies differences are to
be expected, these values indeed fall just below the lowest experimentally
obtained measurements of S-D time constants mentioned above.
Regulated-Current versus Regulated-Voltage Stimulation

Many typical stimulation waveforms are rectangular in shape. A regulated-
current stimulus (of magnitude l) with pulsewidth t will deliver a fixed total
charge Q per stimulus, such that Q = It. Regardless of electrode impedance
and potential shifts (either during a stimulus, or between stimuli), because
the current of the stimulus is regulated, a reproducible applied electric field
can be created within the tissue(s) to be stimulated (see Sect. 2.1). For this
reason, many recently designed implanted neuromuscular stimulation sys-
tems utilize regulated-current stimulus output stages. Transcutaneous neu-
romuscular stimulation systems (i.e., systems that use surface electrodes)
often produce regulated-voltage waveforms. Despite the drawback that

stimulation results may be less reproducible, in this case regulating the
voltage of a stimulus is safer-a regulated-current stimulus passed through
a dislodged or broken surface electrode could produce skin irritation or
even burns due to the resultant high current density.
Biphasic Stimulation
Balanced-charge biphasic (BCB) regulated-current stimuli are used with
many implantable electrode neuromuscular stimulation systems (see Sect.
8.4 below). With such waveforms, a cathodic rectangular pulse is followed
by an anodic phase of equal charge (either rectangular in shape or with the
form of a capacitive discharge) (see Robblee and Rose, 1990). While the
"primary" cathodic phase serves to stimulate excitable tissue(s), the "sec-
ondary" anodic phase performs charge balancing mainly for electrochemi-
cal reasons. Without some form of electrochemical reversal, deleterious
cathodic reactions such as alkaline pH swings, hydrogen gas evolution, and
oxidizing agent(s) formation can occur at the electrode-tissue interface [see
Sect. 8.4 and Robblee and Rose, (1990)]. However, although the intent of
BCB stimulation is to minimize adverse electrochemical reactions, we have
seen in Sect. 4.6 that the anodic current reversal of a biphasic stimulus can
act to abolish an action potential developing in response to the cathodic
phase. This physiological effect has been studied in vitro on isolated
Xenopus Laevis myelinated nerve fibers and in vivo using a cat nerve-
muscle preparation (van den Honert and Mortimer, 1979a). It was found
that a delay between the primary and secondary phases of approximately
loo,us effectively prevented this effect from occurring. Modeling of myeli-
nated nerve BCB electrical stimulation also predicts this result (see Fig.
4.16). Gorman and Mortimer (1983) found that the effect could be used
potentially to advantage for separating the stimulation thresholds of differ-
ing diameter a-motor neuron axons.
Fatigue and Conduction Failure in Response to

Trains of electrical pulses at rates of 10 to 20Hz and at magnitudes that
activate a-motor neuron axons can elicit a long-lasting depression in skel-
etal muscle force output in addition to metabolic deficit fatigue. This "low-
frequency fatigue" phenomenon (Edwards et ai., 1977), which also occurs
in normal physiological use (see Sect. 8.2), is thought to occur (at least in
part) because of a deficit in excitation-contraction coupling. That is, in
response to each muscle action potential, less calcium is released by the
sarcoplasmic reticulum. With increased stimulation frequencies of 20 to
100Hz, "high-frequency fatigue" may be elicited. Motor neuron propaga-
tion failure at axon branch points, neurotransmitter depletion at the motor
endplate, and MAP propagation failure all may contribute to this effect.
Bigland-Ritchie et al. (1986) feel that high-frequency fatigue is avoided in
normal physiological use because motor neuron firing rates remain below
the frequencies needed for propagation failure.
Bowman and McNeal (1986) studied in vivo the response characteristics
of single a-motor neuron axons of cat sciatic nerves when very high-
frequency trains (lOO-lO,OooHz) of balanced-charge biphasic stimuli were
applied. At frequencies of 100 to 1,000 Hz, axonal firing rates were generally
equal to or were subharmonics of the stimulation frequency. Over a 3-min
stimulation period, an axon often typically initially fired at the same rate as
the stimulus train for a very brief time, shifted to firing with every other
stimulus, gradually shifted to firing with every third stimulus, and some-
times then shifted to every fourth. Increased stimulus amplitudes tended to
delay the downward shift in firing frequency with time. Axonal firing
frequencies in this range tend to rapidly deplete supplies of ACH at the
neuromuscular junction. This type of "junction fatigue," first studied by
Wedensky (1884) and Waller in 1885 (see Bowman and McNeal, 1986),
rapidly results in contractile force loss (see McNeal et aI., 1973; Solomonow
et aI., 1983).
At stimulation frequencies of 2,000 to 1O,000Hz, Bowman and MeNeal
found that axons would briefly respond with firing rates of several hundred
Hertz, but stopped firing completely within seconds after stimulation initia-
tion. This type of "electrical conduction block" could be maintained for
periods up to 20 min. with recovery following stimulation termination
occurring within 1 s. The mechanism involved in this type of very high-
frequency conduction block, which has also been studied by a number of
other investigators using rectangular stimuli or AC sinusoidal trains (e.g.,
Cattell and Gerard, 1935; Katz, 1939; Tasaki and Sato, 1951; Tanner, 1962),
is not completely known. Polarization buildup at nodes of Ranvier due to
differing conductivities for inward and outward currents (see Sect. 4.6) may
act to depolarize excitable membranes to the point where strong sodium
channel inactivation results.
Sinusoidal Stimulation Strength-Frequency (S-F) Effects

Nonlinear membrane polarization buildup is also thought to account for the
upward deflection in AC sinusoidal stimulation S-F curves above a mini-
mum level that typically occurs at about 150Hz for myelinated axons (see
Fig. 4.20, right). At stimulation frequencies below about 100 to 150Hz,
rectangular stimulation thresholds for a-motor neurons are independent of
stimulation frequency, and each stimulus tends to produce one action po-
tential. AC sinusoidal trains, however, exhibit an upturn in their strength-
frequency behavior (see Fig. 4.20, left). That is, higher-amplitude sinusoidal
trains are needed to excite the nerve. The well-known "accommodation"
of excitable membrane thresholds (see Sect. 4.6) in response to slowly
increasing stimuli is thought to be responsible for this effect. For example,

Sato and Ushiyama (1950) demonstrated in an in-vitro toad motor nerve
preparation that accurate mathematical prediction of the form of S-F
curves below 40Hz could be performed through consideration of mem-
brane accommodation properties [see Tasaki (1982) for an excellent review
of this and related topics].
As sinusoidal frequencies approach DC, it becomes quite difficult to
stimulate peripheral nerves. Stimulation may be elicited at the onset of
direct current as with any rectangular pulse. However, following the onset
of direct current, additional stimulation of a-motor neurons or skeletal
muscle is not easily achieved except at very high stimulus amplitudes.
Extremely high-amplitude stimuli, at very high frequencies (2,000Hz) or at
DC, have been found to produce "tetanoid" responses directly in isolated
(curarized) frog skeletal muscles (Koniarek, 1989). After initiation of this
type of response, a portion of each muscle fiber appears to reach a constant
depolarization level.
8.4 Functional Neuromuscular Stimulation Systems

Electrically induced stimulation of skeletal muscle for artificially controlled
exercise and restoration of motor function (lost because of injury or dis-
ease) is generally the objective of functional neuromuscular stimulation
(FNS) systems. In addition, FNS electrode and stimulation technologies
recently have also been adapted for use in other applications, such as
dynamic myoplasty. A number of excellent reviews of and texts on FNS,
and the general area of functional electrical stimulation (FES), already
exist (e.g., Vodovnik, 1981; Mortimer, 1981; Peckham, 1987; Agnew and
McCreery, 1990; Rattay, 1990; Peckham and Gray, 1996). In this section, we
will briefly present the current state of FNS and related applications, meth-
ods, models, and electropathology and will introduce basic principles that
can provide greater understanding of accidental stimulation in electrical
accidents.
Objectives and Applications of FNS

Systems and Technologies
It has been estimated that there are over 200,000 spinal cord injured people
in the United States. Spinal cord injuries occur mostly in younger people,
with the most common age at injury being 19 years and the mean age at
injury being 29 years (Young et ai., 1982). A conservative estimate of the
annual cost of care for spinal cord injured persons is one and a half billion
dollars. In the paraplegic person (spinal cord damage in the thoracic or
lumbar region), the freedom to walk, run, or simply stand from a chair to
Functional Neuromuscular Stimulation Systems 313
reach a shelf is restricted or lost. Quadriplegic individuals (spinal cord

damage in the cervical region) will have lost, in addition, some or all of the
ability to manipulate and control their environment. Functional neuromus-
cular stimulation of a-motor neurons that remain intact below the level of
spinal cord damage provides a means by which some control of lost motor
function can be reasserted. At this time, a number of research centers
around the world are developing systems designed to provide upper ex-
tremity control (e.g., Peckham, 1983; Hoshimiya et al., 1989; Smith et al.,
1996; Keith et al., 1996; Triolo et al., 1996) and lower extremity control (e.g.,
Waters et al., 1985; Holle et al., 1984; Kralj et al., 1986; Marsolais and
Kobetic, 1987; Graupe, 1989; Jaeger et al., 1989; Shimada et al., 1996). FNS
can also be used for motor restoration or assistance in other disabled
populations where a-motor neurons remain intact, but central control has
been lost or impaired (e.g., stroke, cerebral palsy).
The term "myoplasty" refers to any clinical procedure wherein skeletal
muscle is transferred within the body (e.g., in plastic surgery). In "dynamic
myoplasty," electrical stimulation is used to activate muscle that has been
surgically relocated (for a recent review, see Grandjean et al., 1996). In
"dynamic cardiomyoplasty," a skeletal muscle (usually the left latissimus
dorsi) is mobilized except for its neurovascular pedicle, internalized
through a window created by partial resection of the second or third rib,
and secured to the heart. Leriche and Fontaine proposed as early as 1933
that skeletal muscle might be used to replace damaged myocardial tissue.
Early efforts in this area were frustrated primarily by the skeletal muscle
fatigue that resulted when the muscle was subjected to cardiovascular work
loads (see Kantrowitz, 1990). With the discovery (see Sect. 8.2) that chronic
electrical stimulation can be used to improved skeletal muscle fatigue
resistance through conversion of type II (fast) fibers to type I (slow),
this application of FNS technologies has seen renewed investigation
(Carpentier and Chachques, 1985). In the dynamic cardiomyoplasty surgi-
cal procedure, skeletal muscle may be wrapped around the heart, used
to replace resected myocardium, or both (see Magovern et al., 1990;
Carpentier et al., 1991). Electrical stimulation ofthe latissimus dorsi muscle
in synchrony with the heart is accomplished using an implanted neuromus-
cular stimulator and electrodes (Grandjean et al., 1991). Worldwide clinical
evaluation of the dynamic cardiomyoplasty approach to cardiac assistance
is now underway (e.g., Moreira et al., 1996; Furnary et al., 1996; Oh et al.,
1996). Alternative approaches to the use of electrically stimulated skeletal
muscle for cardiac assistance, which are also under experimental study,
include creation of "skeletal muscle ventricles" (by wrapping muscle into or
around a pumping chamber much like a ventricle of the heart) and
"aortomyplasty" (where muscle is wrapped around the aorta) (see e.g.,
Chachques et al., 1996; Niinami et al., 1996). Dynamic myoplasty for treat-
ment of fecal incontinence is also under investigation (e.g., Williams et al.,
1991; Baeten et al., 1995).
Electrical Excitation of Skeletal Muscle by

Functional Neuromuscular Stimulation
Excitation of a given motor nerve fiber can be achieved, in general, if an
electric field is introduced (by electrodes) that involves a focalized change
in the extracellular voltage gradient. Equations such as (4.31) imply that a
second-difference function involving extracellular potentials at nodes of
Ranvier [see Sect. 4.4 and Rattay (1986)] characterizes the "driving force"
of electric field distributions. In Eq. (4.31), the term (Ve,n+l - 2V.,n + V.,n+l)
represents a second-difference sampling of the longitudinal extracellular
potential distribution along the fiber. Focal changes in the extracellular
electric field can therefore cause depolarization of the axonal membrane at
nodes of Ranvier and generation of action potentials that will propagate
toward a muscle.
In spinal cord injured people, most motor neurons in the spinal cord will
remain intact if they are below the level of the injury. Electrical stimulation
of their axons can therefore be used to cause muscle contraction. The
restoration of functional movement in general requires the fine gradation of
force so that adequate control can be achieved and damage to the body or
to objects in the environment can be avoided. Just as skeletal-muscle force
output is controlled in normal physiological use by recruitment of a-motor
neurons and by rate modulation of each motor neuron's firing, electrical
stimulation can be used to control the same variables artificially. FNS
system designers often intuitively strive to replace the body's normal con-
trol strategies; however, major limitations still exist in the present ability to
achieve such control.
Electrodes for Functional Neuromuscular Stimulation

An excitatory electric field can be generated by an FNS system using a
variety of electrode types (for reviews, see Mortimer et aI., 1995; Sweeney
et aI., 1996). Stimulation using electrodes located on the skin surface is the
most straightforward possible approach. It was determined as early as 1833
by Duchenne de Boulogne that skeletal muscles could be stimulated
by passing current percutaneously via cloth-covered electrodes (see Licht,
1971; Walthard and Tchicaloff, 1971). Duchenne found that contractions
were most easily obtained by stimulating at a discrete location on the skin
overlying each muscle. He called such areas "points of election" and
theorized that the muscle beneath each point was hyperexcitable. Remak
and von Ziemssen subsequently showed that such locations, now known
as "motor points," actually overlie sites where nerve trunks penetrate
and branch out into the deep face of muscles. Coers (1955) refined this
essentially correct viewpoint. He found that the greatest concentration
of motor end-plates in a superficial muscle is located in the superficial
part of the muscle beneath the skin surface motor point. "Motor lines"
reflect linear projections onto the skin surface of underlying superficially

located motor nerve trunks. Walthard and Tchicaloff (1971) provide exten-
sive maps and tabulations of approximate motor point and motor line
locations.
Stimulus parameters for surface-electrode motor point stimulation vary
from one muscle to another and depend on muscle/nerve anatomy; whether
voltage or current is regulated; the waveform; and the electrode(s) type,
size, placement, and configuration. Vodovnik et aI. (1967) found that for
regulated-voltage rectangular stimuli, while threshold parameters for
stimulation of upper extremity muscles were quite variable, typical thresh-
old values (for 200-,us pulses at 50Hz) were on the order of 20V. Voltages
in excess of 100V may sometimes be necessary to achieve high activation
levels (Vodovnik et aI., 1981). Figure 8.4 presents typical S-D curves for
surface-electrode regulated-current stimulation of four upper-extremity
muscles. These curves were obtained by placing an electrode over the
motor point of each muscle and measuring, for each pulse duration, the just-
threshold pulse amplitude. Crago et aI. (1974) have reported similar results
(e.g., for stimulation of the flexor digitorum sublimis muscle in a typical
subject: 9.9mA threshold, for O.5-ms-duration stimuli at 50Hz). Bajzek and
Jaeger (1987) found that, for triceps surae stimulation, similar or somewhat
20
15 / " Biceps (Right Arm)

<
.s
CD
Triceps Long Head
(Left Arm)
"0
~ (Right Arm)
C. 10
E
~
CD
III
"5
a..
5
/
Upper Trapezius (Right Arm)
o
0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9
Pulse Duration (ms)
FIGURE 8.4. Strength-duration curves for surface-electrode threshold stimulation of

four upper-extremity muscles. (Frequency = 50 Hz; C-3 level spinal cord injured
patient "H.K.") (Adapted from Vodovnik et aI., 1967.)
higher charge parameters were necessary (e.g., on the order of 50 to 100,us

pulse duration thresholds for 60 to 80mA stimuli at 25Hz).
An increase in surface-electrode stimulus intensity above the threshold
level tends to activate a larger volume of the underlying muscle and there-
fore increase force output until maximal muscle activation is achieved.
Equation (4.31) implies that, for a given electric field distribution beneath a
surface electrode, a "shell" will exist under the electrode that separates (for
each size nerve fiber) regions where excitation will occur. Figure 8.5 pre-
sents a family of S-D curves for varying levels of torque about the elbow
given surface-electrode stimulation of the biceps muscle. Each curve essen-
tially represents a constant degree of biceps muscle activation (i.e., a
constant level of neural excitation penetration into the muscle body).
Prediction of the relationship between surface-electrode stimulation pa-
rameters and underlying muscle activation level is nontrivial. Rattay (1988)
has modeled the excitation of nerve fibers under disk-shaped surface elec-
trodes using the "activating function" approach. He found, as have others,
that the tendency for high current densities to arise at the edge of a surface
electrode will ultimately limit the depth to which excitation can be elicited
(i.e., at sufficiently high edge current densities, burns can occur).
30
25
<
g 20
Q)
"0
.a
aE 15
c(
Q)
U)
:;
a.. 10
5
zero ft-Ibs
o
o 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8
Pulse Duration (ms)
FIGURE 8.5. Strength-duration curves for surface-electrode tetanic stimulation of

the biceps muscle at five levels of isometric torque about the elbow. Curve for zero
ft-Ibs reflects the just-threshold response. (Frequency = 50Hz; elbow angle = 75°;
subject "P.c.") (Adapted from Vodovnik et aI., 1967.)
c:
0
c::: 8.5
l!
(,)
Q)
15
E 8
.g
E
~
'E
·s 7.5
a.
...
~
~ 7
0
c::: •
N
.3
6.5
6
15 30 45 60 75 90
Elbow Angle (degrees)
FIGURE 8.6. Location of the motor point of the medial head of the triceps muscle
versus elbow angle for surface stimulation using a 1-in.-diameter electrode (pulse
amplitude = 15mA; pulse duration = O.2ms; frequency = 50Hz; subject "W.C.")
(Adapted from Crochetiere et aI., 1967.)
A fundamental difficulty with surface motor point stimulation is that,

while electrodes are applied to the skin, the motor point is inherently
dependent on anatomical properties of the underlying muscle. Thus, as
shortening or lengthening of a muscle occurs during a motor task, and as
limb positions or joint angles change, the location of a motor point will shift
with respect to fixed locations on the skin surface. Figure 8.6 illustrates this
effect for surface electrode stimulation of the medial head of the triceps. In
these experiments, Crochetiere et al. (1967) estimated that the 2.5 cm shift
in the motor point skin surface location (over an elbow angle range of 15-
90°) corresponded well with a theoretical expected excursion of 2.5 cm in
the underlying muscle.
In a recent study of joint angle and electrode position effects on surface
stimulation characteristics, McNeal and Baker (1988) found that, in activa-
tion of the hamstring muscles, a large intersubject variability existed in
predicting surface regions of greatest excitability. Changes in regions of
excitability also generally occurred as the knee was flexed. Excitability
regions for surface stimulation of the quadriceps muscles showed less
dependence on joint angle. Surface electrode stimulation also tends to be
relatively non-selective because the electric field created is necessarily
widespread. For example, stimulating the innervation of deep muscles with-

out also stimulating overlying superficial muscles is impossible. Despite
these drawbacks, however, the non-invasiveness of surface stimulation has
great appeal; and a number of surface electrode based FNS systems have
been reported. Patients using surface electrode systems for lower-extremity
FNS have been able to stand, to walk at speeds approximately one-quarter
of normal and to climb stairs (Kralj et aI., 1983; Petrofsky and Phillips,
1983). In stimulation systems intended purely for artificially generated
muscle exercise (as opposed to function restoration), surface stimulation is
clearly preferred.
Intramuscular electrodes (Caldwell and Reswick, 1975; Crago et aI., 1980;
Handa et aI., 1989; Memberg et aI., 1994), which can be percutaneously
placed or fully implanted into skeletal muscles, and implanted epimysial
electrodes that are placed on skeletal muscle surfaces (Grandjean and
Mortimer, 1986) enable selective activation of individual muscles (or even
muscle regions). FNS research systems that use intramuscular or epimysial
electrodes are presently limited by the need to implant, maintain, and use a
large number of electrodes if many muscles are to be controlled (e.g., see
Marsolais and Kobetic, 1987). An exciting recent development in intramus-
cular electrode design is the concept of "injectable" microstimulators that
receive power and command signals by inductive coupling (Loeb et aI.,
1991).
So-called "neural" electrode systems place electrodes directly into or
adjacent to peripheral nerve trunks (Naples et aI., 1990) and offer the
possibility of controlling multiple muscles with a single implant. As pre-
sented in Sect. 8.2, in the distal segments of peripheral nerve trunks, indi-
vidual fascicles (bundles) tend to contain the innervation for one or two
skeletal muscles (Sunderland, 1978). Horch and colleagues have developed
and demonstrated the feasibility of an implantable intrafascicular electrode
stimulation system that uses Teflon-insulated 25-,um diameter Pt-Ir wires.
Single pulse charge thresholds for muscle activation using such in-
trafascicular electrodes can be remarkably low-typically about 1 nC
(Nannini and Horch, 1991; Yoshida and Horch, 1993). Tyler and Durand
(1993, 1994) have studied an interfascicular stimulation electrode design
(the "slowly penetrating interfascicular nerve electrode", or "SPINE") that
allows placement of electrodes between but not within fascicles. Epineural
stimulation (i.e., surgical implantation of electrodes onto the epineurium of
a nerve trunk) has been used by one research group (Holle, 1984). All of
these intraneural and epineural electrode systems are potentially suitable
for selective activation of individual nerve trunk fascicles and, because they
are not introduced into or onto muscles, may be more reliable and in some
instances easier to implant than intramuscular or epimysial electrodes.
Nerve cuff electrodes, which involve placement of electrodes (within insu-
lating tubes or sheaths) near but not within the peripheral nerve may offer
similar advantages as intra- or epineural electrodes but without invasive
penetration of the nerve trunk. A peroneal nerve cuff electrode, developed

by Medtronic, Inc., and McNeal and colleagues (Waters et ai., 1975, 1985;
McNeal et ai., 1977), has been used for correction of footdrop in a large
patient population; with some implants functioning well after many years.
McNeal and Bowman (1985) tested in animals multiple-electrode cuffs
intended to provide greater selectivity of muscle activation. Electrode posi-
tion and good contact (i.e., close apposition of the electrodes to the nerve
trunk) were found to be important factors in successfully achieving selectiv-
ity of stimulation with this design. Agnew and coworkers (e.g., Agnew et ai.,
1989) have developed a helical cuff electrode intended to reduce possible
mechanical damage to nerve trunks. Spiral nerve cuffs (Naples et ai., 1988)
suitable for fascicle selective stimulation have been developed by several
groups. These cuff designs enable direct apposition of electrodes to a nerve
trunk surface without eliciting the mechanical damage normally caused by
tightly fitting cuff systems (Naples et ai., 1990; Sweeney et ai., 1990; Veraart
et ai., 1993; Rozman et ai., 1993; Sweeney et ai., 1995). Most recently,
Walter et ai. (1997) have demonstrated the feasibility of producing selective
hand and wrist movements in a raccoon animal model using multielectrode
nerve cuff stimulation.
Recruitment and Rate Modulation in Functional

Neuromuscular Stimulation
With regulated-current BCB rectangular stimuli, recruitment levels are in
general controlled by altering the charge within the primary stimulus (i.e.,
by changing the amplitude or the pulse width). Figure 8.7 illustrates an
example of isometric force recruitment using intramuscular electrodes im-
planted in cat soleus muscle. Recruitment by pulse width and pulse ampli-
tude modulation is compared. Recall from Sect. 4.5 that a minimum charge
threshold is approached for very small values of pulse width. This property
explains the slight shift toward lower charge injection values across most of
the range of recruitment for pulse width modulation in comparison to
amplitude modulation for this example.
Recruitment selectivity in FNS involves two separate aspects-spatial
selectivity and nerve fiber diameter selectivity. In the above discussion, we
introduced the concept that implanted electrodes offer greater selectivity in
activating different muscles than does surface stimulation. This enhanced
"spatial selectivity" occurs because the electric field introduced can be
focused closer to the a motor neuron fibers of interest. An aspect of FNS
not yet discussed, fiber diameter selectivity, relates to the tendency to
stimulate subpopulations of nerve fibers (at the same spatial location) based
on their size. Consider again the form of the spatially extended nonlinear
nodal (SENN) myelinated nerve model equation (4.31). In Eqs. (4.27) to
(4.29) and (4.32a), the terms Cm, G a , and Ii,n are all proportional to fiber
Charge (flC)
o 0.1 0.2 0.3
1.0
0.8
Q)
~
0 0.6
LL
"0
Q) Modulation
.~
til
E 0.4 o Amplitude
0 o
z Pulse Width
0.2
o 2 3
amplitude (rnA)
o 20 40 60 80 100
pulse width (flS)
FIGURE 8.7. A comparison of isometric force recruitment by pulse width and ampli-
tude modulation. Circles: normalized force as a function of pulsewidth (zero to
lOOfts) at a fixed stimulus amplitude (2.7mA); squares: normalized force as a
function of stimulus amplitude (zero to 2.7mA) at a fixed pulse width (lOOfts).
Abscissa also gives the total charge per stimulus (top scale) for either modulation
scheme. Data are for 10-Hz intramuscular stimulation of a cat soleus muscle.
(Adapted from Crago et aI., © 1980 IEEE.)
diameter. These terms occur in Eq. (4.31) only as ratios. As pointed out by
McNeal (1976), the only effect of changing fiber diameter is therefore in the
calculation of the extracellular potential terms. Because internodal length is
proportional to fiber diameter, the extracellular potentials at the nodes will
change with fiber diameter. In one idealized situation where an extracellu-
lar potential distribution rises linearly along a fiber, peaks over a node of
Ranvier, and then declines linearly, Eq. (4.31) predicts that fiber excitation
threshold will be exactly inversely proportional to fiber diameter. This
theoretical dependence of excitability given a linear field along a nerve
matches the experimental results of Tasaki (1953) for studies on single
myelinated fibers from frog sciatic nerve. In that the focalized electric fields
generated by actual stimulation systems rarely approximate such an ideal

distribution, more complex dependencies of stimulus threshold upon fiber
diameter are generally found. McNeal's original model predicted that, for
regulated-current stimuli of 100-,us duration from a monopolar point cath-
ode located 1 mm above a node of Ranvier, large fiber diameter thresholds
were inversely proportional to the square root of fiber diameter. In the
small diameter range, threshold approached an inverse square relationship
with fiber diameter.
Recall from Sect. 8.2 that, in most normal physiological use, the "size
principle" predicts an orderly recruitment of motor units from slow (small
diameter axons from small motor neurons) to fast (large diameter axons
from large motor neurons). In most electrical stimulation of myelinated
fibers, however, there will be a tendency to recruit large axons at small
stimulus magnitudes and then smaller axons with increased stimulus levels.
Such "reversed recruitment" of myelinated nerve fibers with extracellular
electrical stimulation is a well-known phenomenon (e.g., Blair and
Erlanger, 1933; Petrofsky, 1978; Fang and Mortimer, 1987). A major prob-
lem in FNS for motor control restoration is that, even if sufficient spatial
selectivity is achieved (primarily through electrode selection and place-
ment), reversed recruitment of motor units will inappropriately utilize fast,
more readily fatigued muscle fibers for low-force tasks. Slower, fatigue-
resistant muscle fibers will only be recruited at higher stimulus levels. This
also results in an undesirable, steep relation between force output and
stimulus magnitude. Exacerbating this problem is the fact that, with the
muscle disuse that tends to follow spinal cord injuries, many fatigue resis-
tant muscle fibers tend to shift their metabolism towards less oxidative, and
more anaerobic, more readily fatigued mechanisms. To circumvent these
problems in clinical use of FNS systems, electrical stimulation "exercise"
procedures are sometimes employed to alter the metabolic and physiologic
properties of type IIB (FF) fibers and type IIAB (FI) towards the less
readily fatigued type IIA (FR) and type I (S) profiles.
Exceptions to the reversed recruitment principle of electrical stimulation
probably occur at very small electrode-nerve separation distances. Con-
sider the idealized situation where an electric field generated by a point
source very close to a node of Ranvier is so highly focused that, at adjacent
nodes of Ranvier regardless of fiber diameter (and therefore internodal
spacing), no appreciable extracellular potentials occur. In this case, the
extracellular potential second-difference function of Eq. 4.31 simply re-
duces to the term Ve,n for all size fibers. We would therefore expect that all
size fibers would be recruited at about the same current threshold. Model-
ing studies of Veltink et al. (1988, 1989a), Altman and Plonsey (1989), and
Sweeney et al. (1989) predict that as electrode-nerve separations decrease
below the order of internodal lengths, strict reversed recruitment order
disappears. This effect also has been studied experimentally by Veltink
et al. (1989b) using intrafascicular electrodes.
Rate modulation in FNS differs from that in normal physiological use in

one important respect. FNS systems will, by their nature, activate all motor
units synchronously in phase with each stimulus at a predetermined fre-
quency. In normal physiological use, the nervous system asynchronously
activates different motor units. The result is that, at moderate average firing
rates, smoothly controlled force production is possible. In synchronized
FNS activation of motor units, however, smoothly fused contractions are
possible only at higher average firing rates. As we have seen in Sect. 8.2,
however, higher firing rates more rapidly fatigue the muscle fibers.
A variety of control strategies (including feedforward, feedback, and
adaptive) which incorporate different types of machine intelligence (e.g.,
optimization, neural networks, fuzzy logic) have been implemented, tested,
and/or proposed for FNS systems (see Chizeck et aI., 1988; Crago et aI.,
1996). Recruitment modulation (varying pulse width or amplitude at a fixed
stimulation frequency) is generally used to achieve low to moderate force
output control. Rate modulation can then be used to produce high force
levels after all motor units have been recruited. Simultaneous modulation
of both recruitment levels and rate modulation is much more complex but
offers the possibility of finer, more rapid control of force and of changes in
force. In most current FNS systems, control and feedback signals are gener-
ated by externalized artificial sensors which can be difficult to build and
maintain. However, it may eventually be possible to record "natural" sen-
sor signals from implanted peripheral nerve or CNS electrodes for use in
FNS feedback and control (for a review, see Hoffer et aI., 1996).
Electropathology of Functional Neuromuscular

Stimulation Systems
The mechanisms by which electrical stimulation might damage tissues fall
into two general categories: those deriving from electrochemical processes
that enable charge transfer at electrode-tissue interfaces and those that
occur directly or indirectly as a result of current flow within the tissues [for
a review, see McCreery and Agnew (1990)]. An implanted neuromuscular
stimulation system must meet general biocompatability requirements (be-
yond those related to electrical stimulation) as well (Yuen et aI., 1990).
Current flows in metals by electron movement, while current in the body
consists of the flow of charged ions. This flow of charge can be induced by
capacitive or conductive electrode mechanisms [see Robblee and Rose
(1990) for a review of stimulating electrode electrochemistry]. Capacitive
mechanisms (which involve charging and discharging of the electrode
"double layer") are generally not sufficient to deliver the current densities
needed for nerve or muscle simulation. Conductive mechanisms (which are
also known as Faradaic and involve electron transfer across the electrode-
tissue interface via electrochemical redox reactions) are therefore almost
always necessary. "Reversible" conductive reactions involve chemical spe-
Skeletal Muscle Stimulation in Electrical Accidents 323
cies that, ideally, are bound at the electrode surface and can be quantita-
tively reversed by alternately passing equal and opposite amounts of
charge. Examples of such reactions are oxide layer formation/depletion and
hydrogen atom plating. Balanced-charge biphasic (BCB) regulated-current
stimuli (see Sect. 8.3) can therefore use reversible conductive reactions for
electrical stimulation without introducing new electrochemical species into
the tissue. "Irreversible" conductive reactions necessitate formation of new
species and in general may elicit variable amounts of tissue damage depend-
ing on the body's ability to buffer or suppress deleterious effects. Examples
of such processes are electrode corrosion, oxidation of chloride ions, and
the electrolysis of water. Such reactions can bring about the formation of
biologically toxic products and may generate large swings of pH near the
electrode-tissue interface.
Although electrical stimulation ofaxons within their normal physiologi-
cal range does not in general elicit any damage, Agnew et al. (1989) have
reported that prolonged BCB electrical stimulation (in which only revers-
ible conductive reactions were presumably utilized) of cat peripheral nerve
could elicit damage in a fraction of large myelinated nerve fibers. This result
depended on the stimulus parameters (current, pulsewidth, and frequency)
and the total duration of stimulation. For the peroneal nerve trunk,
supramaximal stimulation of Aa nerves for 8 h or more at 50 Hz consistently
produced this early axonal degeneration (EAD) in some fibers. Depending
on the stimulation parameters, damage could sometimes be found with
frequencies as low as 20Hz. In a recent series of experiments, Agnew and
colleagues (see McCreery and Agnew, 1990) found that local anesthetic
applied to peroneal nerves near the stimulation site prevented EAD. This
complete abolition of damage by local anesthesia (which will prevent elec-
trical excitation of the nerve fibers) strongly suggests that hyperactivity of
axons may be at least partially responsible for EAD generation.
8.5 Skeletal Muscle Stimulation in

Electrical Accidents2
The neuromuscular effects of accidental electric shocks have been a subject
of study since the first development of commercial electric current sources.
In particular, two possible neuromuscular responses to accidental electrical
stimulation have received a great deal of attention because they can poten-
tially contribute to injurious or fatal outcomes: the generation of involun-
tary movements by skeletal muscle activation and arrest of respiration. In
this section, we will review existing experimental data on skeletal muscle
2Much of the material in this section is extended from an excellent review of 60-Hz
contact current safety levels by Banks and Vinh (1984).
stimulation in electrical accidents. Section 8.6 analyzes contractile move-

ments of the hand contacting a relatively low-voltage electrical source and
the tetanizing condition of the so-called "let-go" threshold. In Sect. 8.7, we
will review the possible effects of electric stimulation on respiration.
History and Overview

Virtually since the introduction of commercial electricity, it has been recog-
nized that alternating current could produce tingling and pain (see Chapter
7), sensations of tissue heating, stimulation of muscle contractions, respira-
tory difficulty or arrest, and even cardiac fibrillation (see Chapter 6). At
current levels where voluntary skeletal muscle control is lost and involun-
tary movements occur, "freezing" to a live conductor held in the hand is
possible. As c.P. Dalziel, who performed many landmark investigations in
this area, noted in 1956, electric stimulation of the forearm and hand flexors
can generate very powerful grasping contractions that are not easily over-
come. Dalziel reported that "in 1942, a shipyard workman in the San
Francisco Bay area attempted to obtain a better view of an in-plant Christ-
mas Eve party by climbing around an overhead crane. He slipped and, to
keep from falling, grasped the 440-volt crane trolly wires. He immediately
froze to the trolley wires, and his body swung back and forth until someone
opened the switch. He then fell to the ground, fractured his skull, and died.
A similar accident occurred in 1947 at a brass works in the Los Angeles
area. A laborer lost his footing when climbing down from the roof of the
factory. As he fell, he grasped the 220-volt trolley wires of an overhead
crane and froze to the trolley wires. His body dangled from the wires until
one employee opened the switch while another employee caught him as he
fell. He received severe burns but recovered" (Dalziel and Massoglia,
1956).
Early investigators of this effect defined the current level where an indi-
vidual could, through remaining voluntary control, just barely free himself
from a hand-held current source as the let-go value. Experimental study of
this value and its dependence on a number of variables [subject sex and age,
contact(s) size and location, skin conductivity conditions, stimulus wave-
form and parameters, etc.] has dominated investigations on accidental neu-
romuscular stimulation effects.
Let-Go Thresholds for Hand Contact with

Electrical Stimuli
In the 1920s and the 1930s, a number of studies on let-go thresholds (or
similar tests) for 50- or 60-Hz stimulation were carried out (Grayson, 1931;
Thompson, 1933; Dalziel, 1938; Gilbert, 1939; Whitaker, 1939). A summary
of the results of these studies is presented in Table 8.2 (after Banks and
Vinh, 1984). Despite the widely varying methods and end points used, the
TABLE 8.2. Summary of data from early let-go current investigations.
Let-go current (mArms) Frequency
Report Subject Current path Active electrode Endpoint Minimum Average Maximum (Hz)
Gilbert 25 Hand-hand 20-mm brass Release grip NR 15 NR 50 en
~
(1939) Both sexes rod (1l
Whitaker 13 Hand-hand Pliers Release grip 6.0 7.8 10 NR [

(1939) Eo
Thompson 42 men Hand-hand Rotatable metal Rotate handle NR 8.35 20.0 60 ~
cen
(1933) 27 women handle NR 5.15 8.8 60 Co
(1l
Grayson 42 men NR NR Not "serious NR 8 NR NR
en
rl
(1931) discomfort"
Dalziel 56 men Hand-hand Flat metal disk Tolerate NR 13.9 22 60
S·
c
(1938)' sensation ;-
rl
18 men Hand-foot Flat metal disk Tolerate NR 13.5 NR 60 o·
::l
sensation
S·
Dalziel 42 men Hand-hand No.lOAWG Release grip NR 12.6 18 60
tTl
(1938)b copper wire ;-
(")
rl
NR = not reported.
...o·
"First series. Eo
bSecond series. (")
>
(")
Source: Adapted from Banks and Vinh, © 1984, IEEE. 0:
(1l
::l
rl
en
\.;.)
N
Ul
average values for humans fell within a fairly reasonable range of about 8 to
15mArms.
From 1941 to 1956, Dalziel and his colleagues reported in a series of
papers on human let-go currents and voltages. Despite some experimental
and statistical methods that would not be acceptable by today's standards
(Banks and Vinh, 1984), these reports still form the foundation of our
knowledge about the let-go phenomenon and have had great impact in
setting safety standards for line-frequency contact currents (see Chapter
11). The adult experimental subjects in these studies were predominantly
healthy, young, male volunteers aged 21 to 25 years (with an upper age
tested of 46 years). A smaller number of female subjects in their "late teens
to the early 20's" were also tested (Dalziel and Massoglia, 1956). Dalziel
described his female subjects as probably representing a fairly sedentary
population. A small number of children were also tested. In all experiments,
"an individual's let-go current [was] defined as the maximum current he
[could] tolerate when holding a copper conductor by using muscles directly
affected by that current" (Dalziel et aI., 1943). A No.6 or No.7 A WG
polished copper wire was grasped as the active electrode in most studies.
Dalziel and colleagues felt that conductor size and shape did not have a
large effect on threshold measurements. The hands of the subjects were
kept moist with a "weak salt" solution, and any of three indifferent
electrode sites/types were used (usually a brass plate under the opposite
hand, but sometimes a plate under one or both feet, or a conducting
band around the upper arm), apparently without strongly affecting the
results. After allowing each subject to become accustomed to the sensations
that would be encountered, the current was raised slowly to a certain
value, at which point, the subject was commanded to drop the electrode. If
he or she succeeded, the current values were increased by steps until
the subject was no longer able to let go of the wire; if the subject failed, the
test was repeated at a lower value. It is important to note that, at the let-
go threshold current value, "failure to interrupt promptly the current is
accompanied by a rapid decrease in muscular strength caused by pain
and the fatigue associated with the severer involuntary muscular contrac-
tions" (Dalziel and Massoglia, 1956). The end point of experiments
was checked by several trials thought to be insufficient in number to fatigue
the individual but enough to determine the let-go value accurately (Dalziel
et aI., 1941). This process was undoubtedly quite unpleasant (see Fig. 8.8).
Dalziel and colleagues recognized that psychological state and motiva-
tion levels certainly affected results. "Psychological factors, especially
fear and competitive spirit, were the most important causes for the varia-
tions" (Dalziel et aI., 1943). Table 8.3 and Figure 8.9 summarize results
obtained for the 60-Hz let-go currents of 134 men and 28 women. The
99.5 percentile rank levels (9mArms for men, 6mArms for women) were
considered to represent a reasonable criterion for safety (Dalziel and
Massoglia, 1956).
FIGURE 8.8. An experiment subject undergoes let-go current threshold testing.

(From Dalziel, (©) 1972, IEEE.)
TABLE 8.3. Summary of Dalziel's 60-Hz let-go current

threshold data.
0.5 Percentile
Subjects rank Minimum Average Maximum
134 men 9.0 9.7 15.9 22.0
28 women 6.0 7.4 10.5 14.0
Note: Values of let-go current in mArms.

Source: Adapted from Banks and Vinh, © 1984, IEEE.
99.99
99.9 ................... :...............•.......... ···················1························+.·.·•• ···••.....•...•..
99
2"
c: 95
as
II: 90
~
80
"E 70
Q)
~
Q)
50
~
Q) 30
0>
20
~
Q) 10
~
Q) 5
a.
.1
.01
6 8 10 12 14 16 18 20 22 24
Let-Go Current (in rnA rrns)
FIGURE 8.9. Let-go current percentile rank distribution curves for adult males and
females given hand contact with a 60-Hz AC source. Average value for 134 males =
lS.9mArms; extrapolated 0.5 percentile value = 9mArms. Average value for 28
females = 1O.SmArms; extrapolated 0.5 percentile value = 6mArms. (Adapted
from Dalziel, © 1972 IEEE.)
Great interest existed at this time in also establishing 60-Hz safety levels
for small children. In general, Dalziel and colleagues found that obtaining
let-go currents for young children was extremely difficult. In addition to the
reluctance of parents to permit experimentation on their children, Dalziel's
limited experiences with children indicated that children were likely to "just
cry at the higher values" (Dalziel, 1972). Despite these problems, results
were reported for three small boys (Table 8.4), and Dalziel recommended
a "reasonably safe" current level for children as 50% of the 99.5 percentile
value for normal adult males-namely, 4.5mArms.
TABLE 8.4. Dalziel's 60-Hz let-go threshold data for

three boys.
Let-go current
Subject Age (mArms)
1 5 years 7
2 9 years, 3 months 7.6
3 10 years, 11 months 9
Mean 7.9
Source: Adapted from Banks and Vinh, © 1984, IEEE.

fj)
I
<:
E
C
~
::J
U
0
Ol
Percentile curves
i; 40
...J
' \ 99.5
20
~~ 0.5
5 10 50 100 500 1000 5000
Frequency (Hz)
FIGURE 8.lD. The effect of stimulus frequency on the let-go current threshold for
adult males. Values on the left-hand axis are for DC. (Adapted from Dalziel, © 1972
IEEE, using some additional data from Dalziel et aI., 1943.)
The frequency dependence of tetanizing sinusoidal current thresholds

was reported in 1943 (Dalziel et al.) for groups of adult male subjects
ranging in number from 26 to 30 at frequencies from 5 to 10,000Hz as well
as at DC. These results were compared with the 60-Hz study on 134 men.
Figure 8.10 depicts the percentile curves for frequencies ranging from 5 to
1O,000Hz as well as at DC. 3 Tests on steady or gradually increasing DC
currents "produced sensations of internal heating rather than muscular
3 A somewhat unusual statistical method of "correcting" these curves, and DC data,

for variations in the 60-Hz let-go threshold mean value was used. For each group of
subjects tested at a given frequency, a 60-Hz let-go threshold mean was also found.
A "corrected mean" for each frequency was calculated as (the actual mean of the
sample group at the given frequency) X (the 60-Hz mean for the group of 134 men)/
(the 60-Hz mean for the current sample group). These corrected means were also
used in calculation of percentile ranks at each frequency. The correction factors
ranged from 1.00 (at 60Hz) to 1.09 (at 5Hz).
contractions. Sudden changes in current magnitudes produced muscular

contractions, and interruption of the current produced a very severe
shock .... The maximum [level of discomfort] a subject could take and
release was termed the release current. It represents the limit of voluntary
endurance rather than the let-go limit." The DC release current mean of 28
men was 73.7 rnA. After "correction" for variation in the 6O-Hz let-go
threshold for this group, the 99.5 percentile rank "reasonably safe" DC
current value for men was 62mA. Because at 60Hz the mean value for let-
go threshold values of women was about 66% of that for men (Table 8.3),
it was felt that the 99.5 percentile male values at 5 to 10,OOOHz, and at DC,
could be scaled by 0.66 to obtain estimates of reasonably safe values for
women.
8.6 Analysis of the Let-Go Phenomenon4

Analysis of factors that may be involved in the let-go phenomenon first
requires consideration of forearm/hand functional anatomy (especially as
regards power grip generation) and the current paths and impedances of
hand-contact accidents. The fundamental principles of neuromuscular elec-
trical stimulation presented previously in this chapter can then be utilized.
Functional Anatomy of the Hand and Wrist

A brief introduction to the anatomy and kinesiology of the hand and wrist
is essential to an understanding of how powerful gripping actions occur.
More detailed descriptions can be found in Gray's Anatomy (Gray, 1989),
Kaplan's excellent functional anatomy text (1984), and Rasch (1989). The
forearm (antebrachial) muscles are divided functionally into flexors and
extensors (anterior and posterior), although often complex interactions are
possible. Foremost among the flexors of this group that contribute to pow-
erful gripping are the flexor digitorum profundus muscle (the medial part of
which is innervated by the ulnar nerve and the lateral by the anterior
interosseous branch of the median nerve) and the flexor digitorum
superficialis muscle (innervated by the median nerve) (see Figs. 8.11 and
8.12). In a transverse section through the middle of the forearm (Fig. 8.12),
the relatively large fraction of cross-sectional area occupied by these two
muscles (52% of the area taken up by muscle, nerve, and blood vessels)
attests to their potential power. Maximal flexion of the hand thus can
overcome maximal extension. The intrinsic muscles of the hand are catego-
rized into three groups-those of the thumb, those of the little finger, and
those in the middle part of the hand. In gripping an object, the four small
4Portions of this section have been adapted from Sweeney (1993).

Analysis of the Let-Go Phenomenon 331
FIGURE 8.11. Anterior aspect of the left forearm showing deep muscles. (Adapted
from H. Gray, Gray's Anatomy, 30th ed., C.D. Clemente (ed.), Lea & Febiger, 1985,
p.533.)
lumbrical muscles (innervated by the median and ulnar nerves) in particular

provide flexion of the proximal segment of each finger (and under some
conditions, extension of the distal finger segments) (see Gray, 1989). We
can therefore hypothesize that, until the large flexor muscles of the fore-
arm are activated, electrically activated gripping actions using only the
intrinsic muscles of the hand can probably be overcome by use of the
forearm extensors. The let-go threshold should occur at about the point
where significant involuntary excitation of the forearm flexors has been
initiated.
Body Impedances and Current Paths in Hand-Contact Accidents

We have seen in Chapter 2 that the impedance properties of dry skin are
highly nonlinear and time varying. It is likely that the major source of skin
nonlinearity is the corneum. At the level of electrical line voltages, a skin
impedance breakdown occurs fairly rapidly. As depicted in Fig. 2.12, for
example, 50-Hz AC voltage stimulation results in an initial rapid drop in
impedance within a fraction of a minute, followed by a slower drop in the
first minute, and then more gradual decreases in the next 7 to 8 minutes
until a plateau value is achieved [see Sect. 2.2 and Freiberger (1934)]. This
phenomenon may partially explain Dalziel's observation that, near the let-
go threshold, failure to promptly interrupt current flow results in a rapid
inability to do so.
Although the total internal body impedance, pathways, and current den-
sities that occur in sinusoidal voltage electrical stimulation are highly
dependent upon the voltage magnitudes and frequencies as well as
electrode(s) size and placement, some useful generalities can be drawn.
Hand-to-hand and hand-to-foot internal impedances for line voltages and
frequencies are in the range of 500 to 7500hms. As can be seen from Figs.
2.20 and 2.21, about 50% of the total internal body impedance in this
situation can be attributed to the wrists and/or ankles. Cross sections
Flexor carpi Palmaris

Anterior interosseus radialis longus
nerve and artery
Brachioradialis
Flexor digitorum
Radial artery superflcialis
Radial nerve Ulnar artery

Lateral cutaneous Ulnar nerve lying on
nerve of forearm flexor digitorum
profundus
Extensor carpi
radialis longus Flexor carpi ulnaris
Extensor carpi
radialis brevIs Flexor digitorum
profundus
Radius
Ulna
Abductor
pollicis longus
Extensor pollicis
Extensor digitorum and Extensor carpi longus
extensor digiti minimi ulnaris
Flexor
digitorum
Radial artery Ulnar nerve

Flexor poUicis longus Flexor carpi ulnaris
Flexor digitorum profundus
Articular disc
Radius
StylOid process of ulna
Extensores Extensor Extensor

carpi po!licis digitorum
radiales longus and indicis
FIGURE 8.12. Transverse cross section through middle of left forearm (upper) and
wrist (lower.) [Adapted from Gray, 1989, p. 620 (upper) and p. 625 (lower.)]
through these areas (e.g., Fig. 8.12) reveal that relatively poorly conducting
bone, tendon, fat, and ligament dominate. Current densities in these regions
should be highest in the existing higher conductivity soft tissues (nerve
trunks) and blood vessels. We would therefore expect that in hand-contact
electrical accidents, current density will be very focused directly beneath
the electrode in the muscle and nerve of the palm and within the median
and ulnar nerve trunks in the wrist.
To a first approximation, current densities beneath a cylindrical object

held in the hand can be estimated by considering the surface area of the
electrode (although if the rod is not completely enclosed by tissue, current
densities will be highest at the edges of the conductor). The No.6 A WG
wires used by Dalziel and colleagues would have an approximate surface
area of 11.1 cm2 when held by an average adult male, assuming a typical
hand breadth of 8.6cm (McConville et aI., 1980) and No.6 A WG diameter
of 0.41 cm. Given a total current 1 (in rnA), a current density of 0.091 (mAl
cm2) would therefore occur close to the electrode surface. For adult fe-
males, average hand breadth is approximately 7.8cm (Young et. aI., 1983),
so current density would be 0.10 1 (rnA/cm2). Typical hand width of a 5-
year-old boy is about 5.7 cm (Synder, 1975), so current density would be
0.141 (mA/cm2). An estimate of current densities within the median and
ulnar nerve trunks near the wrist can be obtained from Fig. 8.12. Tissues
other than bone, cartilage, fat and tendon in Fig. 8.12 comprise less than 5%
of the wrist cross-sectional area at this level. From anthropometric esti-
mates of wrist circumferences (17 cm for adult males (McConville et aI.,
1980); 15.7cm for adult females (Young et aI., 1983); and about 11cm for
small children), we can roughly calculate wrist cross-sectional areas making
the simple assumption that all wrists have the elliptical shape of Fig. 8.12.
Performing these calculations yields estimates of 21.3, 18.2, and 8.9cm2 for
adult males and females and for small children, respectively. For total
current I, current densities of 0.94 1,1.10 I, and 2.25 1 (mA/cm2) result for
men, women, and children, making the simple approximations that most
current will flow in the nerve trunks and blood vessels at this level, that
these tissues have roughly the same conductivity, and that such tissues
comprise approximately 5% of the total cross-sectional wrist area regard-
less of sex or age. This is in contrast to densities for the more proximal
forearm.
From Fig. 8.12, we see that skeletal muscle dominates the cross-section at
the midforearm, and highly conductive tissues (muscle, nerve, and blood
vessels) comprise about 72% of the cross-sectional area in adult males.
Cross-sections of the male forearm from other anatomy sources give similar
estimates of combined muscle, nerve, and blood area percentages (e.g.,
73% in Sect. 57 of Morton, 1944; 73% in Sect. 56 of Carter et aI., 1977).
Adult females on average have a higher percentage of subcutaneous fat in
the extremities as compared to adult males (25% as opposed to 12% as
reported by Malina, 1975). Analysis of cross-sections in the midforearm of
adult females (level FI in Kieffler and Heitzman, 1979; Sect. 11-21 in Bo et
aI., 1990) as well as anthropometric estimates based on 50th percentile data
for 17-year-old females (McCammon, 1970) both yield skeletal muscle
cross-sectional area estimates of about 61 %. Using this information from
Fig. 8.12, and anthropometric data on typical midforearm circumferences
[23.4cm for adult males (McConville et aI., 1980); 21.2cm for adult females
(Young et aI., 1983); 16cm for 5-year-old children (Synder, 1975)], we can
estimate current densities within the relatively conductive tissues at this

level. Assuming the elliptical shape of Fig. 8.12 is typical regardless of sex or
age, current densities of 0.032 I, 0.047 1, and 0.069 1 (mAlcm2) result for
men, women, and children, respectively (using an estimate of 72% for
combined muscle, nerve, and blood cross-sectional area fraction in children
and adult males and 61 % for adult females).
Excitation Threshold Estimates

Recall from Sects. 4.5 and 8.4 that, to a first approximation, changes in
the extracellular electric field (Le., the change in voltage gradients or
second difference of voltage values) along an excitable fiber provide
the driving force for eliciting transmembrane potential shifts. Given
current density estimates (1), we can calculate voltage gradient values
('Vcf») using the relation 'Vcf> = fQ, where Q is an estimate of bulk tissue
resistivity.
Making use of the current density estimates derived in the previous
section, we can therefore predict analogous voltage gradients if we assume
a certain tissue resistivity value. Although the conductivities of both skel-
etal muscle and peripheral nerve trunks are anisotropic, with both a spatial
and temporal dependence (see Chapter 2), we will use 200Qcm as an
approximate value for the lumped resistivity for muscle, nerve, and blood
vessels with longitudinally oriented current at power line frequencies (see
Table 2.1 and Sances et aI., 1983). Voltage gradients immediately adjacent
to a No.6 A WG wire (passing a current "I" in mA) held in the hand of adult
males and females or small children (5-year-old boys) become 18 I, 20 I, and
28 I (mV/cm), respectively. In the nerve trunks of the wrist, we would
expect voltage gradients of 188 I, 220 I, and 45 I (mV/cm); and at the mid-
forearm, 6.4 I, 9.4 I, and 13.8 I (mV/cm) (for men, women, and small
children, respectively).
As we have seen in Chapter 4, not only must a longitudinally oriented
voltage gradient exist along myelinated nerve fibers for electrical stimula-
tion to occur, but a change in voltage gradient is necessary. The terms
(Ve,n.i - 2Ve,n + Ve,n+;) in Eq. (4.31) imply that a second-difference sampling
of nodal extracellular voltages provides the driving force for excitation.
That is, a change (difference) in two first-difference terms ({Ve,n-i - Ve,n}
and (Ve,n+i - Ve,n}) is needed. Each first-difference term is essentially a
linear sampling of the extracellular voltage-gradient between two nodes of
Ranvier. If the electric field along a nerve fiber is truly uniform, then in
theory, excitation cannot occur. Realistically, in this situation, spatial
orientation changes (e.g., bends) in nerve fibers tend to introduce apparent
second-difference effects. Likewise, terminations of motor myelinated
nerves at end-plates, or sensory axons at receptors, allow excitation within
a uniform field (see Sect. 4.5).
The peripheral myelinated nerve fibers of a-motor neurons range in

diameter from approximately 9 to 20,um (see Sect. 8.2). If we assume a
factor of 100 difference between fiber diameter and internodal length
(Table 4.1), large (20-,um) fibers will have internodal lengths equal to about
2mm, while small (lO-,um) fibers will have internodal lengths of about Imm.
As we have seen in Sect. 8.4, large fibers tend to be the most easily stimu-
lated by fairly diffuse extracellular electric field distributions, leading to the
phenomenon of "reversed recruitment." With either the myelinated nerve
model described in Chapter 4 or compartmental cable models of mamma-
lian myelinated nerve (Sweeney et aI., 1987), the author has found that
second-differences of approximately SOmV are needed to bring lengths of
nerve fibers to threshold with long pulsewidth (rheobase) stimuli. A 20-,um
diameter nerve fiber lying within a uniform voltage gradient of 2SmV/mm
would experience a second-difference of SOmV if the nerve abruptly turned
by 900 with respect to the field (similarly, SOmV/mm will excite a lO-,um
diameter fiber). In Sect. 4.S, it is reported that, in a terminated nerve model
(e.g., roughly representative ofaxons ending at a motor end-plate), a volt-
age gradient of 6.2mV/mm suffices to excite a 20-,um diameter myelinated
nerve (12.3mV/mm for a 10-,um fiber). These lower values presumably
result from the abrupt impedance transition at the cable end. All of the
above voltage gradients apply for square wave stimuli having durations
several times greater than the chronaxie. For 60-Hz sinusoidal stimuli, the
peak threshold current amplitudes are about 17% greater than those of
square-wave stimuli (Fig. 4.1S).
Directly under a hand-held electrode, or within the midforearm, we will
therefore hypothesize that a voltage-gradient of approximately 7.3mV/mm
(117% of 6.2mV/mm) or S.2mV-rms/mm is necessary to achieve threshold
stimulation of the largest (20-,um diameter) myelinated axons terminating
at motor end-plates of the lumbrical muscles in the hand or the large flexor
muscles of the forearm (1O.2mV-rms/mm for lO-,um fibers). For initial
excitation of the median and ulnar nerve trunks within the wrist, we will
hypothesize that a rheobase voltage gradient of 29.2mV/mm (117% of
2SmV/mm) or 20.6mV-rms/mm is needed assuming that a "bending" ofthe
nerve trunks with respect to the electric field occurs as the nerve trunks
spread into branches, and the current distributes within the hand
(41.2mVrms/mm for lO-,um fibers).
Table 8.5 lists the current density and voltage gradient estimates derived
above for an arbitrary current I (in rnA) and the resultant estimates of
absolute current levels needed to achieve threshold excitation in 20-,um and
10-,um diameter nerve fibers. The results in Table 8.5 predict that, at current
values on the order of ImA (l.lmArms for males, 0.94mArms for fe-
males), typical adult males and females should experience initial, relatively
small contractions of intrinsic muscles in the hand (such as the lumbricales)
due to excitation of 20-,um diameter fibers in the nerve trunks of the wrist.
TABLE 8.5. Analysis of Hand-Contact skeletal muscle stimulation.

Current Voltage 20-.um Fiber 10-.um Fiber 60-Hz let-go
density gradient threshold threshold current
(mA/cm2) (mV/cm) (mArms) (mArms) (mArms)"
Directly beneath No.6 A WG electrode in lumbrical muscles of the hand
Adult males 0.091 181 2.9 5.7
Adult females 0.101 201 2.6 5.1
Small children 0.141 281 1.9 3.6
At median and ulnar nerve trunks within the wrist
Adult males 0.941 1881 1.1 2.2
Adult females 0.101 2201 0.94 1.9
Small children 2.251 4501 0.46 0.92
At the large flexor muscles of the midforearm
Adult males 0.0321 6.41 8.1 15.9 15.9
Adult females 0.0471 9.41 5.5 10.9 10.5
Small children 0.0691 13.81 3.8 7.4 7
"Estimates of Dalziel for adult males and females (Dalziel and Massoglia, 1956); single study
on a 5-year-old boy as reported by Dalziel (1943).
Note: I is an arbitrary hand-contact current in milliamperes.
These estimates can be compared with Dalziel's studies on perception

current thresholds for hand-contact electrical stimulation. [Recall from
Chapter 7 and Sect. 8.2, that large diameter Aa sensory nerve fibers will
also be present in the nerve trunks of the wrist. For example, muscle spindle
primary endings are innervated by type IA fibers which average about
17.um in diameter and can be 20.um (see Fig. 3.12)]. Our predictions
correspond quite well with the actual 60-Hz perception threshold
mean values (1.1mArms for adult males and O.7mArms for females)
(Dalziel, 1972). Somewhat surprisingly, current densities directly beneath a
No. 6 A WG electrode should result in initial muscle activation only at
somewhat higher thresholds (e.g., 2.9mA for adult males). These values
probably reflect overestimations of the actual values because, in reality, the
edges and proximal surface of the electrode will carry the majority of the
current.
Dalziel and colleague's mean values for the 60-Hz let-go threshold cur-
rents of adult males and females (i.e., 15.9 and 10.5mArms, respectively)
and even Dalziel's single value for the let-go threshold of a 5-year-old boy
(i.e., 7mArms) lie within the bounds of our Table 8.5 estimates for thresh-
old stimulation of large and small a-motor neuron fibers innervating the
forearm flexors (compare the last three columns of the table for the last
three rows). This is in keeping with our hypothesis that the let-go threshold
should reflect the point at which forearm flexor activation reaches a level
where it cannot be voluntarily overcome. Because of a combination of
anthropometric (i.e., reduced mean forearm circumference) and body com-

position factors (i.e., higher mean percentage of subcutaneous fat), our
estimates predict that adult females should have let-go thresholds about
69% of those of adult males. Recall that Dalziel and colleagues estimated
that female thresholds would be about 66 % of those of males. Our estimate
for the let-go threshold range in a 5-year-old child (3.8 to 7.4mArms) is
reasonably close to the value of 7mArms for a 5-year-old boy as reported
by Dalziel (1943). This current range estimate is 47% of our predicted adult
male range, a factor in good agreement with the 50% factor originally
proposed by Dalziel (1943).
We can in addition predict the effects of physique on let-go threshold
values for adult males and females by constructing theoretical percentile
rank distributions for let-go currents. These theoretical distributions
can then be compared to Dalziel's data (i.e., Fig. 8.8). We first adopt the
mean let-go current values of Dalziel (15.9mArms for males; 1O.5mArms
for females) as our 50th percentile estimates. For adult females, a mean
let-go current value limit of 1O.5mArms (analogous to the lO-,um fiber
threshold value in Table 8.5 for the large flexor muscles of the forearm)
rather than 10.9mArms would have been obtained above if we had used
an estimate of 58.5% rather than 61% for the fraction of highly con-
ductive tissues in the forearm. Next, we can make use of forearm circum-
ference distribution mean and standard deviation data (adult male data
from Meonville et al., 1980 and adult female data from Young et al.,
1983) to construct theoretical let-go percentile rank distributions for
the same population sample size used by Dalziel when he performed the
studies of Fig. 8.9 (i.e., N = 134 for men and N = 28 for women). These
theoretical curves as seen in Fig. 8.13 appear to incorporate much of
the behavior of Dalziel's data set. More accurate modeling of let-go
thresholds for children could also make use of known relations between
anthropometric measures and the child's sex and age. For example, it is
known that during development, the relative composition of bone, fat, and
muscle in the arm tend to change (see Johnston and Malina, 1966). As
Banks and Vinh (1984) have emphasized, a general safety standard for
children would have to take into account that children range in age and size
from infants to adults.
Thus, while Dalziel and colleagues (1956) found that their attempts to
correlate let-go currents with physical measurements of the forearm, wrist,
strength of grip, and so on, were inconclusive, our theoretical estimates
of let-go thresholds based on information from functional anatomy,
anthropometry, bioelectricity, and electrical stimulation theory appear to
indicate that such correlations should be possible. Measurements of let-go
thresholds, or at least single sinusoidal pulse electrical stimulation thresh-
olds, in individuals where radiographs or magnetic resonance imaging
(MRI) scans of the forearm were also available, would enable more de-
tailed future investigation of such correlations.
99.99
99.9 . . . . . . . . . . . . . . . . . . . . . . . .l. . . . . . . . . . . . i........................i........................ ' ..............................................i........................t. . . . . . . . . . ..

. .
99 .............................-............................ ········+·······················t····...............:........................
~-~-~~~;~~~;:~j~~~~-~:i~=:¥~~=
~
c: 95
ttl ..............."\" .......................
a: 90
l!!
80
~ 70
CD
f::!
CD
e:.. 50 ·. ··1··
CD 30
Cl
20 ·············t················
~CD 10 ···········t·················· ................+·····················t·········
f::! 5
CD
a. 1
-_·······t--_···
.1
.01
6 8 10 12 14 16 18 20 22 24
Let-Go Current (in rnA rrns)
FIGURE 8.13. Theoretical let-go current percentile rank distribution curves for 134
adult males and 28 adult females. Linear regression lines are those of Dalziel and are
the same as those in Fig. 8.9. (Adapted from Sweeney, © 1993 IEEE.)
8.7 Effects of Electrical Stimulation on Respiration

It has been long recognized that accidental electric shocks could inhibit or
prevent the ability to breathe. In 1913, lex-Blake (as cited by Lee and
Zoledziowski, 1964) listed three processes that could lead to death from
electric shock: ventricular fibrillation, prolonged respiratory arrest that
outlasts the shock itself, and asphyxia resulting from tetanic contraction of
the respiratory muscles during the shock. Electrical stimulation can also be
used beneficially in high level quadriplegics with respiratory paralysis as
well as in patients with central alveolar hypoventilation to restore respira-
tion (see Nochomovitz, 1983; Creasy et aI., 1996). In fact, electrical stimula-
tion of the phrenic nerves for artificial respiration was postulated over two
centuries ago by Hufeland (1783).
Anatomy and Physiology of Respiration

Normal low-intensity breathing is performed almost entirely by movement
of the diaphragm for inspiration and by passive elastic recoil for expiration.
The diaphragm is composed of skeletal muscle and dense collagenous con-
Effects of Electrical Stimulation on Respiration 339
nective tissue. It is innervated by the phrenic nerve which derives from C3

to C5 cervical plexes. Inspiration can also be achieved by contraction of
skeletal muscles that raise the rib cage (sternocleidomastoids, anterior
serrati, scaleni, external intercostals). The abdominal recti and internal
intercostal skeletal muscles may be used during forced expiration to depress
the chest cage. The neural "respiratory centers" of the brain stem provide
the control and drive for respiration.
Accidental Electric Shocks and Respiration

In several of his review papers on the dangers of electrical shock, Dalziel
reported anecdotally that in his group's experiments on the "let-go" phe-
nomenon, they had observed that 60-Hz electrical stimulation with hand-
held electrodes could inhibit or stop respiration. Dalziel and Lee (1968)
reported that "the muscular reactions caused by commercial frequency
alternating currents in the upper ranges of let-go currents, typically 18 to
22mA or more, flowing across the chest stopped breathing during the
period the current flowed, and in several instances caused temporary pa-
ralysis of the middle finger. However, normal respiration resumed upon
interruption of the current, and no adverse after effects were produced as a
result of not breathing for short periods." This possible danger of accidental
electrical shock, as well as that of cardiac fibrillation (Chapter 6), is often
cited as justifying the importance of let-go thresholds and the establishment
of safety standards at currents well below the let-go range. An individual
who is "frozen" to a conductor by his inability to release it cannot easily
save himself. "Prolonged exposure to currents only slightly in excess of
a person's let-go limit may produce exhaustion, asphyxia, collapse and
unconsciousness followed by death" (Dalziel, 1959).
In the case of small children, as we have seen in Sects. 8.5 and 8.6, let-go
thresholds can be on the order of7 to 9mA or less (see Table 8.4). In 1940,
it was reported that a 4-year-old boy had died after contacting an electric
fence wire (lAEI News Bulletin, 1940). It was assumed that his hand
"froze" to the conductor, although Dalziel recognized that small children
exposed to painful electrical stimulation might sometimes not even attempt
to extricate themselves from an electrical source (Dalziel, 1972). Tests
performed on the electric fence power supply indicated that a current of
9mA was delivered when a load of 100Q was placed across the erminal
(8.4rnA for 1,000Q; 8rnA for 2,OOOQ; 6.8mA for 5,000Q).
In 1961, Lee reported a clinical study of 104 electrical accidents. In 30
cases, individuals were "frozen" to the conductor. "The longer the victim
was held on to the circuit the greater appeared to be his chances of devel-
oping heart and chest symptoms suggestive of impending asphyxia, and of
losing consciousness.... Artificial respiration was administered in two
cases, one of whom was 'held on' and was being asphyxiated." Greenberg
(1940) found that, in forelimb to hindlimb electrical stimulation of dogs,
currents of up to 50mA caused arrest of respiration during the period of

current flow; however, respiration restarted as soon as the current was
removed. Lee and Zoledziowski (1964) studied the effects of electric shock
on the respiration of rabbits. With currents up to about 200mA, respiratory
arrest appeared to be due to muscular contraction in the chest. Only with
exceptionally high limb-to-limb current levels, or with current paths that
include the head, does it appear possible to electrically bring about an arrest
of respiration that persists after the current is switched off (see review of
Lee, 1964). It is presumed that this effect is caused by electrically induced
damage to the brain stem respiratory centers.
Artificial Respiration by Electrical Stimulation

Dr. William Glenn and colleagues at Yale University, starting in 1959,
developed the first practical long-term phrenic nerve stimulation system for
phrenic "pacing" (e.g., Glenn and Phelps, 1985; Glenn et aI., 1986 and 1988)
of quadriplegics. Several commercially available phrenic nerve stimulation
systems now exist which utilize implanted electrodes on or adjacent to the
phrenic nerves-with over 1,000 people worldwide having received such
implants (Creasy et aI., 1996). An approach wherein phrenic nerve stimula-
tion electrodes might be placed within the diaphragm (Peterson et aI., 1986
and 1994a,b) has also been proposed and studied. Epidural spinal cord
stimulation can be used to activate the intercostal muscles (e.g., DiMarco
et aI., 1987 and 1994). Geddes and colleagues (Geddes et aI., 1985, 1988
and 1990; Riscili et aI., 1988 and 1989a,b) have also developed an
"electroventilation" method that uses surface electrodes located in the
axillary region. It is thought that this technique also primarily stimulates the
phrenic nerves. The same research group has also proposed the use of
electromagnetic stimulation of the cervical phrenic nerves as a possible
means of inducing artificial respiration (e.g., Geddes et aI., 1991).
9
Stimulation via Electric and
Magnetic Fields
9.1 Introduction
Electromagnetic (E-M) fields are ubiquitous in modern society. Radio fre-
quency fields, for example, are produced by a variety of communications
devices. Extremely low-frequency fields are produced by power transmis-
sion and distribution lines, as well as by home wiring and appliances. Some
environmental exposures can result in measurable short-term reactions.
The electric fields from high-voltage transmission lines, for example, can
provide the opportunity for electric shock under some circumstances. The
new generation of magnetic resonance imagers press the limits where per-
ceptible effects might be caused by their time-varying gradient fields. Nu-
merous research efforts have focused on biological effects associated with
both short-term and chronic exposure to low-level E-M fields. Although
chronic exposure issues are beyond the intended scope of this book, bio-
physical mechanisms proposed to account for chronic effects are treated
briefly in Chapter II.
This chapter concentrates on short-term human reactions to electric and
magnetic fields whose wavelengths are long in comparison with the dimen-
sions of the human body. By concentrating on long-wavelength phenom-
ena, it is possible to treat electric and magnetic field effects separately. This
approach differs from the study of high-frequency electromagnetic radia-
tion, where the electric and magnetic components are intimately connected.
In the long-wavelength regime, the term "electromagnetic field" is often
replaced by "electric and magnetic fields."
9.2 Electric Field Induction Principles

A steady electric field produces surface charges on conducting objects,
including the human body. Positive charges are accumulated on the side of
a conducting body nearest to the negative source of the field, and negative
charges on the side nearest the positive source. If the field is alternating, the

342 9. Stimulation via Electric and Magnetic Fields
positive and negative charges alternate in position, resulting in an alternat-

ing current within the biological medium.
It is important to distinguish between direct and indirect E-field induction
effects. The former refers to charges and currents induced directly in the
body. The latter refers to current or charge transfer between a person and
a conducting object within the field. Before proceeding further, consider
physical induction mechanisms.
Induction Mechanisms
Consider a conducting object situated above a ground plane in a vertically
oriented electric field. If the object is electrically connected to ground, it will
acquire a net charge, which, for simple geometric shapes (flat plate, hori-
zontal cylinder, sphere), can be expressed by (Deno, 1975a, 1975b).
Q = VsCo (9.1)
where Co is the capacitance between the object and ground, and Vs is the
space potential that would be present at the centroid location of the object
if it were not present. In a low-frequency alternating field where the electri-
cal wavelength is much greater than the dimensions of the object, one can
ignore propagation effects and resort to the so-called quasistatic solution
(Kaune, 1981). In that case, the induced alternating current is given by the
time derivative dQldt, which, for a sinusoidal field, is
(9.2)
where f is the frequency of the field, and j is the complex phaser operator
(indicating 90° phase shift with respect to the E-field). In a uniform field, an
equivalent expression for Eq. (9.2) is obtained by noting that Vs = Eh:
Is = j2nfEhCo (9.3)
where E is the electric field strength, and h is the height above the ground
plane at the electrical centroid of the object. The terms in Eq. (9.3) can be
conveniently regrouped as
(9.4)
where Co is the permittivity of free space (8.85 x lO- 12 F/m). The phasor
operator j has been dropped, because we will be dealing with magnitudes.
The second term in Eq. (9.4) expresses the area of a parallel-plate capacitor
with small separation and having capacitance Co (see Eq. 2.4). This leads to
the equivalent-area concept for describing induced current (Deno, 1975b):
(9.5)
For irregularly shaped objects, Ae is the area of a flat plate that would
result in the equivalent induced current. Methods for evaluating Eq. (9.5)
Electric Field Induction Principles 343
have been determined for a variety of objects and shapes (Deno, 1975a,
1975b; Deno and Zafanella, 1982; Reilly, 1979a, 1979b, 1982; Reilly and
Cwicklewski,1981).
Another parameter of interest is the open-circuit voltage V o, i.e., the
voltage that would be measured between the object and the ground plane in
the absence of current conduction to ground. Vo can be determined from
the short-circuit current by
(9.6)
where Zo is the impedance measured between the induction object and
ground. For an object that is perfectly insulated from ground, Zo is deter-
mined by its coupling capacitance to ground. In this case, Vo achieves its
maximum possible value, given by
Vo{max) = _Is_ (9.7)

21CfCo
where Vo and Is apply to magnitudes, without regard to phase [the phasor
operator j has been dropped from Eq. (9.7)].
Table 9.1 provides examples of the capacitance, normalized induced
current (lIE), and normalized open-circuit voltage (VolE) for several
objects in a 60-Hz field of strength E. References cited in the table discuss
measurements involving these objects in E-fields.
The value of Is given by Eqs. (9.2) to (9.5) expresses the so-called short-
circuit current (i.e., the current that would flow in a grounded conductor
that contacts an otherwise insulated object). When a grounded person
provides the ground path, the current that flows at the points of contact may
result in a perceptible electric shock. An insulated person touching a
grounded connection would also result in current flow, but here the induc-
tion object would be the person.
TABLE 9.1. Example induction parameters of various

objects in 60-Hz electric fields (object on insulated
surface).
liE VIE CO
Object (A V- 1m) (m) (pF)
Person (1.8m tall) 1.8 x 10-8 0.48 100-150
Auto (midsize) 9.1 X 10-8 0.21 200
Auto (large) 1.2 X 10-7 0.16 2,000
School bus 3.7 x 10-7 0.26 3,700
Commercial bus 4.9 X 10-7 0.45 2,900
Fence wire (ht = 1 m, 2.7 X 10-7 1.0 720
length = 100m)
Source: Data from Bracken (1976), Reilly (1982), Reilly

(1979a).
We can gain some insight into the induction on a person by using the
theoretical expression for short-circuit current in an upright cylinder
(Reilly, 1978a):
(9.8)
where h is the length of the cylinder, r is its radius, d is the distance from the
bottom of the cylinder to the ground plane, f is the frequency of the field,
and E is the strength of a vertically oriented E-field. [Units in Eq. (9.8) are
amperes, Hertz, meters, and volts/meter]. We have verified Eq. (9.8) with
measurements on cylinders placed on a ground plane within a 60-Hz elec-
tric field (Reilly, 1978a).
If we model a person as an upright cylinder with a height-to-radius ratio
of 12 and standing on a ground plane (d = 0), then Eq. (9.8) predicts a value
of short-circuit current given by:
(9.9)
When evaluated at 60Hz, Eq. (9.9) is identical to the empirical formula
determined by Bracken (1976) for people within transmission-line electric
fields.
Whereas Eq. (9.9) expresses the short-circuit current at the point of
ground contact, an investigator may wish to know the current density in a
person who is immersed in an alternating E-field. Experimentors, using
conducting models of the human body, have mapped the distribution of
internally generated currents for both grounded and ungrounded humans
standing within vertically oriented 60-Hz E-fields (Deno, 1977; Kaune and
Forsythe, 1985). Calculations over a wider frequency range and for more
complex field environments have also been carried out (Takemoto-
Hambleton et aI., 1988; Hart, 1992; Gandhi and Chen, 1992). Consider, for
example, a person standing within a vertically oriented alternating E-field
and grounded through the feet. In this case, the amount of current passing
through any given horizontal cross section of the body increases as the
cross-section is moved from the head toward the feet, with the greatest
current passing through the feet as illustrated in Fig. 9.1 (adapted from
Deno, 1977). The horizontal axis expresses the fraction of maximum cur-
rent, given by Eq. (9.9), passing through a cross-section at the indicated
height. The outermost curve applies to a person grounded through the feet;
the other curve applies to an ungrounded person standing 13 cm above a
ground plane. The maximum current in the grounded person flows through
the feet. For an ungrounded person, maximum induced current flows
through the midsection of the body.
The current density in a homogeneous model is the induced current
divided by the cross-sectional area. One can infer from Eq. (9.9) that the
Person
grounded
through feet
.8
1:
.2>
Q)
.r:;
.6
>-
B
'0
c
.2
U
~ .4
u..
1.0
Fraction of Is
FIGURE 9.1. Distribution of current flowing in a person standing in a vertically

oriented alternating E-field. Horizontal axis describes fraction of maximum short-
circuit current passing through cross section at indicated height. (Adapted from
Deno, 1977.)
current density induced in the upright body would not depend on a person's
height, as long as the ratio hlr is held constant. This inference can be drawn
by noting that the induced current and cross-sectional area of the model
cylinder both grow in proportion to the square of height.
Body Surface Effects

Environmental E-field strengths are typically reported as undisturbed field
levels (i.e., the intensity that would be measured in the absence of a biologi-
cal subject). The presence of a conducting object will distort the undis-
turbed field, such that the field incident on the surface of the object will be
significantly modified. The effect is illustrated in Fig. 9.2 for a person in a
vertically oriented uniform field. The orientation of the E-field is indicated
FIGURE 9.2. Perturbation of environmental electric field by a person. External elec-

tric field indicated by "flux lines" outside body. Arrows inside body roughly indicate
direction and intensity of induced internal electric field and body current. Direction
of arrows reverses everyone-half cycle. Field lines are illustrative and not drawn to
scale.
by the direction of the flux lines, and its intensity by their density (not drawn
to scale in the figure). The flux lines are normal to the surface of the body.
The lines drawn within the body roughly represent the internally induced
current.
As suggested by the figure, the E-field on the surface of the body will be
enhanced in certain regions, particularly on the upper body, and reduced in
other regions. The average field, E a, over the entire body is related to the
short-circuit current by (Kaune, 1981)
(9.10)
where Ab is the body surface area. Notice the similarity of Eqs. (9.5) and
(9.10).
Ea will be exceeded at particular regions of the body, particularly at
elongated surfaces on the uppermost body. Table 9.2 expresses field-
enhancement factors applying to the average, or to the topmost region of
humans and animals. The table also includes normalized values of short-
circuit current, applicable to 60-Hz fields. The enhancement factor indicates
the value of the incident field as a ratio of the undisturbed field. For
example, the field at the top of a person's head will be 18 times the undis-
turbed field measured in the absence of the person. If the person were to
raise one hand above the head, the electric field on the raised hand would
be enhanced even further.
Table 9.2 applies to the situation where the source of the E-field is several
body dimensions distant from the sUbjects. If, instead, the source were close
to the subject, there would exist a further enhancement of the incident field
because of the interaction of the charges on the body surface, and the
charges on the conducting object.
Induced Electric Shock Principles 1

Figure 9.3 illustrates a conceptual model of ELF electric field induction. An
AC high-voltage source, such as an electric power transmission line, is
capacitively coupled to a charging object through Cs • The object is also
capacitively coupled to ground through Co, and may include leakage resis-
tance, indicated by Ro. The maximum sinusoidal current that would flow
in a low-impedance connection to ground is the short-circuit current Is
discussed above.
TABLE 9.2. Electric field enhancement and induced

current for several species.
Field
enhancement lJE
Species Avg Top [A/(Y/m)]
Human (standing) 2.7 18.0 1.6 X 10- 9
Swine 1.4 6.7 7.0 x 10- 10
Rat (resting) 0.7 3.7 1.2 X 10- 11
Rat (rearing) 1.5 * 2.4 X 10- 11
Horse 1.5 * 2.7 X 10- 9
Cow 1.5 * 2.4 X 10- 9
* Data not available.

Note: "Avg" refers to field enhancement averaged over body.
"Top" refers to field enhancement at top of body. I)E is
normalized short-circuit current at 60Hz.
Source: Data from Kaune (1981).
1 Portions of this section have been adapted from Reilly and Larkin (1987).
High-voltage
source
Charging object
(e.g., vehicle, fence)
Discharging object
(e.g., person)
FIGURE 9.3. Equivalent circuit model of electric field induction.
The voltage and charge developed across Co can be discharged through a

conductive path to ground, such as through a grounded person, represented
by the impedance Zp. The capacitor discharge may be conducted through
direct contact with the skin, or, with sufficient voltage, through a spark. The
critical field strength required to support an electrical discharge in air
depends on the material of the two conductors, and, in accordance with
Paschen's law, on the product of electrode separation and atmospheric
pressure (Howatson, 1965). The voltage needed to conduct a spark dis-
charge to dry, intact human skin is in the vicinity of SOOV (see Sect. 2.5).
After direct contact is made with a charged object, a sinusoidal current Ip
will flow into the body. For many cases of practical interests, Ip "" Is (Reilly,
1979b), where Is is the short-circuit current.
Figure 904 shows examples of voltage waveforms measured on an object
energized by a 60-Hz source and contacted by a subject's finger. These
examples were produced using equipment that simulates waveforms ob-
served in field conditions (Reilly and Larkin, 1987). Each record shows
a combination of capacitor spark discharges, indicated by abrupt dis-
continuities in the voltage trace and a period of continuous sinusoidal
current while the finger rests on the electrode. Current spikes can be seen
in coincidence with the abrupt changes in voltage, but the magnitude of
these spikes is not represented accurately, because the sampling rate in
these records (100,us) was relatively slow. In Fig. 904a, Vois 1,000Vrms. A
capacitor discharge occurs upon contact with the electrode at t "" 20ms. In
Fig. 904b, Vo is 2,000 V rms. Again, there is a single prominent contact dis-
charge, but there are also several discharges just as the finger is withdrawn
(t "" S7ms). Expanded waveforms of capacitive discharges are shown in
Figs. 2.29 to 2.33. If the voltage at the moment of contact is high enough
(about SOOV), the discharge occurs through a spark; below the critical
voltage, the discharge occurs only at the moment of physical contact with
the electrode.
Both the capacitor discharge voltage and the sinusoidal contact current
are important components of the AC field-induced stimulus. These quanti-
ties are related in a well-defined way:
= Zo Va (9.11)
Is
where Zo is the impedance of the charging object. For many cases of
interest, Cs « Co and Eq. (9.11) can be written:
Approach Contact Withdrawal
(a) Vo = 1000 V.
-2.0 0
100
Elapsed time (ms)
4.0 I
.1 '
I
I
!
2.0
>
.>.!
I
Q)
0 ,
...'"'"
(5
\
I I
> I
I
-2.0 \ i , I
1 J "I
\ i \1'
(b) Vo = 2000 V. U
-4.0
0 100
Elapsed time (ms)
FIGURE 9.4. Example waveforms for AC field-induced stimuli, with tapped finger
contact of energized electrode Co = 200pF.
(9.12)
where f is the sinusoidal oscillation frequency of the electric field. Absolute

values are indicated because Eq. (9.12) applies to the magnitude of Vo and
Is without regard to their relative phases.
The condition of greatest induced voltage on the charging object occurs
when leakage resistance is large, that is, 21CfRoCo » 1. In this case, Eq.
(9.12) reduces to
VoI-= 21CfC
II: 1
o
(9.13)
In many practical cases, however, Romay be small enough that the induced
voltage is substantially below its theoretical maximum (Reilly, 1979b). In
such cases, the ratio Volls, given in Eq. (9.12), becomes smaller than its
theoretical maximum (i.e., the open-circuit voltage becomes less promi-
nent). From a sensory point of view, the sinusoidal current in the contact
phase of the stimulus may take on greater significance than the capacitive
discharge component.
In this chapter, the effects of leakage resistance are expressed in terms of
the factor k = IVollsl(21CfCo). A functional expression for k can be obtained
by mUltiplying Eq. (9.12) by 21CfCo:
k = [1 + ------:-1]-1/2 (9.14)
(21C fRoCot
For a well-insulated object, k = 1, which is the value attained as Ro ~ 00. The
value k = 112, for example, implies that leakage resistance reduces the
induced voltage to one-half that of a well-insulated object.
A very slow approach to an AC-charged source provides conditions in
which a sequence of spark discharges can occur. Both the average number
of discharges and the average voltage at which they occur will increase as
the approach velocity is reduced and as the peak sinusoidal voltage is
increased. Figure 9.5 illustrates typical discharge patterns associated with
very slow approaches using a device that could move an electrode toward a
stationary fingertip with a fixed, controllable velocity. In each example, the
charging voltage is 1,900Vrms, and the capacitance is 100pF. Each discon-
tinuity in the waveform represents a single spark discharge. Notice that the
discharges tend to occur at lower voltages in successive cycles of the sinusoi-
dal waveform, but that nearly all of the discharges extinguish at a minimum
voltage of about 600 V associated with the spark plateau described in Sect.
2.5. The number of discharges is greatest at the slowest approach. In each
(a) Velocity = 2 mmls
,~
II
2.0 ,ii 2.0
:>
!r' :~
I'
= :>
, I.
/~
=
' I
i II I . ,"-,
!l, 0 0
e i !l,
"0
>
, I
Il
. I
i,~
j\ i \1
.i
Jl 0,15 e
"0
>
0.15
-2,0 I'
I,}
1 \/ -2.0
'J
(b) Velocity = 8 mmls (e) Veloeijy = 16 mmls
FIGURE 9.5. Examples of stimulus patterns for various electrode approach veloci-
ties. Energizing voltage is 1.9kV rms. Horizontal axis is elapsed time; Co = lOOpF.
(a) Velocity = 2mm/s. (b) Velocity = 8mm/s. (c) Velocity = 16mm/s. (From Reilly
and Larkin, 1987.)
trace, discharges begin to occur singly on the individual half-cycles of the

60-Hz waveform. Then, as electrode distance is shortened, discharges occur
in clusters of increasing number and decreasing peak amplitude.
To gain some perspective about the factors that influence multiple
discharges, we recorded a large number of wave-forms while subjects
tapped electrodes of different sizes and energized at different voltages. The
number of discharges occurring at finger approach and withdrawal were
counted, and each discharge was measured. These counts and measure-
ments were also made with an electrode moving at a controlled low velocity
toward the skin. The results of these measurements were compared with a
computational model, in which a grounded electrode was assumed to ap-
proach an energized electrode in random phase with respect to the 60-Hz
charging voltage (Reilly et aI., 1983).
Table 9.3 illustrates averages of withdrawal and approach data for two
subjects, each experiencing 1,200 tapped electrode trials. The important
findings for a tapped electrode can be summarized as follows: (1) Below
900V, multiple discharges are rare. Typically, only one discharge is ob-
served, in the approach stage. (2) As the AC voltage is raised above 900V,
the number of discharges increases. The increase is greater in finger
withdrawal that in approach. (3) There are generally more withdrawal
TABLE 9.3. Statistical summary of discharge number for

AC field stimuli.
Mean number of discharges per tap
Voltage Ball electrode 1-mm electrode
(Vrms) Approach Withdraw Approach Withdraw
600 0.95 0.23 0.95 0.69
900 0.97 0.37 0.94 0.93
1,200 1.05 1.04 1.01 1.38
1,500 1.14 1.39 1.12 1.54
1,800 1.23 1.75 1.15 1.90
2,400 1.57 2.71 1.36 2.98
Note: Electrode tapped with the fingertip, using light tap force.
Source: Reilly and Larkin (1987).
discharges for a small electrode than for a large one. There is little depen-
dence on electrode size in the approach stage. (4) When multiple discharges
occur, they cluster within individual quarter-cycles of the sinusoidal
waveform.
For very slow approach velocities (less than 20 mm!s) , and adequate
voltage (>SOOV peak), a series of spark discharges may be obtained on
individual half-cycles of the alternating induced voltage function. Table 9.4
lists statistical data on the number of discharges for voltages from 1.09 to
TABLE 9.4. Stimulus characteristics for very slow

approaches to electrode.
Voltage Velocity
(kVrms) (mm/s) nd n,
1.09 2 20.8 7.6

4 11.7 3.9
8 4.4 2.0
16 1.7 1.3
1.31 2 27.2 8.8
4 18.0 5.2
8 10.1 3.1
16 4.0 1.9
1.57 2 49.5 16.7
4 24.8 7.6
8 16.4 4.7
16 6.2 2.2
1.89 2 60.6 18.2
4 33.8 9.8
8 20.9 5.5
16 8.3 3.4
Note: nd = average number of discharges during stimulus

pattern; n, = average number of waveform half-cycles during
which discharges were present.
Source: Reilly and Larkin (1987).
Direct Perception of ELF Electric Fields 353
1.89kVrms. In Chapter 7 we show how the sensory system integrates these

repeated stimuli.
9.3 Direct Perception of ELF Electric Fields

Mechanisms for Human Detection
Although power frequency E-fields induce internal body currents, such
currents can he easily dismissed as having any ability to stimulate excitable
tissue. This statement can be easily demonstrated by considering the E-field
necessary to excite peripheral nerves. Consider, for example, a person of
1.8m height standing within a 60-Hz field, and grounded through both feet.
H we consider the combined cross-section of the two lower legs (less bone)
as 35cm2 , and the minimum current density threshold as 0.lmA/cm2 (see
Sect. 7.6), then we calculate from Eq. (9.9) that an undisturbed field
strength of 200kV/m would be required to stimulate peripheral nerves in
the ankles. Considering that the undisturbed field is enhanced at body
surfaces (18 times, for example, at the head, as noted in Table 9.2), the body
would be in a state of severe corona at the stated field level.
Nevertheless, low-frequency E-fields can be perceived directly by hu-
mans and animals through other means. The mechanism for human detec-
tion is thought to be primarily the result of the vibration of body hair or
clothing. Hair has been observed to vibrate at both 60- and 120-Hz rates in
a 60-Hz E-field; these rates suggest two different mechanisms of hair vibra-
tion (Deno and Zafanella, 1982; Cabanes and Gary, 1981). When the hair is
hydrated, it appears to be sufficiently conductive that charges induced on
the surface of the skin are free to move along the hair shaft. This phenom-
enon would result in a mutual repulsion between individual hair follicles,
with a maximum force occurring on each one-half cycle of the electric field.
The result would be a 120-Hz vibration rate in a 60-Hz field. Dry hair, in
contrast, is a very poor conductor, on which environmental ions may be
held relatively immobile. The force on dry hair would consequently re-
spond directly to the E-field, resulting in a 60-Hz vibration within a 60-Hz
field. Low temperature and humidity decrease human sensitivity to ELF
field detection, whereas moisture caused by humidity or perspiration in-
creases sensitivity. At low stimulation levels, subjects typically perceive a
60-Hz E-field as a gentle breeze on exposed body surfaces (Reilly, 1978;
Kato et aI., 1989). A tingling sensation between body and clothes is also
observed. At higher levels, subjects describe a distributed tingling or crawl-
ing sensation on the skin.
Tests on human perception of 60-Hz E-fields were carried out at the high-
voltage research facility known as "Project DHV." During two months of
testing, approximately 122 men and 8 women walked within an electric field
produced by an overhead transmission line, stopping at selected points
corresponding to a gradually increasing electric field. At each stopping

point, participants considered the sensation On their raised hands, head,
head hair, or tingling between body and clothes (hands at sides). The
participants recorded one of three responses: no feeling, perception, or
annoyance. Figure 9.6 (from Reilly, 1978b) summarizes the responses of the
men. A few participants reported perception at remarkably low field levels
(:52kV/m). It would require more carefully controlled tests to be sure that
participants actually respond to low field levels rather than confuse other
stimuli such as a breeze. For the group of 8 women, the median threshold of
stimulation On the raised hands was 17.5kV/m, versus 6.7kV/m for a group
of 60 men tested On the same day. The higher threshold for women is
thought to be because their body hair is typically shorter than men's. The
fact that E-field detection thresholds are indistinguishable in male and
female mice (Stern and Laties, 1985) tends to make other gender-related
explanations less plausible.
Other tests conducted with 50-Hz fields at an indoor test facility
(Cabanes and Gary, 1981) have found higher thresholds for field perception
than shown in Fig. 9.6. In tests with 75 subjects (65 male, 10 female), median
95
90 • • Raised hands
• • Tingling
80 0----0 Head hair
70
g>
60
=g
8. 50
...
(/)
Q)
Q) 40
a.
0
Q) 30
Co
15 20
1:
e
Q)
Q) 10
Co.
0.5
1 2 3 4 5 6 7 8 910 15 20 30
Electric field intensity at 1-m height (kV/m)
FIGURE 9.6. Response thresholds of men to the direct effects of 60-Hz electric field.
Temperature = 50-65 F, both dry and wet conditions included. (From Reilly, 1978.)
0
-- -----_.
------ --
15 -----~---------
" ...
/,I'
Tingling (25th percentile)
,./,
,. Head-hair
Raised hands
o 50 100 150 200

Frequency (Hz)
FIGURE 9.7. Median values of threshold of perception of hair sensations of people as

a function of the frequency of the electric field. For "tingling" the 25th percentile
also is given. (From Deno and Zafanella, 1982. Copyright 1987. Electric Power
Research Institute. EL-2500. Transmission Line Reference Book: 345 k V and
Above, 2nd ed. Reprinted with permission.)
thresholds were found to be 2SkV/m with the hands at the sides, and
12.SkV/m with one raised hand. The fact that these thresholds are higher
than those for outdoor tests (Fig. 9.6) might be explained by a less humid
climate in the indoor facility, although relevant climatological data were not
published for the indoor tests.
Figure 9.7 illustrates the connection between the electric field frequency
and perception levels (from Deno and Zaffanella, 1982). Perception thresh-
olds were least at the lowest tested frequency (30Hz), with a gradual loss
of sensitivity as the frequency was increased. For the data represented in
Fig. 9.7, the rate of hair vibration was not reported; as noted earlier, it could
be either at the frequency of the applied field, or twice the frequency,
depending on the degree of hair hydration.
Cutaneous vibration sensitivity has a frequency response counter to that
indicated for E-field sensitivity. Hair receptors are examples of rapidly
adapting mechanoreceptors that are responsible for detection of vibration
on hairy skin. These receptors have a V-shaped threshold curve when

plotted against the frequency of excitation, as shown in Fig. 3.18. A com-
plete model for electric field detection may have to include the mechanical
compliance and electric impedance properties of the hair follicles.
Human perception is greatly enhanced when a strong DC field is com-
bined with an AC field, as shown in tests by Clairmont and colleagues
(1989). Test subjects recorded reactions within electric field environments
produced by conductors energized at various AC or DC potentials. Indi-
viduals recorded sensation levels at measurement locations where the AC
field, DC field, and ion current densities were monitored. Sensation levels
were reported on a six-point scale defined as follows: (0) not perceptible; (1)
just perceptible; (2) definitely perceptible, but not annoying; (3) slightly
annoying; (4) very annoying; (5) intolerable.
Figure 9.8, illustrating head hair perception, shows that alternating and
static fields have a synergistic effect-the combined fields can produce a
much stronger reaction than either field taken separately. Such synergism
might be explained by a mechanism where ions are produced by the DC
conductors and collected on body surfaces; the alternating component of
the fields would provide an oscillatory force for hair vibration. Although
Ql
> 3
.!l!
c
0
'"@
en
c
(J)
(f) 2
-20 o 20 40
DC Electric field (kV/m)
FIGURE 9.8. Reaction levels for exposure to mixed DC and 60-Hz AC electric fields;
stimulation of head hair while standing in field. Indicated field levels apply to
undisturbed field. Horizontal axis applies to DC field strength; curve parameters
apply to AC field strength. (From Clairmont et ai., ©1989, IEEE.)
such an explanation appears plausible, note that even a pure static field is
perceptible. Detection of a static field may also be the result of hair vibra-
tion, which has been reportedly observed beneath a DC high-voltage trans-
mission line. Such hair movement is not clearly understood but may be
associated with corona on the tips of the hair shafts. Laboratory studies by
Kato and colleagues (1986) failed to detect hair movement within a static
field. Differences between laboratory and environmental observations
might be associated with the density of ambient air ions, and relative
humidity.
In other laboratory experiments, Kato and colleagues (1989) exposed the
hairy arm and the palm of the human hand to SO-Hz electric fields. The
median perception threshold was SOkV/m on days that the relative humid-
ity ranged from 6S% to 78%. On dry days with humidity ranging from 22%
to 33%, nearly all subjects failed to perceive fields up to l1SkV/m, although
one exceptional subject detected a field at l1SkV/m. In all cases, perception
was associated with hair movement. When the hairless hand was exposed,
the field could not be detected up to the maximum level used in the
experiments (l1SkV/m).
The field levels cited in Kato's experiments may seem large in com-
parison with those indicated in Figs. 9.6 to 9.8, but may be accounted for by
differences in the definition of exposure levels. In Kato's laboratory experi-
ments, the reported fields are those on the surface of the skin; in the
environmental measurements (Figs. 9.6 to 9.8), the reported levels apply to
the undisturbed field (i.e., the rms value of the field in the absence of a
person). The difference between these two definitions is significant, as
suggested by Table 9.2.
Various studies have examined detection of ELF fields by laboratory
mammals (Stern et aI., 1983; Stern and Laties, 1985; Sagan et aI., 1987;
Weigel et aI., 1987). Weigel's experiments with cats provide significant
insight into mechanisms of transduction in E-field perception. In these
experiments, both normal and hairless cats' paws were exposed to strong
60-Hz E-fields. Thresholds as low as 160kV/m were noted, where the field
value applies to a measurement directly on the surface of the exposed body
part. That threshold value is similar to the perception threshold noted in
Fig. 9.S for human detection, if one accounts for the field enhancement on
local body parts.
Weigel and colleagues isolated receptor types and their receptive fields in
the cat's hind limb by monitoring action potential activity from the spinal
dorsal roots that supply those regions. They concluded that the most sensi-
tive mode of transduction was via rapidly adapting receptors. It was thought
that hair vibration could activate hair follicle receptors directly, or could
vibrate the skin and activate other types of nearby receptors, such as
pacinian corpuscles. In some cases, the later mode appeared to be the more
sensitive mechanism for transduction. But even in the hairless paw, electric
field stimulation was still noted, although at somewhat increased threshold
levels. Stimulation in the hairless paw was thought to be from vibrational

displacement of the skin due to dielectric forces across the corneum, similar
to the dielectric mechanism discussed in Sect. 7.2.
One should exercise great caution in comparing electric field detection
thresholds of humans and laboratory mammals because of differences in
body hair, as well as body shape, size, and orientation. Comparisons with
nonmammalian experimental species require even more caution, because
of the possibility of differences in detection mechanisms. Certain highly
sensitive fish, for example, are able to detect E-fields within the marine
environment as low as SnV/cm (Kalmijn, 1990). The existence of a unique
receptor and signal-processing mechanisms are said to be responsible for
this remarkable capability.
9.4 Human Reactions to AC Electric

Field-Induced Shock2
Stimuli produced through alternating E-field induction are more complex
than the single capacitor discharges discussed in Sect. 7.3. We have seen in
Sect. 9.2 that AC stimuli may contain one or more capacitor discharges
when physical contact with the induction object is made and broken, as well
as sinusoidal current during the period of maintained physical contact.
The individual discharges of the AC field-induced stimulus are identical
to the capacitor discharges discussed in Chapter 7. Thus, the factors that
affect sensitivity to individual discharges are also important with AC
stimuli. With AC field-induced stimuli, however, additional factors are
important, including multiple discharges and sinusoidal contact current.
The strength of the individual discharge, relative to the contact current, will
depend on both the capacitance and leakage resistance of the AC field-
charged object, as suggested by Eq. (9.12). It is therefore important to
evaluate AC field stimulus sensitivity throughout a range of capacitance
and leakage-resistance values.
Perception Thresholds
The sinusoidal contact component of the AC stimulus has a peak amplitude
that is considerably below the peak of the capacitive discharge current.
Nevertheless, the contact component can be perceptually potent-experi-
mentally, we find that 60-Hz sinusoidal currents can be detected with large-
area fingertip contact at an average current of about 0.2S rnA (Reilly et aI.,
1983). This value is consistent with 60-Hz touched-contact thresholds noted
in other studies (see Chapter 7).
2Parts of this section have been adapted from Reilly and Larkin (1987).
Human Reactions to AC Electric Field-Induced Shock 359
To further compare sensitivity to stimuli induced by alternating E-fields

with single-discharge stimuli, we tested the two kinds of stimuli randomly
interleaved during the same testing session. Capacitance was also varied
randomly, among values of 100, 800, and 6,400pF. At each capacitance
value, we determined the individual's threshold ratio, defined as the voltage
threshold for a single capacitor discharge divided by threshold rms voltage
for the AC field stimulus. For the six subjects tested, there were wide
differences (about 2:1) in absolute levels of sensitivity. Nevertheless, the
threshold ratios were highly similar at each capacitance. The implication of
this result is that, by knowing an individual's sensitivity to one type of
waveform, it is possible to determine his or her sensitivity to another
waveform using a simple multiplicative constant. The constant will, of
course, depend on the details of the two waveforms being compared.
Figure 9.9 plots the perception threshold voltage for contact with an
object charged by a 60-Hz E-field; thresholds decrease with increasing
4. a r----r----r---r---r--r---.-------r-....,....---.----.
1.0
Vi
E
>
.>.!
III
0'>
co
~
0
>
I~; I
"0
0
~ 0.1 k = w Co
'"~
~
I-
10 3 104
Discharge capacitance (pF)
FIGURE 9.9. Perception threshold for 60-Hz AC field-induced stimuli. Tapped elec-
trode. Curves for k = 1/2 and 1/4 indicate effect of leakage resistance. (From Reilly
and Larkin, 1987.)
capacitance, and with decreasing leakage resistance. [See Eq. (9.14) for a
definition of the leakage-resistance parameter k.] The resistance effect
stems from the greater sensory contribution of the sinusoidal contact cur-
rent component as leakage resistance is reduced. The fact that thresholds
are lower with reduced leakage resistance does not mean that reduced
resistance makes the shock exposure worse. If the induction E-field is held
constant, decreased leakage resistance reduces the induced voltage even
more than it does the threshold. For example, comparing k = 1/4 with k = 1,
the induced voltage is reduced by the multiple 0.25, but the perception
threshold at C = 3,200pF is reduced by the multiple 0.52 (see Fig. 9.9).
Superthreshold Effects
Suprathreshold reactions were tested for alternating E-field induction
stimuli in a manner similar to that described in Chapter 7 for single capaci-
tive discharges. In the magnitude-estimation procedure, perceived magni-
tude for 60-Hz E-field stimuli on the fingertip was found to grow at the 2.7
power of induced stimulus voltage at 6,400pF, and at the 2.2 and 2.8 powers
at 800 and 200pF, respectively. These exponents are not very different from
those observed for single-discharge stimuli (Chapter 7).
Table 9.5 lists, for six subjects, the mean affective response category
thresholds as multipliers of individual perception thresholds. The multiples
are for 60-Hz field-induced stimulation of the fingertip in active electrode
touching. Compared with the corresponding multiples for single discharges
(Table 7.2), 60-Hz stimuli reach the "unpleasant" category at approxi-
mately the same multiple; however, "pain" and "tolerance" categories are
reached with generally lower multiples.
Multiple Discharges
When a person grazes a metallic object that is energized by an alternating
E-field, there is a possibility for multiple discrete discharges. As demon-
TABLE 9.5. Multiples for superthreshold categories with

60-Hz E-field-induced stimuli.
Response Capacitance
category 200pF 800pF 6,400pF
Unpleasant 2.1 2.2 1.8
Pain 2.9 2.9 2.4
Tolerance 4.3 6.2 3.5
Note: Tabulated values indicate multiples relative to percep-

tion thresholds for given response categories. Data apply to
fingertip stimulation.
Source: Reilly and Larkin, 1987.
Human Reactions to AC Electric Field-Induced Shock 361
strated in Sect. 9.2, even a seemingly slow approach produces stimulus

patterns not unlike a deliberate direct touch with the fingertip. It is very
difficult to generate more than two to four discharges with a deliberate
grazing motion of the limb. However, a continuous stream of discharges can
be easily produced by pressing a shirt sleeve against an energized electrode.
In our tests, continuous sparking through a 0.1S-mm-thick fabric was easily
produced at 1.3 kV and above. Multiple discharges occur because the fabric
prevents electrical contact with the skin, which would preclude further
capacitive discharges. The duration of such discharges depends largely on
the reaction delay for the person to pull the arm away. A variety of studies
on human reaction time show that about 0.25 s is needed to move the hand
in response to a stimulus (Woodworth and Schlosberg, 1954). This period
corresponds to 30 half-cycles of a 60-Hz stimulus.
To produce multiple discharges reliably in the laboratory, a device was
used to move an electrode toward the skin with controlled motion. Figure
9.5 gives examples of the discharge patterns. Sensory magnitude estimates
for repeated discharges were obtained in a manner similar to that described
for single-discharge stimuli (Chapter 7). On each test trial, electrode ap-
proach velocity was randomly set at either 2,4, S, or 16mmls.
Figure 7.11 illustrates perceived sensation magnitude (geometric means
for three subjects), plotted against n c, the number of 60-Hz half-cycles
during which capacitive discharges were produced. The magnitude esti-
mates were related to the effective categories "neutral" and "pain" in a
separate procedure. Figure 7.11 suggests that a perceptible but neutral
stimulus can become highly unpleasant if presented as a repetitive train. A
rationale for representing nc as the independent variable is that there is a
tradeoff between the amplitude and number of pulses in each half-cycle
packet (Fig. 9.5), and that it approximately characterizes the overall stimu-
lus. The values of nc are the means for each velocity that was used in the
procedure. The straight lines in this format indicate power functions with an
exponent of 0.S5, indicating nearly a one-to-one correspondence between
nc and the mean perceived magnitude.
The dynamic range from the threshold of perception to pain is rather
narrow (Table 9.5). The limited dynamic range is especially striking when
one considers that a stimulus of low sensory potency can become quite
unpleasant when presented as a repetitive train. Consequently, conditions
giving rise to perceptible alternating field-induced shocks may, under some
circumstances, also present conditions for creating unpleasant effects.
Extrapolation to Other Frequencies

If field strength is held constant (resulting in constant Yo), the induced
current will be proportional to frequency [refer to Eq. (9.13)]. At the 50-Hz
power frequency common in Europe, the contact current would be reduced
by a factor of 50/60 (S.3 %) relative to that at 60 Hz. This would result in an
upward shift of 50-Hz thresholds, but the shift should not exceed 8.3% of
the values plotted in Fig. 9.9.
If the induction frequency is increased beyond 60 Hz, the sinusoidal
contact current takes on increasingly greater sensory importance in com-
parison with the capacitive discharge component, and, for sufficiently high
frequencies, will become the dominant sensory factor. Above 1 kHz, per-
ception thresholds for sinusoidal currents rise approximately in proportion
to frequency (Sects. 4.6 and 7.5), reaching a plateau at about 100kHz where
heating effects predominate (Fig. 7.12). Since the current induced by the
alternating electric field also rises in proportion to frequency, the E-field
level for perception of induced contact current should be nearly constant
between about 1 and 100kHz.
The opportunity for multiple discharges can increase with the induction
frequency. When individual capacitive discharges can be perceived, mul-
tiple discharges can greatly increase sensory potency, as noted in Fig. 7.11.
Sensory effects with multiple discharges may be intensified at frequencies
somewhat above 60 Hz because, sensory magnitude is encoded in part
through the action potential repetition frequency (see Chapter 3), but this
effect would be expected to saturate at an AP rate on the order of 200/s.
Skin Erosion from Low-Energy Discharges

The energy in a spark discharge to dry skin can be dissipated in a very small
volume because of the small effective diameter of the spark contact and the
fact that the resistivity of the dry corneum is high. The volume of dissipation
is especially small where the corneum is thin. Because of the high concen-
tration of dissipated energy, relatively small discharge energies are capable
of causing skin erosion. The effects of single or low frequency discharges to
the corneum are discussed in Sect. 2.6, where it is shown that discharge
energy typical of carpet sparks is capable of minor skin erosion when
repeatedly presented to the same spot on the dry skin of the leg. Radio
frequency discharges are discussed in Sect. 11.4, where it is shown that
relatively low voltage contacts (=200 V peak) can produce minor erosion of
corneal tissue in a small area. These contacts may be associated with a
stinging sensation. Such sensations may be due to the stimulation of thermal
receptors due to a localized heat rise. Another possible explanation
involves nonlinear impedance characteristics of the skin, which produces
low frequency modulation currents capable of exciting sensory neurons
electrically (see Sect. 4.6).
9.5 Time-Varying Magnetic Field Induction

Magnetic fields, produced by all current-carrying devices, induce electric
currents in conductive material, including the human body. Under most
conditions of environmental exposure, the induced currents are far too
Time-Varying Magnetic Field Induction 363
small to stimulate excitable tissue. Nevertheless, fields capable of stimula-

tion may be associated with certain industrial or research facilities, or with
medical diagnostic devices.
In this chapter, magnetic fields will be characterized in terms of magnetic
fluX density, B, having units of tesla (T). That quantity is related to magnetic
field strength, H, by B = pH, where fl is the permeability constant of the
medium, and H is the field strength in units of amperes per meter (Aim).
The permeability of air (flo) in this system of units is 4n X 10-7. The
permeability of biological tissue differs little from that of air (see Chapter
2). Consequently, the value of B within the biological medium may ordi-
narily be taken as that measured in air in the absence of the biological
specimen. The system of units used above is the International System (SI).
Flux density, B, is sometimes reported in the cgs system in units of gauss. To
convert between the two units, note that 1 gauss = 10-4 tesla.
To calculate the induction in a volume conductor, such as the human
body, a simple approach treats the volume as if it were made up of concen-
tric rings normal to the direction of the field, as illustrated in Fig. 9.1Oa. In
this simple picture, the field is assumed to be oriented along the long axis of
the body. The magnetic field induces an E-field within the body, having
roughly circular paths, as indicated in the figure. The induced E-field, in
turn, produces circulating eddy currents, which follow the path of the elec-
tric field in a medium of homogeneous conductivity.
In accordance with Faraday's law, the internally induced electric field E
is related to the time rate of change of flux density B by
(9.15)
The first integral is taken over a closed path, and ds is the element of area
normal to the direction of B. Each bold-face term is a vector quantity;
elsewhere in this chapter the standard fonts are used to represent scalar
quantities. If B is uniform over the region inside a closed path of radius r,
the induced electric field strength calculated from Eq. (9.15) is
E = -!... dB (9.16)
2 dt
where the direction of the induced E-field is along the circumference of
the circle. In some applications the magnetic field varies sinusoidally as
B = Bosin2njt, and Eq. (9.16) becomes
E = (rnJBo)cos2njt (9.17)
where the term in parentheses is the peak induced E-field during the sinu-
soidal cycle, and Bo is the peak magnetic field. Frequently, calculations for
sinusoidal magnetic fields express only the magnitude term in Eq. (9.17),
leaving off the cosine term. For induction in a concentric ring model, such
as that shown in Fig. 9.1Oa, Eq. (9.16) suggests that the outermost rings
would have the greatest E-field strengths. According to this simple model,
Magnetic field
0000
Magnetic field
(a) Longitudinal field (b) Frontal field
FIGURE 9.10. Distribution of internally induced electric fields from whole-body

exposure to time-varying magnetic field. Magnetic field direction is parallel to
long axis of body in part (a), and perpendicular to front of body in (b). Direction
of internal E-field and induced current reverses every half-cycle of alternating
magnetic field.
the maximum E-field for whole-body exposure would be computed from

Eq. (9.16) with r being the maximum circle radius that can be drawn on the
body in a plane perpendicular to B.
When the magnetic field is perpendicular to the long axis of the body, the
pattern of induced fields and eddy currents is somewhat more complex, as
suggested in Fig. 9.1Ob. Treating the human body as a prolate spheroid as
in Fig. 9.11, a long-wavelength solution was published by Durney and
colleagues (1975), and expressed in applied form by Spiegel (1976). The

long-wavelength assumption is that the field wavelength is much greater
than the dimensions of the biological subject. The complete solution for the
induced E-field because of electromagnetic exposure involves one compo-
nent as a result of the incident electric field, and another as a result of the
incident magnetic field. The internally induced electric field for three orien-
tations of a sinusoidal magnetic field is given by
-j21CfBx(a Zya z - bZza y)
E1 = (9.18)
(a z+ b Z)
j21CfBr(a zxa z - bZza x )

(9.19)
(a Z + bZ)
(9.20)
where E 1, E z, and E3 are the magnitudes of the sinusoidal electric fields that
are induced by magnetic field exposure along the x, y, and z axes, respec-
tively; B x, By, and B z are orthogonal magnetic field magnitudes; a and bare
semimajor and semiminor axes of the spheroid; ax, ay , and a z are unit vectors
along the x, y, and z axes; f is the frequency of the field; and j is the phasor
Bz
z
JJ.z
a
b
~----~y ~-----I-=-- Y
B~
x x E =37.7 VIm
(a) field perpendicular to long axis (b) Field parallel to long axis
FIGURE 9.11. Prolate spheroid model of uniform magnetic field induction. Bold
arrows indicate direction of induced electric field at outer boundaries axis a = 0.4 m
and semiminor axis b = 0.2 m.
operator (indicating 90° phase shift with respect to the sinusoidal phase of
the B field). Equations (9.18) to (9.20) define spatial vector components
of the induced electric field in terms of the magnitude of the sinusoidal
variation [the term in parentheses in Eq. (9.17)]. The vector magnitude
in Eq. (9.20) is identical to the amplitude term of Eq. (9.17) because r =
(~ + lil2 along a circular path.
A general expression for pulsed fields of arbitrary waveshape can be
derived from Eqs. (9.18) to (9.20) using arguments based on Fourier synthe-
sis. Designating An as the magnitude of the nth Fourier component of flux
density, its derivative is j2n!"An- It follows that the electric field induced in
an elliptical cross-sectional area is
E = - dBw [a ua
2
v - b2 va 1
u
(9.21)
dt a2 + b2
where au and a. are unit vectors along the minor and major axes respec-
tively, (u, v) is the location within the area, and Bw is the magnetic flux
density in a direction perpendicular to the elliptical cross section. For a
circular area, a = b = r, and Eq. (9.21) is equivalent to Eq. (9.16) on the
perimeter of the circle.
Consider, for example, a spheroidal representation of the torso of a large
person with a = OAm and b = O.2m, and a uniform 60-Hz magnetic field of
strength 1 T. If the field is oriented along the z axis, then Eq. (9.20) specifies
that the induced electric field at the equator of the ellipsoid is 37.7 Vim. For
orientation along the x axis, Eq. (9.18) specifies that the field is 30.2V/m at
the upper extreme of the spheroid, and 60.3 Vim at the lateral extreme. The
induced E-field components for this example are illustrated by bold arrows
in Fig. 9.11.
The internally induced electric fields result in circulating eddy currents.
The current density, J, is related to the induced E-field by the conductivity
of the medium, a, in accordance with J = Ea. For induction in living
subjects, calculation of the current distribution is complicated by the widely
differing conductivities of various body components (muscle, bone, blood
vessels, fat, etc.). The model used in this chapter treats the body as if it were
of homogeneous conductivity. The force responsible for stimulation is
taken to be the induced E-field within the biological medium (Chapter 4),
and the actual current density is of secondary interest.
In some applications, a knowledge of the current density is essential. One
such application concerns the distribution of internal heating caused by E-
M field exposure. Here, the distribution of current density must be known
in order to determine the patterns of power deposition within the body.
Gandhi and colleagues (1984) evaluated the distribution of power deposi-
tion resulting from magnetic field exposure by modeling the body as a series
of rectangular current loops, as illustrated in Fig. 9.12. Each loop incorpo-
rates the complex impedance appropriate to the tissue in a particular loca-
~ ~
Z2.S
G:
~ Z4,S
~ ZS.8
ZS.6
G:
FIGURE 9.12. Impedance mesh model for calculating magnetically induced currents
in nonhomogeneous medium. For simplicity of notation, only nine magnetic loops
are indicated. Induced electromotive force in each loop is calculated according to
Eq. (9.15). (Adapted from Gandhi et aI., 1984, © 1984 IEEE.)
tion of the body. The induced electric field in each loop can be calculated in
accordance with Eq. (9.15); the current value is thereby inferred using the
complex conductivity (including anisotropic properties) of the elements
in the loop. By simultaneously solving the loop equations for the entire
matrix, one can determine the overall current distribution.
Figure 9.13 illustrates the power deposition, calculated with the above
method, through a cross section containing the liver cut in a plane perpen-
dicular to the long axis of the body. The magnetic field is assumed to be
perpendicular to the cross section (as in Fig. 9.lOa), and to have a sinusoidal
variation at either 27.1 or 13.6MHz. The figure indicates that power depo-
sition generally rises when the measurement location moves from the
center of the body to its outer perimeter. This overall behavior is a con-
sequence of the radial dependency of the induced field, indicated in Eq.
(9.16). For a perfectly homogeneous medium, the power absorption would
vary as r (since power is proportional to the square of current). In Fig. 9.13,
deviations from an r dependency arise from variations in conductivity. The
25r-----~----~-----r----~----~------r---~
Cl 15
.::£
~
Q)
"§
c
.Q
.e
0(/)
10
.0
~
Distance from center of body (cm)
FIGURE 9.13. Average radial dependence of power deposition in cross-section of

human liver; cross-sectional plane perpendicular to long axis of body; distance
measured from center of body. Magnetic field perpendicular to cross section; mag-
netic flux density is 56,uT; frequency of field indicated on curves. (From Gandhi
et ai., 1984, © 1984 IEEE.)
sudden drop near the edge of the body (r = 12cm) results from a sudden
decrease in conductivity at the outer layers comprising fat and skin.
9.6 Principles of Excitation by Time-Varying

Magnetic Fields
This section emphasizes excitatory effects of magnetic fields in which the
exposure is over a large portion of the body, such as the entire torso or
whole head. Magnetic stimulation may also be provided over localized
Principles of Excitation by Time-Varying Magnetic Fields 369
areas of the body, using pulsed fields from small magnetic coils. Applica-
tions of local magnetic stimulation are treated in Sect. 9.9.
Pulsed Fields: Large Area Exposure

The subject of magnetic field excitation has been extensively studied with
application to patient exposure in magnetic resonance imaging (MRI). In
these imaging systems, three types of fields are applied at levels of intensity
much greater than what typically is found in most other environments. MRI
fields consist of an intense static field, a radio frequency field, and a
switched gradient field. The biological effects and mechanisms of inter-
action associated with these fields are very different from one another.
Biological mechanisms associated with MRI static field include magnetohy-
drodynamic effects, which are treated in Sect. 9.10. Radio frequency expo-
sure concerns in MRI have largely centered around thermal effects, which
are discussed in Sects. 9.5 and 11.5. Switched gradient field effects are
mainly associated with excitation of peripheral nerves, which is the main
focus of this Section. A treatment of various hazards in MRI is given by
Shellock and Kanal (1994). We shall also treat other applications of large
area exposure to magnetic fields, not necessarily confined to MRI.
Pulsed magnetic waveforms induce in-situ E-fields that are proportional
to the time rate of change of flux density, dB/dt. Figure 9.14 shows examples
of pulsed gradient field magnetic waveforms used in magnetic resonance
imaging (MRI) consisting of gated trapezoidal and sinusoidal functions and
associated dB/dt waveforms. Typical MRI gradient field pulse duration's
may be tenths of milliseconds to several milliseconds long, and may include
sequences of pulses separated by intervals from tenths of milliseconds to
several milliseconds long.
(a) (b) (c) (d)
0~d~~'~ ~
dB/dt waveforms
FIGURE 9.14. Example of flux density, and dB/dt waveforms.

To determine magnetic thresholds for nerve and heart excitation, we

must consider several factors, including the rheobase thresholds, the appro-
priate strength-duration (S-D) time constants, and the anatomical arrange-
ment of the excited tissue.
Minimum Excitation Thresholds

As seen in Chapter 4, the E-field aligned with the long axis of a nerve
or muscle fiber governs excitation. Since the induced E-field is proportional
to dBldt, the spatial and temporal properties of dBldt will determine the
excitation potential. With magnetic induction, dBldt is necessarily bipha-
sic and charge-balanced (i.e., the area under the positive and negative-
going phases are equal). However, the dBldt waveform of a pulsed
B-field can be treated as monophasic for the purposes of nerve excitation
if the biphasic reversal of the induced field is gradual or delayed, as seen
in Fig. 9.14 (a) and (d). Recall from Chapters 4 and 6 that nerve and heart
S-D relationships dictate that E-field thresholds are lowest for stimuli
that are long in comparison to the S-D time constant, and that thresholds
for monophasic pulses are typically lower than for biphasic waveforms
of equal duration. Consequently, minimum magnetic excitation thresholds
can be analyzed using an assumption of a monophasic square-wave dBldt
pulse.
As noted in Sect. 4.5, a nerve fiber may be excited at one of three loci: a
nerve terminus, a sharp bend, or a location where the spatial gradient of the
E-field is sufficient (Fig. 4.15). With magnetic exposure over a large portion
of the anatomy, the spatial derivative of the induced E-field is typically
small, and stimulation at a nerve terminus will determine the lowest thresh-
old. Based on a myelinated nerve fiber situated in a uniform field oriented
along the long axis of the fiber, E-field thresholds are calculated as 6.2, 12.3,
and 24.6 VIm for fiber diameters of 20, 10, and 5,um respectively (Table 4.2).
The threshold applicable to a 20-,um fiber will be used here for a conserva-
tive calculation.
Recall from Chapter 6 that the cardiac excitation threshold is a minimum
during the unexcited (diastolic) interval of the cardiac cycle; the fibrillation
threshold is a minimum during the vulnerable (partially refractory) period.
The relatively lower excitation threshold will be used here to evaluate
magnetic stimulation of the heart. For prolonged exposure, the cardiac
excitation threshold at the one- and 50-percentile ranks are estimated at 5
and 10V/m respectively with large-area electrical stimulation (Sect. 6.10).
Note that the presumed one-percentile cardiac excitation threshold is very
close to the median nerve threshold. To facilitate the calculations in this
section, a reference threshold of 6.2 V1m will be used for both the nerve and
the heart.
Strength-Duration Time Constants

The S-D time constants of cardiac and nerve tissue differ widely. As noted
in Chapters 4 and 7, a theoretical model for myelinated nerve with uniform
field excitation indicates 7:, = 0.12ms, a value consistent with in-vivo nerve
stimulation in animal experiments. As we shall see presently, that theoreti-
cal value is close to experimental time constants derived with large area
magnetic stimulation. In contrast, the experimental average of 7:, for cardiac
excitation with large-area exposure is about 3ms-a factor more that 10
greater than that for nerve stimulation. Large area exposure is assumed for
most cases of magnetic induction because of the nonfocal nature of the
induced electric field.
The S-D time constant determines the thresholds for stimuli that are
brief in comparison to 7:,. In the short pulse regime, minimum thresholds are
defined by the peak value of B rather than dBldt. As a result of S-D
principles defined in Sect. 4.2, it follows that Bo = Eo 7:., where Eo is the
minimum dBldt threshold for t « 7:., and Bo is the rheobase B threshold for
t» 7:,.
Anatomical Considerations
Body geometry is illustrated in Fig. 9.15 for a large man. Body dimensions
in applied situations may differ from this example, and calculations
would vary accordingly. Furthermore, the location of the heart within the
torso will vary as the body position is changed from prone to erect, and
also will move during the cardiac pumping cycle. Calculated E-fields
have been carried out by representing body cross sections as ellipses, as
indicated in Fig. 9.15. Numbered points indicate positions where the
E-field (shown in Vim) has been calculated according to Eq. (9.21) with
dBldt = 100 Tis.
Calculated E-fields generally increase as the location is moved toward the
perimeter of the body, and are especially great along the minor axis of the
ellipse. Points of maximum E-field on the perimeter of the body are points
14,9, and 3 in the three cross sections; the maximum E-field on the heart
occurs at points 10, 6, and 1.
An underlying assumption in the above is that the incident magnetic field
is constant in both magnitude and phase over the torso cross-section. A
somewhat more conservative calculation can be made by assuming the
entire body is uniformly exposed to the B field by using an equivalent body
ellipse equal to the height of the person. A calculation for sagittal exposure
was made, using semimajor and semiminor axes of 0.9 and 0.17m respec-
tively. In this case, the calculated E-field was somewhat elevated relative to
the case where only the torso was exposed. The increase was 11 % at point
(10), and 14% at point (14).
v v
W
-..J
tv
;.0
Internal organs
CIl
~
a· Heart 3'
b • liver s::
c· Stomach 6J
d . Diaphragm c.
e • Large intestine o
f • Small intestine ::s
g. Vertebra $.
h· Ribs ~
i . Pectoral muscle and pleura
~
(1l
(7) 5.4 2t
n'
~
::s
0-
h s:::
~
::s
(1l
~
·v
n'
~
(1l
5:
rJl
(1)87
(a) Sagittal cross section (b) Frontal cross section (C) Longitudinal cross section
(a = 0.4, b = 0.17) (a 0.4, b 0.2)
= = (a = 0.2, b = 0.17)
FIGURE 9.15. Example of position of the heart within three cross sections of the body; a and b
are semimajor and minor axes of equivalent cross-section ellipsoids. Numbered locations are identified
with calculated electric field intensity (Vim) for magnetic exposure within the ellipse of dBldt = lOOT/s.
Direction of field is perpendicular to the cross-section. View of internal organs shows conductive paths
through the heart
Another implicit assumption is that the internal conductivity of the body

is homogeneous. In reality, the body is composed of various organs having
diverse conductivities, as indicated in Fig. 9.15, and these variations will
distort the internal E-field with respect to a homogeneous structure. For
instance, in the sagittal cross section, the E-field will be enhanced between
the skin and the spinal column as induced eddy currents are crowded
around the low-conductance vertebrae and the curvature of the torso in
the lower lumbar region. Indeed, Schaefer and colleagues (1994) report
that sensory perception from magnetic stimulation is enhanced in regions
where boney projections are just below the surface of the body. It should
be clear from Fig. 9.15 that the heart is part of a large electrically con-
ductive circuit that includes the blood vessels, diaphragm, liver, and
intestines, and should not be treated as an isolated conducting organ
suspended in non-conductive air (lungs) as some have opined. With longi-
tudinal magnetic exposure (Fig. 9.15c), the heart is part of a circumferential
electrical circuit that includes the pericardium and muscle lining the
thorax.
Calculated Nerve and Cardiac Excitation Thresholds

In accordance with the preceding, a reference value of 6.2 V1m will be
assumed as the median rheobase threshold for nerve excitation, and also as
the one-percentile rheobase threshold for cardiac excitation. Table 9.6 lists
dBldt needed to induce 6.2 Vim at selected points on the body model of Fig.
9.15 according to Eq. (9.21). The table also lists Bo> the minimum threshold
TABLE 9.6. Magnetic field parameters necessary to induce threshold E-field at

selected body loci (calculated by elliptical cross-section model).
Minimum threshold
B-field Ellipse b,a Body dBldt B
orientation (cm, cm) locus (TIs) (mT)
Longitudinal 17,20 (1) heart 71.3 213.9
(3) nerve 62.6 7.51
Frontal (torso) 20,40 (6) heart 68.1 204.3
(9) nerve 38.8 4.66
Sagittal (torso) 17,40 (10) heart 48.8 146.4
(14) nerve 43.1 5.17
Sagittal (whole 17,90 (10) heart 44.0 132.0
body) (14) nerve 37.8 4.54
Calculations are for human geometry of Fig. 9.19. Minimum dBldt thresholds based on
E = 6.2 VIm for long-duration pulses. Minimum B thresholds determined by Bo = Eo re for
short-duration pulses, with re = 0.12ms (nerve), and 3ms (heart); Eo is the minimum dBldt
threshold.
for short pulses, using Bo = Eo 7:., where Eo is the rheobase dBldt threshold;
7:e= 0.12 and 3.0ms for nerve and heart respectively. In the case of sagittal
exposure, an additional conservative calculation is shown for exposure to
an elliptical representation of the entire body.
Figure 9.16 illustrates dBidt and B thresholds as strength-duration curves
for a magnetic field normal to the sagittal cross section of the human body;
the S-D time constant Te is shown as a parametric variable. The curves apply
,$'
, "...
--- ,
-------------------- - _... - ... ..., ,~
,
"
~
, ...
2
-------------------- ... -" ',"
-------- ----------- - - Nerve

- - - - Heart
10~-----~
4 ~_L~-L~~_~~~~~~~~~~~~~~~~~
.01 .1 10 100
Pulse duration (ms)
30
.01 .1 10 100
Pulse duration (ms)
FIGURE 9.16. Calculated excitation thresholds (E = 6.2 VIm) for sagittal exposure of
torso to pulsed magnetic field. Curve parameter ie applies to strength-duration time
constant of excitable tissue. Upper panel: peak flux density; lower panel: peak
dB/dt. (From Reilly, 1992.)
to the most sensitive loci on the heart or peripheral nerve, namely points
10 and 14 on Fig. 9.15. At long durations, dBldt thresholds (lower panel)
converge to a minimum value (rheobase) independent of Te. At short
durations, B thresholds (upper panel) converge to a minimum value that is
directly proportional to Te. Consequently, choosing a small value of Te
provides a conservatively low estimate of the B-threshold.
As a general rule, the more central position of the heart within the torso
in comparison with the more eccentric position of peripheral nerves
requires a greater magnetic field to achieve a given induced E-field. The
relative geometric advantage of the heart depends on the direction of the
incident field with respect to the body. Nerve and cardiac thresholds are
predicted to be most disparate when the incident field is perpendicular to
the frontal cross-section of the body, and least disparate when the field is
perpendicular to the sagittal cross-section. The lowest absolute thresholds
are predicted for peripheral nerves with frontal exposure, and for the heart
with sagittal exposure.
A further separation of cardiac and nerve excitation thresholds can be
realized by taking advantage of the relatively longer excitation time con-
stants of cardiac tissue. Cardiac and nerve tissue thresholds become more
disparate as the stimulus phase duration is made shorter than the S-D time
constant of the heart (about 3ms).
Sinusoidal Fields: Large Area Exposure

Excitation thresholds, when plotted versus sinusoidal frequency, describe a
strength-frequency (S-F) curve. Excitation thresholds for sinusoidal and
rectangular stimuli have been compared in Chapter 4 using a theoretical
model for myelinated nerve fibers. At low frequencies (=100Hz), sinusoi-
dal thresholds are minimum, and exceed the pulsed monophasic stimulus
thresholds by 17%, which would place the minimum sinusoidal E-field
threshold at 7.25 Vim for a 20-,um fiber. Thresholds rise at both lower and
higher frequencies according to the S-F relationships described by Eq. 4.35,
in which the parameters fa and Ie describe the low- and high-frequency
corners, respectively. At high frequencies, E-field thresholds eventually
rise in proportion to frequency. The low-frequency corner was found to be
fa = 10Hz. As for the high frequency corner, available data suggest a wide
spread of values. The myelinated nerve model indicates fe = 5,400Hz
for axonal stimulation (Chapter 4). Electrocutaneous sensory stimulation
indicates a range of values, having a median of about 500Hz (Chapter 7).
With cardiac excitation, the median experimental value of Ie is 120Hz
(Chapter 6).
The upper transition frequency Ie is significantly lower for electro-
cutaneous stimulation as compared with the myelinated nerve model.
That observation is consistent with the finding that experimental values
of Te are higher with electrocutaneous data. In developing criteria for mag-
netic stimulation, the relevance of electrocutaneous data is not clear-

possibly a model based on axonal stimulation may be preferable. This
speculation is suggested by the good agreement of model and experiment
when an axon is excited subcutaneously (Reilly et ai., 1985). Unfortunately,
available frequency sensitivity data, reviewed in Sect. 9.7, are inadequate to
provide clear guidance on the most appropriate value of t. for magnetic
stimulation.
Figure 9.17 illustrates S-F curves for magnetic excitation of nerve and
cardiac tissue for sagittal exposure, with the parameter t. varied. As with
Fig. 9.16, the curves apply to the most sensitive loci on the heart or periph-
ery of the body, namely points 10 and 14 on Fig. 9.15. At low frequencies,
dBldt thresholds converge to a minimum value independent of t. (lower
panel); at high frequencies, B thresholds converge to minimum value that is
directly proportional to Ie (upper panel).
Criteria for Nerve and Heart Excitation:

Large Area Exposure
The S-D curves of Fig. 9.18 have been used in to develop standards for
patient exposure to pulsed and sinusoidal fields in magnetic resonance
imaging (see Sect. 11.3). For pulsed stimuli, the horizontal axis is inter-
preted as the duration of a single monophasic phase of dBldt; for sinusoidal
stimuli, it is the duration of a half-cycle of the sinusoidal wave. The figure
indicates dBldt values required to induce 6.2V/m at points on the body
model (Fig. 9.15) as indicated in Table 9.6. Recall from Sect. 4.6 with
sinusoidal stimulation of multiple cycles, thresholds eventually converge to
that of a single monophasic pulse of duration equal to the half-cycle time of
the sinusoid. Consequently, one can conservatively use Fig. 9.18 as thresh-
old criteria for both pulsed and sinusoidal fields, regardless of the number
of cycles of sinusoidal variation. As discussed in Sect. 11.3, the criteria
associated with these figures formed the basis for MRl exposure standards
of the FDA in the United States, the NRPB in the United Kingdom, and the
lEC in continental Europe.
Standards pertain to exposure of patients without implanted metallic
devices. With such implants, an enhancement of the induced E-field can
occur in regions near the implant (Reilly and Diamant, 1997).
Body Size Scaling

One can apply the methods presented above to subjects of various body
sizes using an elliptical representation of the torso dimension of the subject,
and solving Eq. 9.21 for the value of dBldt needed to induce the threshold
E-field. One can simplify this procedure with an approximate method based
~~
---. ---.
---. ---.
fa =50 Hz
100
~~~ ------
~-~
~---~--___ 100
I='
.E-
~~- ----------
---. ---
~~-
---
200
~~-
m
..I<:
-----
ra
Q)
Q.
- - Nerve 500
10 - - - - Heart
10 102 103 104 105

Frequency (Hz)
,
,, ,, ,,,
2000
, ,
1000 - - Nerve
, ,, ,, ,, ,
,
- - - - Heart ,, ,, ,,
, , 100 200
e.
(j)
fa =?O Hz, , , ,
,
~ , ,, ,, ,,,
,, ,,, ,,
"'C
..I<:
,, , ,,
ra
,, ,,, ,,,
Q)
Q.
100 ,, , , ,
... "
'~---... :: ... " "
30
10 102 103 104 105
Frequency (Hz)
FIGURE 9.17. Calculated excitation thresholds for sagittal exposure of torso to sinu-
soidal magnetic field. Curve parameter !e applies to upper frequency constant of
excitable tissue; lower frequency constant!o = 10Hz. Upper panel: peak flux den-
sity; lower panel: peak dBldt. (From Reilly, 1992.)
on body weight. Note that the induced E-field is directly proportional to the
dimensions of the ellipse (i.e., if we halve the ellipse dimensions, the maxi-
mum induced E-field will be halved as well). Further, note that body weight
varies as the cube of the linear dimensions of the subject. Therefore, assum-
1000
Longitudinal
Exposure
100
30
1000
Frontal
Exposure
~
~
-c
100
30
1000
Sagittal
Exposure
100
30
.01 .1 10 100
FIGURE 9.18. Calculated thresholds for dBldt exposure of human torso, with
induced E-field of 6.2 Vim as a criterion; reference points identified in Fig. 9.15.
ing that subjects have a similar body shape, we can scale B-field thresholds
from one subject size to another by using a body weight scaling factor
K B-- (WO J
-
1/3
(9.22)
W
where Wo is a reference body weight for which the threshold is known, and
W is the weight of another subject. For the subject depicted in Fig. 9.15, one
may ascribe a body weight of about 90kg. If we apply the body weight
Experimental Investigations of Magnetic Excitation 379
scaling formula to a small child of body weight lOkg, we obtain KB = 2.08.

Consequently, thresholds appearing in Table 9.6 would be multiplied by
that factor for application to the small child, provided that the excitation
threshold of 6.2V/m applies equally to both the large and small subject.
One might question that assumption, since cutaneous electrical thresholds
are subject to a body size scaling law (Eq. 7.8) in which small individuals
have lower thresholds for contact currents than large individuals. The role
of body size in magnetic thresholds has not been systematically studied.
9.7 Experimental Investigations of Magnetic

Excitation: Large Area Exposure
When I first published theoretical models for magnetic excitation of periph-
eral nerve, most available experimental data were derived from small coil
systems in which exposure was limited to a small area of the anatomy
(McRobbie and Foster, 1984; Polson et aI., 1982a; Polson et aI., 1982b; Veno
et aI., 1984; Irwin, 1970). My review of these early studies (Reilly, 1989;
1992) suggested that if one properly accounts for factors associated with the
temporal and spatial distribution of the magnetic stimulus, the experimen-
tal findings are seen to be reasonably consistent with theoretical models.
Since publication of my early predictive models, numerous experiments
have been conducted using large coil systems; most of those studies were
motivated by the desire to verify acceptability limits for patient exposure in
MRI examinations. As will be demonstrated below, experimental findings
for both nerve and cardiac excitation are surprisingly close to the predictive
models-an outcome that I ascribe to equal measures of inspiration and
good luck, considering the many simplifying assumptions used in the
models.
Thresholds for Nerve Excitation

Table 9.7 summarizes experimental data on nerve excitation by large area
magnetic exposure. The first column roughly describes the excitation wave-
form. The phase duration (tp) is that for a single phase of the dBldt wave-
form-for sinusoidal waveforms, f = 1/(2tp). For experiments in which tp
was varied, the largest value used in the experiments is listed. The magnetic
field direction is that of the principal spatial component with the convention
that x, y, and z indicate magnetic field directions perpendicular to sagittal,
frontal (coronal), and longitudinal (axial) cross sections of the subject,
respectively. The threshold dB/dt applies to the listed value of tp; the first
listed value indicates the mean or median, ±indicates a standard deviation,
and ( ) indicates a minimum observed threshold.
The last two rows of data in Table 9.7 apply to excitation of the phrenic
nerve, leading to respiratory reactions. The magnetic thresholds for
w
00
TABLE 9.7. Nerve excitation thresholds for large area magnetic exposure of human and animal sUbjects. 0
phase mag. field

Waveform duration direction reaction threshold :0
of dBldt (liS) subject n (x,y,z) locus reaction dBldt (Tis) Reference I:n
Damped 1,190 Human 20 z Torso, Sensation 75 (45) Bourland et aI., 1990 3'
-
sine, 1 ~ extremities S
~
Pulse 200 Human 13 x Torso Sensation 99.7 ::':: 13.5 Schaefer et aI., 1994 o·
-
i:l
Y 86.7::':: 15.9 -<
z 101.8 ::':: 12.3 SO'
tI1
Pulse 368 Human 11 y Torso Sensation 63.7 ::':: 13.5 Schaefer et aI., 1995 (;
n
z 94.9::':: 22.4 ....
(;'
-
Sine,4.7ms 394 Human 10 z Abdomen Sensation 60::':: 6.0 Budinger et aI., 1991 ~
i:l
Sine, 46ms 238 Human 2 z Torso Sensation, 61.0 Cohen et aI., 1990 0-
twitch ~
~
(JCj
Sine, 32 ~ 400 Human y Head Sensation 71.0 Yamagata et aI., 1991 i:l
(1)
Pulse 1,000 Human 52 y Torso Sensation 14.4* Bourland et aI., 1997 -( ;'
Damped 572 Dog x Torso Twitch 92 Nyenhuis et aI., 1990 "I1
(P'
sine, 1 ~ Y 114 0:
z 179 '"
Pulse 560 Dog, 12 z Torso Twitch 370 ::':: 357 Bourland et aI., 1991a
17-32 kg
Pulse 530 Dog z Torso Inspiration 1,000 (450) Bourland et aI., 1991b
50% vol.
Pulse 450 Dog, 21kg 1 z Phrenic Inspiration 450 Mouchawar et aI., 1991
nerve
Notes: First dBldt value indicates mean or median; ::':: indicates standard deviation; ( ) indicates minimum; - indicates information not supplied.
Phase duration is maximum used in experiment. Direction: x = sagittal; y = frontal; z = longitudinal. * Revised metric (see text).
respiratory effects are, as expected, greater than those for peripheral nerve
excitation because of the more central anatomical location of the stimulated
organs.
To compare the experimental data with theoretical predictions, we must
account for several factors, including the waveshape and the size of the
subject or of the exposure area. Consider, for instance, the second row of
data from the experiments of Schaefer and colleagues (1994). If we consider
the S-D time constant 7:e = 120,us as suggested by the theoretical model,
then by applying the S-D law (Eq. 4.6), we expect that rheobase dBldt
would be a factor of 1.23 below the listed threshold values, namely 80.8,
70.3, and 82.5T/s for x, y, and z exposure directions respectively. If we
assume that 7:e = 150,us as indicated by experimental data (described
below), the multiple would be 1.36, and we would estimate rheobase thresh-
olds as 73.4, 63.9, and 75.0T/s for x, y, and z directions respectively. Com-
pare those values with the theoretical rheobase predictions in Table 9.6 of
48.8,38.8, and 62.6T/s for a large man and for a monophasic square-wave
dBldt pulse. Further adjustments for body size and waveform factors would
suggest even closer correspondence of theory and experiment.
As a second example, consider the row of data applying to Nyenhuis and
colleagues (1990). Using an estimated body weight of 20 kg for the canine
subject, the scaling factor to a 90-kg man (Eq. 9.22) would be 1.65. Applica-
tion of that scaling factor to the canine dBldt thresholds results in thresh-
olds of 55.7,69.1, and 108T/s for application to a human.
The experimental conditions applying to Table 9.7 do not exactly corre-
spond to the assumptions in the theoretical models. For instance, the theo-
retical model assumes a monophasic square wave dBldt pulse, and constant
value of the peak flux density and its phase over the exposure area. Experi-
mental deviations from those assumptions would tend to produce higher
experimental thresholds. Furthermore, the theoretical model assumes a
large person of homogeneous conductivity, whereas actual subjects vary in
size, and internal conductivities are not homogeneous. Experimental devia-
tions from those assumptions would produce lower or higher experimental
thresholds. Nevertheless, even a rough interpretation of the experimental
data shows that the predicted thresholds do not differ greatly from experi-
mental findings.
The most recent data from Bourland and colleagues (1997) indicate a
much lower dBldt rheobase than previous data from the same laboratory.
This results from a change in the exposure metric used by the investigators
(Nyenhuis et aI., 1997). In earlier studies, the investigators reported the
peak field, whereas in recent studies they reported a spatial average.
Of course, the relevant exposure metric is the E-field induced within the
subject. Under conditions where the induced E-field could be effectively
determined, the rheobase E-field for stimulation of the forearm was found
to be 5.9 Vim (Havel et aI., 1997)-a value that is quite close to the theoreti-
cal minimum value of 6.2 Vim.
Thresholds for Cardiac Excitation

The experimental data discussed here apply to acute effects with large-area,
short-term exposure of humans and animals. In addition to these effects,
there is evidence of more subtle cardiac reactions with much lower intensi-
ties and longer durations of exposure. A statistically significant slowing of
the human heart by a few percent has been reported with simultaneous
exposure to a 60-Hz environmental electric field of 9kV/m and a magnetic
field of 20.uT, but without effect at lower (6kV/m; lO.uT) or higher
(12kV/m; 30.uT) fields (Graham et aI., 1994). At present, these findings
cannot be explained on the basis of understood biophysical models.
Experimental data on magnetic excitation of the heart have been avail-
able since about 1991, with the introduction of experimental devices for
advanced MRI echo-planar devices. As might be expected, human data on
magnetic excitation of the heart does not exist; experiments to date have
been performed with dogs as summarized in Table 9.8. As with stimulation
of the heart by conducted current (Chapter 6), magnetic stimulation is
most effective when it is delivered during the T -wave of the cardiac cycle
(Yamaguchi et aI., 1994).
Thresholds for two of the experiments in Table 9.8 (Mouchawar, 1992;
Yamaguchi, 1992) are given in terms of the induced E-field on the heart,
rather than the incident magnetic flux density. The authors each calculated
those E-fields for their coil systems situated over an infinite, homogeneous
medium. The data of Mouchawar show an average excitation threshold of
124 Vim at a stimulus phase duration of 530.us. If we considered the stimu-
lus as a monophasic dBldt pulse, and use a value of ie = 3ms for cardiac
tissue, we would infer an average rheobase of 21.5 VIm, which is a factor of
1.7 greater than the rheobase threshold of 12.4 Vim used as the median
cardiac excitation threshold in the model described in Sect. 9.2. For reasons
that have been analyzed previously (Reilly, 1993), even the inferred
minimum value of 21.5V/m in Mouchawar's experiments, as well as
Yamaguchi's threshold of 30V/m are thought to exceed attainable
rheobase. One contributing factor is the biphasic nature of the waveforms
used by both experimenters. The phase durations in these experiments
(0.571 and 2.13ms) are below the expected S-D time constant of 3ms.
Under such conditions, a biphasic reversal, such as evident in the wave-
forms used by the experimenters, would increase thresholds relative to a
monophasic pulse, or a biphasic pulse with a delayed or more gradual phase
reversal.
Excitation thresholds for the heart should be substantially greater than
that for nerve as long as the pulse duration is sufficiently less than the S-D
time constant of the heart (3ms). Consequently, the avoidance of periph-
eral sensations in a patient should provide a conservative safety margin with
respect to cardiac excitation. The margin between nerve and heart excita-
tion thresholds was examined in canine subjects by Bourland and associates
TABLE 9.8. Cardiac excitation parameters for magnetic stimulation of animal subjects.
phase mag. field threshold threshold
Waveform duration direction dBldt E-field nerve ex.
of dBldt (#s) subject n (x,y,z) (Tis) (Vim) multiple Reference
Pulse 530 Dog, 12 z 4,242::': 678 14.1 ::': 6.7 Bourland et aI., 1991a
tTl
17-32 kg ~
(l)
Pulse 530 Dog, 10 z 3,900 Bourland et aI., 1991c ....
21-32kg S·
(l)
Dog 9 z 2,135 ::': 457

::s...,.
Pulse 530 Bourland et aI., 1992
e:..
Pulse 571 Dog 11 x 124 Mouchawar et aI., 1992 ......
::s
<:
Damped 540 Dog 12 z 2,155 ::': 286 Nyenhuis et aI., 1992 (l)
~
sine*, qq'
1~ ~
Dog, 2 x 30 Yamaguchi et aI., 1992
o·
Damped 2,130 ::s
sine 10, 15kg o
'"....,
r = 4.2ms
s:::
Damped 490 Dog,9kg x 159 Rosono et at., 1992
sine (l)
~...,.
r;'
Notes: First dBldt value indicates mean or median; ::': indicates std. dev.; - indicates info. not supplied.
Phase duration is maximum used in experiment.
~
O.
...,.
* Static field of 1.5 T added. :>:>
....
Direction: x = sagittal; y = frontal; z = longitudinal. o·
::s
~
00
~
(1991a), who reported a threshold ratio of cardiac to neuromuscular excita-

tion of 14.1 ± 6.7. We can compare that ratio with one calculated for human
subjects using as rheobase a threshold of 62.6 Tis for nerve, and 142.6 Tis for
the heart as suggested in Table 9.6 for longitudinal exposure. (As suggested
by Table 6.12, the listed 1 % cardiac threshold has been multiplied by 2 to
give the 50% cardiac threshold). From the S-D formula, with 'te = 0.12 and
3.0ms for nerve and heart respectively, we derive a theoretical heartlnerve
excitation ratio of 14.0, which is quite close to the experimental ratio.
Stimulus Duration and Frequency Effects

The strength duration time constant, 't., allows us to scale from the rheobase
threshold at long stimulus durations to the threshold at shorter durations
through the S-D law (Eq. 4.6). As seen in Chapter 4, 'te differs substantially
not only with respect to tissue type, but also with the spatial distribution of
stimulus field. It is therefore important to assess 'te under appropriate con-
ditions of stimulation.
Table 9.9 summarizes 'te data from experiments that used magnetic stimu-
lation of nerve and cardiac tissue in humans and animals. Some of these
data were obtained with much smaller coil systems than used in the experi-
ments listed in Tables 9.7 and 9.8. But even for these smaller coils, the
spatial derivative of the induced E-field is sufficiently small that the experi-
mental 'te values can be attributed to large coil systems. In Table 9.9, the
TABLE 9.9. Strength-duration time constants determined with magnetic stimulation.

Magnetic
Exposed field T,
Subject n area device Reaction (us) Reference
Human 10 Motor 8.5cm EMG-arm 152 :t 26 Barker et al., 1991
cortex dia. coil
Human 12 Wrist 8.5cm EMG-arm 150:t 55 Barker et aI., 1991
dia. coil
Human 4 Wrist 4.8cm Perception 146 :t 7.5 Mansfield & Harvey,
dia. coil 1993
Human 14 Forearm llcm Perception 395 :t 79 Havel et aI., 1997
dia. coil
Dog 12 Thorax z-coil Muscle 148 :t 49 Bourland et aI.,
twitch 1991a
Dog Thorax x-coil Twitch 291 Nyenhuis et aI., 1990
y-coil 213
z-coil 240
Dog 2 Heart 22-cm Ec. beat =3,000 Yamaguchi et aI.,
dia. coil 1992
Direction: x = sagittal; y = frontal; z = longitudinal.

r-----r-----r-----r-----r-----~----~----r130
16
15 120
14
110 ~
t=' 13 f = 1270 Hz t:..
§. "'C
"'C
12 100 "0
"0 .r:
.r: CJl
CJl ~
~ 11 90 £;
£; subject 1, Y gradient ....
III 10 :!2
III
80 "'C
9 subject 1 and 2, z gradient
8 70
7 60
0 10 20 30 40 50 60
number of oscillations
FIGURE 9.19. Variation of perception threshold with number of cycles of magnetic

stimulation at 1,270Hz. Z-gradient exposure was constant over longitudinal cross
section of body; y-gradient varied over cross-section. (From Budinger et ai., 1991.)
majority of fe values do not differ greatly from the value of 120.us derived
with the SENN model. However, the data of Nyenhuis and of Havel (both
with the same laboratory) indicate larger fe values for reasons that remain
unexplained. fe determined by Yamaguchi and colleagues for cardiac exci-
tation is close to the value used in the theoretical models.
Another aspect of duration sensitivity was studied by Budinger and col-
leagues (1991). In their experiments, 10 healthy men were positioned with
the pelvic region in the maximum field of a z-gradient coil. A sinusoidal
magnetic waveform could be varied in frequency as 600, 960, 1,270, and
1,950Hz; also varied were the number of cycles of duration. The variation
of perception thresholds with the number of stimulus cycles, illustrated in
Fig. 9.19, shows characteristics of the theoretical SENN model as in Fig.
4.19-the threshold is maximum with one cycle of stimulation and gradually
diminishes to a minimum plateau as the number of cycles is increased.
Similar dependency on the number of cycles of sinusoidal stimulation has
been reported by Schmitt and colleagues (1994), who demonstrated a mini-
mum plateau at 25 cycles of stimulation at a frequency of 1.0-1.3 kHz, and
by Yamagata and associates (1991), who reported a minimum plateau at 30
cycles of stimulation at 1,250Hz.
Budinger and colleagues also evaluated the variation in perception
threshold with the frequency of stimulation. For a field in the y direction,
the threshold expressed in B units diminished with increasing frequency,
reaching a minimum plateau at high frequencies (=1,950Hz) as expected
by the theoretical models discussed above. However, for fields in the x
or z direction, a plot of thresholds vs. frequency showed sharply declining

thresholds at the higher frequencies, in contrast to theoretical expectations.
Others who have tested perception sensitivity thresholds have found fre-
quency dependency in accord with theoretical models, in which a minimum
B-field plateau is found at frequencies above about 3kHz (Mansfield and
Harvey, 1993).
Suprathreshold Nerve Excitation Reactions

Suprathreshold reactions can be reported as multiples of perception thresh-
olds. For cutaneously applied currents, ratings of unpleasantness or pain are
typically elicited at a perception threshold mUltiple of about 2 for sinusoidal
or repeated stimuli (Table 7.3). Suprathreshold reactions were investigated
by Budinger and associates (1991) using z-coil exposure to the torso of two
human sUbjects. The ratio of pain to perception was 1.3 when averaged over
stimulus frequencies of 960, 1,270, and 1,950Hz. Similar findings were re-
ported by Bourland and associates (1997) using a much larger sample size
(n = 52). For instance, with y-axis exposure, Bourland's subjects reported
definite discomfort at threshold multiple of 1.44, and intolerable pain at a
multiple of 1.89. These reported multiples for discomfort and pain are
considerably smaller than that in electro cutaneous experiments. They
might be explained if subjects were to experience more aversive reactions
associated large area magnetic stimulation as compared with more focal
electrocutaneous stimulation, although this conjecture remains unproven.
9.8 Visual and Auditory Reactions to

Electromagnetic Exposure
Direct Perception of Magnetic Fields
In certain animal species, specialized mechanisms for magnetic field detec-
tion can be demonstrated (Tenforde, 1989). If the dominant mechanism of
human perception is due to peripheral nerve stimulation, then magnetic
thresholds can be determined with the models of Sect. 9.6. At 60Hz, for
example, magnetic thresholds for the most sensitive neural structures are
estimated to be 120mT (peak) forfrontal exposure of the torso and 190mT
for longitudinal exposure.
Magnetic perception thresholds were tested in now-classic experiments
by Tucker and Schmitt (1978). These experiments established that humans
cannot detect 60Hz fields as high as 0.75mT when applied longitudinally to
the whole body, or as high as 1.5mT when applied locally to the head. An
interesting finding was the extreme care that must be exercised to avoid
artifactual perception due to extraneous cues in the experimental proce-
dure, such as vibration, acoustic cues, or dimming of lights. The exposure in
Visual and Auditory Reactions to Electromagnetic Exposure 387
these experiments was below theoretical nerve excitation thresholds by a

factor of about 200. Apparently, if some detection mechanism exists in
humans other than peripheral nerve stimulation, it requires whole-body
exposure above 0.75mT to be effective.
Visual Sensations
A particularly sensitive mode of detecting magnetic fields occurs through
the perception ofphosphenes, which are perceived light patterns induced by
nonphotic stimuli, such as pressure on the eyeballs, or electric energy. They
are often referred to as electro- and magnetophosphenes when induced by
electric currents or by magnetic stimulation, respectively. In an interesting
historical account, Marg (1991) reports the initial discovery of electro-
phosphenes in 1755 through the discharge of static electricity from a
Leyden jar, but the first report of magnetophosphenes was not made until
the late 1900s by d' Arsonvai.
A unique feature of electric phosphenes is their low excitation threshold
and sharply defined frequency sensitivity as compared with other forms
of neural stimulation. Figure 9.20 compares thresholds for magneto- and
electrophosphenes from the experiments of Lovsund et ai. (1980a, 1980b).
Magnetophosphenes were elicited in individuals with the head placed
between the poles of a large electromagnet (temple to temple), and
electrophosphenes from current through electrodes placed on the temples.
Magnetophosphenes are shown in Fig. 9.20 on an absolute scale (Lovsund
et aI., 1980b); electrophosphenes are shown on a relative scale (Lovsund et
aI., 1980a). Somewhat lower magnetophosphene thresholds have been re-
ported by Silny (1986). The threshold of phosphenes depends on the level
of background illumination (Barlow et aI., 1947b), and this parameter must
be controlled in experimental procedures.
A third curve in Fig. 9.20 shows electrical thresholds divided by the
frequency; this display facilitates a comparison of the curve shapes of
magneto- and electrophosphenes. A rationale for this representation is that
the induced E-field (and current density) is directly proportional to the
frequency of the magnetic field, as suggested by Eq. (9.17). Electrical
thresholds, modified in this manner, ought to conform to the curve shape
for magnetic thresholds. The two curves do, in fact, appear quite similar,
suggesting that the magnetic effect is the result of induced current, rather
than a direct action of the magnetic field.
As shown in Fig. 9.21 electrophosphene thresholds tested by Adrian
r,
(1977) increase as with n = 3.5, in contrast to n = 1 for peripheral nerve
excitation (Chapter 4); the figure also shows thresholds for auditory effects,
which will be discussed presently. Adrian also used stimuli consisting of two
frequencies. Phosphenes were produced as long as the difference frequency
was near the most sensitive single frequency (=20Hz), even though the
individual frequencies would have been ineffective if presented singly. This
70 7
Magneto Phosphenes (left axis)
Electrophenes (right axis)
60 Electrophosphene+ frequency ,. 6
I
,, "
50 , , /" 5
f='
,,
/'
g ::2
,, 0
. ,,
>. .t::
<Il
.~
c
40 4 ~
I .;;
CIJ
,,.1
,,,
"0 "iii
x u
:E
u
.~
30 ,
I 3 ~
Qi
c I CIJ
I
Cl I .~
~'" \
\ /
I
..
iii
m
.. ....
'
20 \
\
\
I 2 a:
I
\ , '
......
I
:-"....' .'
~ .'
10
00 0
30 40 50
Frequency (Hz)
FIGURE 9.20. Comparison of phosphene thresholds for magnetic and electric excita-
tion. Magnetic threshold on absolute scale (left vertical axis); electric threshold on
relative scale (right vertical axis). Background illumination at 3cd/m2 • (Adapted
from Lovsund et ai., 1980b.)
finding points to a nonlinear mechanism for the activation of action poten-

tials. Although the neural membrane is highly nonlinear when depolarized
near the action potential threshold (Chapters 3 & 4), it is only slightly
nonlinear when substantially below threshold. Whether the nonlinear as-
pects of the neural membrane can explain the findings of Adrian has not
been explored.
Using Lovsund's observed magnetophosphene thresholds, we can esti-
mate corresponding induced E-fields using an ellipsoidal model of the head
as in Eq. (9.21). Such a treatment is only approximate because the experi-
mental field in Lovsund's experiments was not uniform over the entire
cross-section of the head, as is assumed in the theoretical treatment. Con-
sider sagittal exposure of a head model with semimajor and semiminor axes
of 0.13 and 0.1 m respectively, and a minimum threshold of lOmT at 20-Hz
frequency. With these parameters, the maximum E-field induced within
the head is calculated to be 0.079V/m. At the location of the retina, the
calculated field is 0.053 V/m, which is consistent with the current density
threshold of 0.008A/m2 at the retina determined for electrophosphenes
(Lovsund et aI., 1980b), assuming that the conductivity of the brain is
0.15S/m (Table 2.1).
A somewhat greater E-field threshold was determined in experiments by
Carstensen and associates (1985), in which sinusoidal current was intro-
duced into the eye through a saline-filled cup electrode held against the
eyeball. Carstensen calculated the E-fields within the head using a 3-D finite
element model of the head and diverse tissues. At the most sensitive fre-
quency tested (25Hz), the current threshold was 0.04mA, which corre-
sponded to an E-field at the retina of 0.2V/m.
Phosphene thresholds are quite unlike those for peripheral nerve excita-
tion. For one thing, the internal E-field corresponding to phosphene per-
ception at the optimum frequency is a factor of 100 or so below minimum
1.0
/
/
i 10< f3.~/
./ i / ,/
. /'
// Io<f
10 Frequency (Hz)
FIGURE 9.21. Perceptual thresholds for auditory and visual sensations from
trans cranial stimulation by electric current. Electrode placement (1.75cm dia.) as
indicated. Insert showsr slopes with n = 1 and n = 3.5. (From Adrian, 1977.)
thresholds for neural stimulation (Table 4.2). Furthermore, the frequency

sensitivity differs substantially from that for neural stimulation (Fig. 9.17).
These observations suggest that some mechanism other than afferent
neural stimulation is responsible for the production of phosphenes.
Mechanisms for Phosphenes

Although visual sensations from stimulation of the visual cortex has been
demonstrated (Brindley and Lewis, 1968; Brindley and Rushton, 1977;
Ronner, 1990), experimental evidence demonstrates that phosphenes with
much lower electrical thresholds can be traced to excitation of structures in
the retina, rather than the optic nerve or the visual cortex of the brain. The
visual field for phosphenes occurs in the opposite quadrant relative to the
area of stimulation on the retina, whether through magnetic (Barlow et ai.,
1947) or contact current stimulation (Brindley et aI, 1955), as would be
expected if the site of excitation were in the retina. A subject who was blind
because of the removal of both eyeballs could not experience phosphenes
(Lovsund et ai., 1980a). However, another subject who was blind as a result
of retinitus pigmentosa, a disease of the photoreceptors and epithelium, did
experience electrical phosphenes. This subject was thought to have had
functional bipolar and ganglion cells in the retina.
Brindley and colleagues (1955) performed a remarkable series of experi-
ments to explore the mechanism of electrical phosphenes. One has to
marvel at his dedication in accepting the passage of electric current through
electrodes positioned on his retina-all in the pursuit of scientific inquiry.
Brindley found that phosphenes were most sensitive to current in a direc-
tion radial to the retina, but were difficult to produce when the current
traveled in a tangential direction. This would imply that the affected struc-
tures are most likely cells with a radial orientation, possibly photorecep-
tor cells, although the aforementioned experiments with a blind subject
would suggest bipolar or ganglion cells might be likely (see Fig. 3.24).
From Brindley's data with capacitor discharges, we can infer a threshold
charge density of about 0.09 flC/cm2, which is about a factor of 100 below the
nerve excitation threshold for cutaneously applied capacitor discharges
(Fig. 7.5).
A likely locus of stimulation is the presynaptic junction of cells within the
retina (Lovsund et ai., 1980a; Knighton, 1975a; 1975b). This view seems
reasonable in light of the fact that small changes in the potential of presyn-
aptic cells can be greatly magnified in the potential and excitability of post
synaptic cells (see Sect. 3.7). That could account for the low excitation
threshold of phosphenes in comparison to nerve excitation. Knighton esti-
mated that depolarization of the presynaptic membrane by only 60flV is
sufficient to evoke a visual response. In contrast, in order to excite a periph-
eral nerve fiber, membrane depolarization by about 15 mV is necessary
(Chapter 4)-a factor of 250 times greater.
Duration and Frequency Relationships for Visual Effects

The sensitivity of phosphenes to temporal aspects of a retinal stimulus
differs markedly from that for the nerve and heart. Experimental data show
that the chronaxy or S-D time constant for phosphenes using electrodes on
the temples is about 14ms (Baumgardt, 1951; Bergeron et at, 1995). An
interesting aspect of phosphene S-D curves is that the threshold does not
monotonically decrease with increasing stimulation duration. Rather, the S-
D curve has a long-duration plateau as with neural S-D curves, but exhibits
a dip in the threshold by as much as 30% just before the rising phase at short
durations. The slope of the S-D curve from Bergeron's experiments was
approximately C°.5, in contrast to C 1 for nerve and muscle S-D characteris-
tics, where t is the duration of a monophasic stimulus. This finding is
complicated by the fact that Bergeron used a regulated voltage rather than a
regulated current device for stimulation, which can affect the S-D slope
because of the non-linear impedance properties of the skin (see Sect. 2.2).
In the experiments described previously, Knighton developed S-D
curves for electrically evoked responses, and found S-D time constants, 7:.,
in the range 14 to 36ms (Fig. 9.22). These values are consistent with the
phosphene data described above, but are about 100 times greater than
corresponding values for excitation of peripheral nerve (Chapters 4 and 7),
1000
,,
,,
500 , 1: = 14 ms
'r e
200
<'
.;;
Q)
"C
100
~
.c:
C)
as 50
E 36 ms
If)
::l
"3
E 20
~
10
5
0.5 2 5 10 20 50 100
FIGURE 9.22. Strength-duration curves for electrically evoked potentials in the

retina of the frog's eye. Curves represent various experimental procedures.
(Adapted from Knighton, 1975.)
and are about lO times greater than those applying to excitation of cardiac
and skeletal muscle (Chapters 6 and 8). The relatively large values of 7:e for
visual responses are consistent with strength-frequency data for phos-
phenes, which show an upper transition frequency ([. = 20Hz) that is much
lower than that for nerve (=1,000Hz) or cardiac (=100Hz) excitation.
Phosphenes were tested in five human subjects by Budinger and col-
leagues (1984) using a pulsed magnetic field having a biphasic dBldt wave-
form with a phase duration of lOms. The threshold dBldt was found to be
1.3T/s for younger subjects, and 1.9T/s for older men; subjects were most
sensitive to a repetition rate of 15 per second. These dBldt thresholds are
roughly a factor of two higher than those obtained by Lovsund with sinusoi-
dal magnetic fields having a similar phase duration. If the S-D time constant
is assumed to be 14ms (the lower value obtained by Knighton), one would
conclude that Budinger's thresholds are a factor of 2.0 above a presumed
rheobase for the same subjects.
One should distinguish between the stimulation of the retina and of the
visual cortex, both of which can elicit phosphenes. Phosphenes induced
through stimulation of the visual cortex have 7:e values typical of nerve
excitation (i.e., on the order of O.2ms) (Ronner, 1990). In addition,
rheobase thresholds for visual cortex stimulation are typical of direct
neural stimulation, in contrast to the much lower thresholds with retinal
stimulation.
Electrical phosphenes persist beyond the duration of the applied stimu-
lus, in contrast to neural or cardiac action potentials, whose duration are
on the order of 1 and 200ms respectively. The duration of phosphene
persistence increases with the level of the both electric and magnetic
stimuli (Barlow, 1947). At threshold levels, phosphenes persist less than 1s;
if the stimulus is increased by a factor of 4, durations may exceed 16s.
Following the cessation of phosphenes, there is a refractory period in which
thresholds for subsequent stimulation is increased. With the application of
strong stimuli, refractory periods up to 60s have been observed (Barlow,
1947).
Implications of Phosphenes for eNS

Synaptic Interactions
Although photoreceptors are not found in the central nervous system
(CNS), the brain and spinal column are rich in neuro synaptic junctions. It
is logical to inquire whether CNS interactions are possible at the low phos-
phene thresholds pertaining to retinal stimulation. While we lack a clear
answer to this question, experimental evidence shows that CNS interactions
are indeed possible with magnetic stimulation of the brain at intensities well
below levels necessary to excite neurons. For instance, Silny exposed the
human head to sinusoidal magnetic fields, and recorded visual evoked
potentials (VEP) on the surface of the scalp in response to a visual stimulus.

With a 50 Hz stimulus at a flux density of 60 mT, Silny observed significantly
altered YEP patterns in 12 of 15 test subjects. It is remarkable that these
alterations persisted for as long as 70 minutes after cessation of the stimu-
lus. In addition, his subjects reported headaches and "indisposition" above
60mT exposure. We may compare these values with excitation thresholds
of cortical neuron through exposure of the head by a 50-Hz magnetic field.
For example, the excitation threshold of cortical neurons would be approxi-
mately 1.4 T-peak, assuming a l1-cm spherical diameter for the brain, and
a threshold E-field of 12.3 Vim for a lO-.um neuron (Table 4.2).
The ability of subexcitation fields to alter neuronal response has also
been reported by Bawin and associates (1984; 1986) who exposed hippo-
campal slices from the rat brain to magnetic fields. In these experiments,
neuronal excitability was inferred with measurements of the compound
action potential evoked by a monophasic stimulus. It was found that neu-
ronal excitability could be significantly altered by applying sinusoidal fields
at 5 and 60Hz with in situ intensities as low as 0.5V/m-rms. Both increases
and decreases in excitability could be effected by the sinusoidal fields. The
authors noted that long-term increases in excitability were of comparable
magnitude for 5 and 60Hz fields, and generally lasted from 10 minutes to
hours after the cessation of the sinusoidal field.
Cook and colleagues (1992) reported changes in auditory evoked brain
potentials in human subjects exposed simultaneously to a 9kV/m electric
field and a 20.uT magnetic field. The authors concluded that changes in
exposure level may be more important than the duration of exposure. It
should be noted that these effects were elicited at exposure levels well
below the synaptic effects thresholds discussed above.
Auditory Sensations
The ability to electrically stimulate auditory nerves within the cochlea
has lead to a well-developed technology of cochlear implants for hearing
impaired individuals (Leake et aI., 1990). It is also possible to stimulate
auditory sensations with less invasive electrical means, which will be
reviewed here.
Stimulation at Audio Frequencies

Auditory perception from stimulation at audio frequencies has been
called electrophonics (Flottorp, 1953; Adrian, 1977). Thresholds for
electrophonics determined by Adrian are shown in Fig. 9.21, along with
electrophosphene thresholds for the same electrode arrangement. It
can be seen that the slope of phosphene thresholds versus frequency
is similar to that for neural stimulation and considerably distinct
from electrophosphenes. A further finding of Adrian was that threshold
sensitivity responded to the difference frequency when two frequencies

were mixed in the stimulus, much like similar properties for phosphenes as
noted above.
One cannot rule out the possibility that Adrian's electrophonics may
result from neuromuscular excitation that couples mechanically into nor-
mal auditory channels since the electrophonic thresholds shown in Fig. 9.21
are close to neurosensory and neuromuscular stimulation thresholds with
the indicated electrode size (Chapters 7 and 8). In an investigation of a
variety of methods for stimulating electrophonics, Flottorp (1953) con-
cluded that the mechanism of transduction was of extra choclear origin.
One method requires a dry electrode to be rubbed against dry skin adjacent
to the ear. The underlying mechanism in this case is likely due to electro-
static vibration of the skin, as described by Eq. (7.1). Other mechanisms
involving electrodes external to the ear were thought to be the result of
neuromuscular stimulation. The fact that the perceived electrophonic
frequency was double the stimulating frequency would be consistent with
this explanation.
Stimulation at Microwave Frequencies

So-called microwave hearing was first reported in World War II by radar
operators who heard clicks or buzzing sounds when they were within the
beam of pulsed radar energy. Since that time, much research has been
conducted on the properties and mechanisms of microwave hearing. A
review of the subject and original research have been presented by Lin
(1990). This research persuasively establishes that microwave hearing is the
result of a small transient temperature rise in brain tissue, which, through
thermoelastic expansion, launches an acoustic wave that reaches the inner
ear through normal auditory processes. The temperature rise is extremely
small (=1O- 6o C), as is its duration (=lO,us). Nevertheless, the result is
sufficient to be detected by the inner ear.
Auditory thresholds depend on pulse width, as seen in Fig. 9.23, which
shows perceived loudness as a function of pulse width for radiation at
800 MHz. The auditory threshold is inversely proportional to perceived
loudness. The figure shows resonances which arise from the roughly spheri-
cal resonant acoustic cavity formed by the head. In this example, the first
resonance occurs at about 50,us. For short pulses, thermoelastic displace-
ment or pressure in the head is approximately proportional to the product
of power density and pulse width (i.e., the energy density per pulse). A
"short" pulse is defined such that 2nftd « 1, where f is the fundamental
resonant frequency, and td is pulse duration.
The frequency of vibration is independent of the microwave absorp-
tion pattern within the head and depends only on the equivalent radius of
the brain and its acoustic properties. Theoretical models with an ideal
spherical skull establish the fundamental frequency of vibration as
15
Ul
Ul
Q)
c:
'0
::I
0 10
..J
Q)
>
~
"'ij)
II:
5
O-fooooo:=:......J---JL....L..L..L..LLLL_.....L.......L.....L...L....L..LJLLL..._...L......L...JL...I....L.U.U
10 100 1000
Pulse width (I1S)
FIGURE 9.23. Perceived auditory loudness from microwave exposure at 800MHz as

a function of pulse width, constant incident power density. Data from 18 subjects.
(From Lin, 1990.)
f = knlr, where n is the velocity of propagation (=1,460mls) r is the

radius of the sphere, and k = 0.72 and 0.5 for a constrained and uncon-
strained brain mass, respectively. Experimental values closely follow this
relationship. For instance, with a brain radius of 8cm, the vibration fre-
quency is calculated to be 9.1kHz for a constrained mass. It is significant
that microwave hearing occurs only in individuals with the ability to
hear sound through bone conduction at frequencies above 8kHz (Lin,
1978).
Table 9.10 lists microwave hearing thresholds as a function of pulse
width for various measures of exposure. It can be seen that auditory thresh-
olds occur at a constant value of incident energy density [i.e. product of
peak power and duration (third column) or absorbed energy per gram
of tissue (last column)]. The lowest threshold of peak power density
reported by Lin at an optimum pulse width was 250mW/cm2 , which
would correspond to a specific absorption rate (SAR) of about 0.1 Wig.
These exposures are quite large when compared with standards for elec-
tromagnetic exposure (reviewed in Chapter 11), which limit rms SAR
typically to 0.4 W/kg, and rms power density in certain frequency regimes to
lOmW/cm 2•
TABLE 9.10. Threshold for microwave-induced auditory

effect in human subjects, 45 dB background noise,
2,450MHz carrier frequency.
Pulse Peak Energy Peak Absorbed
width power density SAR energy
(Jls) (W/cm2) (JlJ/cm 2) (Wig) (mJ/g)
1 40.00 40 16.00 16
2 20.00 40 8.00 16
4 10.00 40 4.00 16
5 8.00 40 3.20 16
10 4.00 40 1.60 16
15 2.33 35 0.93 14
20 2.15 43 0.86 17
32 1.25 40 0.50 16
Notes: SAR calculated for equivalent spherical model of the

head. Quantities are per pulse.
Source: Lin (1990).
9.9 Local Magnetic Stimulation3

In recent years, considerable attention has been devoted to using localized
time-varying magnetic fields for stimulating excitable tissue without the use
of electrodes. The main advantage of this technique is that it provides a
noninvasive means of stimulation without causing pain (Barker et ai., 1985).
One potential application is magnetic stimulation of the brain. Recent
results from magnetic brain stimulation include measurement of central
motor conduction time (Barber et ai., 1987; Cracco, 1987; Hess et ai., 1987;
Mills and Murray, 1985), monitoring of spinal cord function during spinal
surgery (Levy, 1987; Shields et ai., 1988; Krauss et ai., 1994), and the ability
to produce motor twitches for a single digit (Amassian et ai., 1988).
Nonmotor brain areas have been stimulated as weli. Some examples of
stimulating cognitive areas are the suppression of visual perception by
stimulating the visual cortex (Amassian et ai., 1989a) and eliciting a sense of
movement in a paralyzed limb (Amassian et ai., 1989b). Stimulation of
peripheral nerve is also possible, but has not received the same attention as
brain stimulation, except in a few special cases such as facial nerves (Evans
et ai., 1988; Maccabee et ai., 1988b; Schriefer et ai., 1988). Magnetic stimu-
lation of lumbosacral roots has also shown utility (Tsuji et ai., 1988;
Chokroverty and DiLullo, 1989). Applications have also included magnetic
stimulation of the heart (Nyenhuis et ai., 1992, 1994) and of the inspiratory
nerves (Voorhees et ai., 1990).
3This section was written by H. A. C. Eaton and J. P. Reilly.

Local Magnetic Stimulation 397
Small Coil Stimulators

The circuitry used to produce a magnetic stimulus is straightforward, but
there are many variations (Davey et aI., 1988; Hallgren, 1973; Merton and
Morton, 1986; Polson et aI., 1982a; Barker, 1994) Typical devices rapidly
discharge a capacitor into a small coil placed near the excitable tissue as in
Fig. 9.24. The figure omits the charging circuit necessary to produce the
initial charge on the capacitor and the inherent series resistance of the
coil, thyristor, capacitor, and cabling. These resistances are of considerable
practical concern and cannot be neglected. Resistor Rs is used only to limit
the current flowing into the gate of the thyristor during firing. Figure 9.24
shows the simplest possible trigger control; many other trigger methods are
possible.
After the switch is pressed in Fig. 9.24, the thyristor is triggered, discharg-
ing the capacitor through the coil. Current that flows through the capacitor
until the voltage at the capacitor terminals reaches zero is described by a
series RLC equation:
(9.23)
Where '1 is the sum of the series resistances of the coil, capacitor, thyristor,
and cabling. This equation is valid for the time interval from t = 0 (when the
thyristor is triggered) until t = ts when the diode begins to conduct, where
ts is given by
tan-l1[2L~1/ LC - (1i/2L f ]1
(1i - 2'2)
(9.24)
ts = --O.~r1/=;=L=C=_=(r,=1/2=L=f~-"-
where '2 is the sum of the parasitic resistances to the right of the diode. A
four-quadrant arctangent function must be used for Eq. (9.24). The time ts
is slightly later than the time at which the current is maximal. After tS' the
current circulates through the diode, thyristor, and coil and is governed by
a series RL circuit equation:
(9.25)
Where Is is the coil current at the time that the diode begins to conduct and
is given by
+
+
c D L
Thyristor
Derivative
150 5
.-.. 120 4
'"
~ 90 3
8
~ 60 2
;g
""0 30
o 0
-30 -1
100 200 300 400 500
Time (microseconds)
Derivative
150 5
120 4 ~ ~ Coil current
90 3
60 2
30
o
~
-;:;- 0 - i ' " - - - r - - " " . . - - - - , - - - - - ' ' r - - - - - - -
-30 -:::: .. 1
25 125
-60 -2
-90 -3
-120 -4 Time (microseconds)
FIGURE 9.24. Upper: a simple magnetic stimulator circuit employing a thyristor

switch. The coil, L, is the stimulus coil located near the excitable tissue. Middle: The
coil current waveform from the circuit, and its derivative. Lower: The coil current
and derivative waveforms that result when the diode is removed from the circuit.
(9.26)
Figure 9.24 shows the shape of the coil current as well as its derivative. The
scales are for a hypothetical circuit having C = 200,uF, Vc = 750V, L =
5,uH, and r 1 = 50mQ. These are realistic circuit parameters, although far
different values are possible. The coil current rises rapidly, followed by a
long decay period. The induced electric field in the tissue, proportional to
dlldt, is biphasic; such is always the case with magnetic stimulation because
the coil current always returns to zero. With the circuit of Fig. 9.24, the
stimulus is a short-duration, high-amplitude pulse. The energy stored in the
capacitor is dissipated primarily as heat in the coil resistance. The diode
prevents reverse charging of the capacitor.
If the diode in Fig. 9.24 is removed, the coil current will be governed by
Eq. (9.23) alone, which is valid since the thyristor will allow current to flow
in only one direction. If the diode is removed from the circuit in Fig. 9.24,
the thyristor switch discharges the capacitor through the coil, which then
recharges the capacitor in the opposite polarity. Using this technique, much
of the energy originally stored in the capacitor is returned to it, and less heat
is dissipated in the coil. Figure 9.24 also shows the waveform of coil current
and its derivative when the diode is deleted from the circuit. The stimulus
and repolarization pulses are of roughly equal amplitude and duration.
For a practical circuit, the capacitor may be several hundred ,uF, and may
be charged to over 1kV. The inductor may range from a few,uH to hun-
dreds of ,uH, depending on the coil arrangement. Peak current of several
kA is typical. Special care must be taken to keep the inherent series resis-
tances of all components to a minimum (usually under 0.1 Q), so that the
LC circuit is as underdamped as possible. Stray inductance in the cabling
connecting the circuit must be kept as small as possible, especially in the
(typically long) cable between the stimulator electronics and the coil. The
ringing frequency of the RLC circuit is usually in the 1- to lO-kHz range.
Considerations of the S-D curve of the target tissue aid in choosing the
optimal ringing frequency (see Sect. 9.6). Because this frequency is gov-
erned only by the values of r, L, and C, it is difficult to adjust the stimulus
duration while maintaining a constant stimulus amplitude. Usually, a single
fixed duration is used, in which case the amplitude of the stimulus is easily
adjusted by changing the initial voltage on the capacitor.
A large amount of energy is stored in the capacitor before each discharge.
Most circuits used to recharge the capacitor build up this charge slowly,
resulting in a long interpulse period (typically a few seconds). Considerable
heating of the stimulating coil results after delivery of many pulses, and coil
temperature is usually monitored on commercial magnetic stimulators.
Multiple action potentials are difficult to produce. One or more seconds are
typically required to recharge the energy storage capacitor, which prevents
stimulation at higher rates. Allowing an LC circuit to "ring" continuously is

not an effective means for producing multiple action potentials. If a low
ringing frequency is used, the magnitude of dlldt will be low unless very
high peak currents are used. Such currents are difficult to handle, and
resistive losses will damped the oscillations quickly, making this scheme
impractical. If a high ringing frequency is chosen, the excitation threshold of
the nerve is increased. More sophisticated circuits for charging and dis-
charging of the capacitor coupled with special coil designs can produce
magnetic pulse trains capable of producing multiple action potentials
(Davey et aI., 1988).
Induced Electric Field

It is not possible to design a coil that would focus stimulation to any desired
location, depth, and orientation of the induced current. The inherent shape
of various body structures places limitations On the possible orientations
of the induced currents. For example, in the near-spherical regions of the
head, it is impossible to induce significant radially oriented (perpendicular
to the skull) currents with a magnetic stimulator (Eaton, 1992). Despite this
limitation, it is possible to excite fibers of diverse orientations within the
brain's cortex because of the positions of neurons within the convolutions
of the cortex, as suggested by Fig. 9.25. The figure shows that an induced
tangential E-field may be oriented for preferential stimulation at the ends
and bends of the cortical neurons, in accord with mechanisms described
Cranial bone
matter
FIGURE 9.25. Anatomical structures and induced E-Field with magnetic stimulation
of the brain. Neurons have diverse orientations with respect to tangential E-field
because of convolutions of the cortex.
in connection with Fig. 4.15. Although one cannot focus excitation at

will within the brain, it is still possible to target excitation to small regions
of the cortex using dual coil systems (Veno, 1994), as will be described
presently.
Coil current, geometry, and the tissue boundary conditions determine the
pattern of the induced currents. For the case of a magnetic stimulator, the
coil size and its distance from the point where the field is being computed
are much smaller than one wavelength. This allows the approximation that
the phase shifts due to propagation delays can be ignored. Tissue conduc-
tivities are small enough that the skin effect can also be ignored. With these
approximations, the induced electric field is computed as (Eaton, 1992)
E = -flo
di
dt L
Oil
dI
4nR - VV (9.27)
where R is the distance between the current filament and the point in the
medium where E is being computed, i is the instantaneous current, and dl is
an element of the current-carrying conductor. The current is assumed to be
confined to a thin filament centered on the coil wire. The term VV arises
from the charge density appearing on the boundaries between different
tissues, or other inhomogeneities. In this analysis, induced currents in the
tissue are assumed to have a negligible contribution to the total magnetic
field, a quite realistic assumption for biological materials. In a homoge-
neous, isotropic medium, either of infinite extent or having appropriately
symmetric boundaries for the coil under consideration, the charge density
will be zero and, consequently, only the integral term above contributes to
the electric field. Several investigators have examined the induced electric
fields for the special case of zero charge density for different coil arrange-
ments. Much work in the area concerns producing a focal region of high
electric field (Maccabee et aI., 1988a; RosIer et aI., 1989).
The boundary conditions are found from the following expression:
aV.E + E·Va = -£V. JE - JE .V£ (9.28)

Jt Jt
In a region where £ and a do not vary spatially, this leads to an exponential
decay of V E and charge density. Because the integral term in the expres-
sion for E already has zero divergence, V2V must equal zero in such a
region. Thus Laplace's equation is solved to find V in a homogeneous
region. On the boundaries of the homogeneous region, E must satisfy the
above relation. For most purposes, the boundary conditions used to deter-
mine V can be computed for the stationary case of dildt = 1, and the
resultant V-field is then multiplied by dildt. This neglects the dEldt terms
above. Omitting these terms neglects the time lag between the application
of the field and the resultant rearrangement of charge; however, this ap-
proximation is good, even for highly capacitive tissues (Eaton, 1992). The
result of this simplification is that the spatial distribution of the field does
not vary with time.
Equation (9.27) can be used to predict the induced electric field inside the
body for nearly any practical coil arrangement. The closed line integral is
performed along the path of the coil winding. For the case of a circular coil,
there is no closed-form solution to the integral, but it can be represented in
terms of the complete elliptic integrals (Jackson, 1962). If simple body
models are used, the determination of the V-field is not difficult. The finite-
element approach has been used for simple-body models (Ueno et aI.,
1988), and can be applied to more complicated problems as well.
Local Stimulation of the Brain Cortex

A simple model for the brain cavity is that of a uniform, spherically shaped
conductor bounded by an insulator, representing the skull. Various theo-
retical models have been applied to this problem (Roth et aI., 1991;
Grandori and Ravizzani, 1991; Eaton, 1992). Using the model of Eaton, Fig.
9.26 illustrates the magnitude of the electric field along a great circle that
encloses the cortex at a distance 2cm below the center of the coil (1.5cm
below the surface of the scalp). The cortex is assumed to have a diameter of
11 cm. An insert in the figure shows the top view of a spherical skull on
which is placed one 6-cm coil (upper panel) or two 5-cm coils (lower panel).
The vertical axis gives the induced E-field on an arbitrary scale that depends
on the rate of change of current in the coil. The field magnitude is repre-
sented without regard to its direction. The E-field just beneath the coil
mainly follows circular patterns approximating the shape of the coil when
shown in the top view, and is primarily circumferential when shown in the
side view. At greater depths the E-field would exhibit reduced peak E-
fields, and would be less focal (since the depth of affected tissue is at a
greater proportion of the coil dimension).
To illustrate the vertical scaling in the upper panel, a peak current
of 3,700A with a rise time of O.2ms into a single winding (dI/dt = 1.85 X
107 Als) would induce a peak E-field of 18.75V/m, which is the theoretical
E-field threshold (according to the SENN model) for a 10-.um fiber excited
by a 0.2ms pulse. The lower panel shows E-field patterns induced by a
so-called figure-of-eight (F8) coil, in which the current is in a reinforcing
direction at the interface of the two coils. To scale this figure, note that a
peak coil current of 2,600A with a rise time of 0.2ms (1.3 x 107 A/s) in each
coil loop would produce a peak E-field of 18.75V/m. Not only is the F8 coil
more efficient, but it is more focal, i.e., the E-field pattern is concentrated in
a smaller region of the cortex. In this example, we have cited thresholds
applicable to a terminated fiber having a diameter of 10 -.urn. In the human
pyramidal tract, 1.7% of myelinated fibers have diameters of 10.um or
greater-some as large as 20.um (Lassek, 1942). Consequently, the cited
thresholds would apply to the minority of fibers. For smaller fibers,
thresholds are expected to rise in inverse proportion to fiber diameter (see
Chapter 4).
30
£l
·2 X
::I
~ 20
e
:1:
-e
~
"0
ai
~
"0
<D
10
0
::I
"0
.E
0
-10 -5 0 5 10
Distance along cortex (cm)
40
dual 5-cm coils Y
£l
·2
30 ,
alongY _ _ ,'
::I
, X
,,
~
~
:e~ 20 ,
"0
ai
along X ---f-.
~ I
I
"0
<D
0
::I
10
"0
.E
o
-10 -5 o 5 10
FIGURE 9.26. Magnitude of peak E-field along cortex of brain 2cm beneath stimulat-
ing coils. Upper: single 6-cm coil; lower: dualS-cm coils. Vertical axis is on relative
scale that depends on coil dJ/dt.
For many diagnostic applications of brain function, it is often important

to confine the area of stimulation to as small a region as possible (Krauss et
aI., 1994). In attempting to increase the focality of magnetic stimulation,
one confronts limitations dictated by the physics of magnetic induction. We
cannot necessarily achieve greater focality by simply making the coil
smaller, because the induced E-field fields at the cortex may actually
become less focal as a result of the fact that the target tissue is at a greater
depth in proportion to the coil size. Furthermore, there would be a less
efficient coupling area of a small coil, thereby increasing the requirements
on peak coil current. While the F8 coil does achieve greater focally and is
more efficient than a single coil, it may be desirable to increase the focality
even more for clinical applications.
One can achieve greater focality with a unique F8 coil design that makes
use of the nonlinear electrodynamics of the neural membrane (Reilly et aI.,
1993). A subthreshold rectangular conditioning pulse (CP) is provided to
one loop of an asymmetric F8 coil either before or during a sinusoidal
stimulus, which is applied to the second loop. The interaction effects of a
conditioning pulse with a sinusoidal stimulus are shown in Fig. 4.24, which
7 x 10-4
..-, 6 x 10-4
~
..-,
5 x 10-4
'S
'0 -+--++-+fi--il--..... x
.c
CIl
~ 4 x 10-4
;;
@
'E 3 x 10-4
~
::;
2 x 10-4
()
'0
()
.:i.
~ 10-4
,,
0
-10 -5 0 5 10
FIGURE 9.27. Excitation threshold requirements with dual-function stimulation. S-

cm coil has 0.2ms current ramp @ 75% threshold. Vertical axis shows inverse of
current in 2.5-cm coil needed to excite neurons at various locations along brain
cortex as indicated on horizontal axis. Small coil has sinusoidal current at 80 kHz
frequency.
TABLE 9.11. Threshold response with various coil configurations.

Coil configuration
y y y y
coil size
coil 1 (cm) 6 5 6
coil 2 (cm) 5 2.5 2.5
waveform
coil 1 0.2 ms ramp 0.2 ms ramp 0.2 ms ramp
coil 2 0.2 ms ramp 80kHz, t=1 ms 80kHz, t=1 ms
Min. threshold current

coil 1 (A - pk) 3703 2597 2800
coil 2 (A- pk) 2597 2632 1538
Spatial res. @ 130% T

ill( (cm) 2.52 2.12 4.65" 1.22
tJ.y{cm) 3.85 3.19 2.32 2.12
illct..Y (cm 2) 9.70 6.76 10.79 2.59
* width encompassess two peaks beneath coil
is derived from the SENN model. Results from a particular implementation

of this technique is shown in Fig. 9.27, in which a 0.2ms CP is applied to a
5-cm diameter coil at 75 % of the excitation threshold (2,800 A peak current,
0.2ms duration) of a lO-,um nerve located in the highest induced field (see
Fig. 9.25), and a sinusoidal wave at 80kHz and lms duration is applied to an
adjacent 2.5-cm coil. The vertical axis in Fig. 9.27 indicates the inverse of
peak current in the small coil needed to bring a lO-,um fiber to threshold if
it is located at the position indicated on the abscissa. It is assumed that the
fiber is a terminated one, with an orientation optimized for stimulation. For
instance, the current threshold for a fiber located at the peak in the figure
would require a peak sinusoidal coil current of l,500A in the 2.5-cm coil to
achieve threshold.
As a measure of focality, we extracted the width of the excitation region
when the stimulus is raised to 130% of the threshold of a lO-,um nerve
in the optimum location. A justification for this arbitrary metric is that one
cannot precisely target the threshold of a most sensitive neuron at the peak
of the induced E-field, and that a group of neurons with diverse thresholds
and orientations would likely be brought to excitation over a region of the
cortex. Results using this metric are shown in Table 9.11 for several stimulus
arrangements, including a 6-cm coil, a 5-cm F8 coil, a 2.5-cm coil, and a

dual function stimulator with 6- and 2.5-cm coils. The table indicates the
coil sizes, the current waveforms in each coil, the peak threshold current
for a lO-,um neuron in the most sensitive location, the spatial width of
the excitation region when the stimulating current is raised to 130% of its
minimum threshold value, and the approximate excited area. This last
parameter indicates the focality of the various schemes. The spatial resolu-
tion of excitation with the dual function stimulator is significantly smaller
than with the other three configurations. For instance, the excited area with
the dual function apparatus is 0.38 times that of the 5-cm F8 coil.
9.10 Scales of Reaction: Power Frequency Magnetic

Field Exposure of the Head
Figure 9.28 illustrates a scale of reactions to power frequency (50, 60Hz)
magnetic field exposure. It is assumed that the adult human head is uni-
formly exposed to the field. The right-hand column indicates reactions that
may occur with stimulation of neurons within the retina or the brain; the
magnitude of the stimulus is placed at a vertical position such that the
relevant threshold can be read on four scales that indicate various metrics of
the exposure. The right-most numerical scale indicates the magnetic field
flux density of a 60 Hz field expressed in milligauss (mG) and Tesla (T) units;
the next scale to the left shows the temporal derivative of the field (units:
Tis); the next scale expresses the maximum induced E-field (units: Vim) at
the site of stimulation; the left-most scale expresses the current density
(units: Nm 2 ) at the site of stimulation, assuming a brain conductivity of
0.15S/m.
The thresholds indicated in Fig. 9.28 are approximate rms values at the
median of a subject distribution under the stated conditions. The statistical
distribution of thresholds among a large sample of subjects is lacking, and
one can only estimate the median value and its variation. However, if the
variation of thresholds for synaptic effects parallels that for peripheral
nerve excitation (Table 7.10), one would anticipate that sensitivity at the
one percentile rank among a large sample of subjects would be about a
factor of 2 or 3 below the median. Indeed, Silny (1986) reported the lowest
phosphene thresholds at 5mT, applying to a frequency of 20Hz, in contrast
to the presumed median value of lOmT at the same frequency from the data
ofLOvsund.
The four scales of Fig. 9.28 can be tied together only for magnetic expo-
sure with a specified frequency and spatial distribution-in this case, uni-
form exposure of the head by a 50 to 60 Hz field. If we were to change the
frequency, or the spatial distribution of the applied field, the thresholds and
the interrelationships among the four scales would change, in accord with
the principles expounded in Sect. 9.5 and 9.6.
Scales of Reaction 407
Max. 50-Hz 50-Hz

Current Max. dB/dT B-Field
Density in E-Field Applied Applied
brain in brain to Head to Head
(Alm 2 ) (VIm) (TIs) (mG) (T) BrainlCNS Reaction
I- Seizures (ECT)
100 -
10 - 1000 I-
10 I- - Stimulation of 10-llm brain neurons

1 - 100 -
100 mT
- VEP alteration
1 I-
0.11- 10 -
- Phosphenes, synaptic activity
alteration
10mT
0.1 I-
1-
0.11-
1000 100ilT
100 (Typical Environmental Exposures)
_ - - -Max. field in HVTL ROW t

} Typical fields near appliances
10
1· 0.1 IlT _____ Typical home, not near

appliances
FIGURE 9.28. Human responses to power frequency magnetic exposure of the adult
human head; spacially uniform magnetic flux density. Reaction thresholds are
approximate rms values for median response.
The lowest reaction indicated is that of visual phosphenes. The phos-

phene threshold for 50 to 60 Hz stimulation is shown at a flux density of
20mT-rms. Experimental evidence discussed in Sect. 9.8 strongly suggest
that the responsible mechanism is the alteration of synaptic activity within
the retina. A similar mechanism might be active at neural synapses within
the brain, and therefore the phosphene threshold might apply to other
central nervous system interactions. Evidence for this conjecture was pro-
vided by Silny (1985), who reported that 50Hz magnetic exposure to the
head can alter visual evoked potentials (VEP) at flux densities of 60mT-
somewhat above the phosphene threshold, but well below neural excitation
thresholds. Silny also reported that his experimental subjects experienced
headaches and "indisposition" with 60mT exposure at a frequency of 50Hz.
Other evidence for CNS effects with sub-threshold electrical stimulation
has been developed with experiments using in-vitro animal brain prepara-
tions (Wachtel, 1979; Bawin et aI., 1984, 1986). Those studies demonstrated
that an in-situ E-field below the threshold of neural excitation is capable of
altering the excitability of brain neurons.
The next threshold shown in Fig. 9.28 applies to stimulation of neurons
having an axon diameter of lO,um, for which the rheobase rms E-field
threshold is 8.8 Vim (12.4 Vim peak) according to the myelinated nerve
model described in Chapter 4. The excited fiber is assumed to reside in the
cortex, where the induced E-field would be greatest within the brain. The
assumed fiber diameter is a relatively large one among the distribution of
fibers to be found in the brain, although a small minority of fibers as large
as 20,um have been reportedly found in the human pyramidal tract (Sect.
3.4). Excitation thresholds are inversely proportional to fiber diameter in
accord with nerve excitation models (Sect. 4.5).
The highest threshold shown in Fig. 9.27 applies to the induction of
seizures. Although I am not aware of any reports of seizures from magnetic
stimulation, I have estimated the seizure threshold based on experience
with surface electrodes used in electro convulsive therapy (ECT). For this
purpose, I estimated the E-field threshold within the brain using the current
density distribution data derived from models of surface electrodes on the
head (Rush and Driscoll, 1968; Weaver and Rush, 1976), along with the
applied current thresholds reported for ECT (Sackeim et aI., 1987).
The resulting magnetic threshold is a conservative estimate. Indeed, mag-
netic stimulation of the brain without untoward results has been demon-
strated by many researchers (Deno, 1994a), as described in Sect. 9.9.
While ECT has been useful in treating depression, the treatment has
some significant side effects. As an alternative, repetitive trans cranial mag-
netic stimulation (rTMS) has been explored for treatment of depression
without the inducement of seizures (George et aI., 1995). The authors
reported significant improvements in some drug-resistant patients when
rTMS was applied to the left prefrontal cortex of the brain using an F8 coil
(see Sect. 9.9). The level of stimulation was 80% of the threshold needed to
evoke a motor twitch of the abductor pollicis brevis muscle when the
stimulator was placed over the motor cortex. Twenty-minute treatments
were performed daily over several weeks. The main side effect, observed
in two patients (n = 6), was the development of mild headaches that
responded to aspirin or acetaminophen.
In principle, it is possible to create measurable short-term thermal effects
within the brain with magnetic field exposure. For instance, consider the
current density of 45 A/m2 at the seizure threshold shown in Fig. 9.28. The
corresponding specific absorption rate (SAR) is 12 Wlkg in accord with Eq.
Magnetic Forces on Moving Charges 409
(11.8). As indicated by Eq. (11.11), a temperature rise of 1 °C would require

a duration of exposure of 333s at the indicated current density. In this
example, aloe temperature rise requires a very high level of magnetic
exposure for a long duration. However, if the frequency of the magnetic
field is increased, the same temperature rise would be associated with a
smaller magnetic field, since the induced E-field increases with frequency.
Thermal considerations with high-frequency magnetic field exposure are
treated in Sect. 11.5.
At the lower portion of Fig. 9.28 are listed typical environmental expo-
sures in nonoccupational settings. The examples show field levels that may
be encountered within the rights of way of high voltage transmission lines,
near household appliances (Gauger, 1985), or within a typical home. These
environmental exposures are several orders of magnitude below the lowest
reaction threshold shown in the figure. In certain occupational settings,
however, much higher magnetic field exposures may be encountered. For
instance, workers in the electrosteel industry or welding industries can
encounter magnetic fields at power frequencies that approach the indicated
phosphene thresholds (Lovsund et aI., 1982).
9.11 Magnetic Forces on Moving Charges

Magnetohydrodynamic Effects
When a charged particle moves within a magnetic field, it experiences a
force
F= Qv xB (9.29)
where F is the force, Q is the charge on the particle, v is its velocity, and B
is the magnetic flux density; bold type indicates vector quantities. Accord-
ing to the cross-product expressed in (9.29), the vector quantities are in a
mutually perpendicular orientation (i.e., velocity, field, and force oriented
along x, y, and z directions, respectively). The magnitude of the force is
given by
F= QvB sinO (9.30)
where 0 is the angle between v and F. This force, known as the Lorentz
force, causes moving charges to drift within a magnetic field in a direction
perpendicular to their flow. Because of this drift, there results a separation
of charges, which can be detected as a voltage known as the Hall effect. The
magnitude of the Hall effect voltage is
v = vB d sinO (9.31)
where d is the distance between two plates used to sense the voltage. This
principle has lead to many useful devices, including a blood flow meter
(Kolin, 1952). In a biological medium, blood flow is an example of moving
charges, which, in a magnetic field, would be expected to experience a

Lorentz force, and a Hall voltage. To illustrate the application ofEq. (9.31),
consider a flow rate of 0.6mJs in the human aorta within a field of 1 T.
Assuming a vessel diameter of 2.5 cm, the maximum induced potential
across the aortic cross section is 15mV according to Eq. (9.31).
The Hall effect voltage can be observed in an ECG as an enhancement of
the T -wave amplitude in rats exposed to strong static magnetic field (Gaffey
and Tenforde, 1981). The average minimum field strength at which a mea-
surable enhancement occurred was reported as O.3T; at 2T, the average
enhancement was 408%. The authors determined that this enhancement
was caused by the superposition of a Hall effect potential on the natural
ECG signal during the repolarization (T-wave) phase of the heart. During
this phase, ejection of blood into the ascending aorta would give rise to a
flow potential according to Eq. (9.31), thereby augmenting the T-wave. It
was observed that the T-wave returned to normal immediately after cessa-
tion of the magnetic field, even with exposures up to 5 h. The authors
observed no arrythmias, changes in heart rate, or changes in respiration rate
for exposure up to 2 T. Similar studies with primates demonstrated no
measurable changes in blood pressure for exposure up to 1.5 T (Tenforde et
aI., 1983). Others have demonstrated no significant difference in cardiac
excitation thresholds with pulsed magnetic stimulation with or without a
concurrent 1.5T static field (Nyenhuis et aI., 1992). A transient increase in
blood flow was observed in mice exposed to an 8Tfield (Ichioka et aI., 1998).
Other investigators (Jeneson et aI., 1988) were able to detect a statisti-
cally significant increase in the cardiac cycle length of 17% in resting human
subjects within a longitudinal 2-T static field. The authors suggested that
these effects might have been seen in human subjects but not animal sub-
jects as a result of the relatively larger human dimensions. The authors
opined that the observed effect is probably harmless in healthy subjects, but
that its safety in dysrhythmic patients was not certain.
Magnetohydrodynamic forces can also exert a drag on blood flow in very
intense static magnetic fields. Theoretical models predict drag forces will
affect vascular pressure by less that 0.2% in a field of lOT (Keltner et aI.,
1990); others have predicted a change in axial velocity of blood flow by a
few percent in a 1 T field (Dorfman, 1971).
Additional effects attributed to magnetohydynamic forces include ver-
tigo, taste sensations and nausea in human subjects exposed to static fields
of 4.0T (Schenk et aI., 1992). Sensations of vertigo and taste were also
reported by a few subjects in a 1.5-T field. In all cases, the reported sen-
sations were associated with rapid head movement within the field. Some
subjects also reported phosphenes during rapid eye movement in the 4-T
field. For subjects lying stationary within the field, the effects were absent.
The cause for the reported effects was not considered in this study, but one
may hypothesize mechanisms involving mechanical forces on the inner ear,
or induced electrical potentials.
Magnetic Forces on Moving Charges 411
Ion Resonance
An unrestricted ion traveling within a magnetic field will travel in a circular
path because of the Lorentz force. If a static field is combined with an
alternating field, a resonant condition will exist when the alternating field
has a frequency
(9.32)
where Ie is the frequency of the alternating field, (q/m) is the charge-to-mass

ratio of the ion, and Bo is the magnitude of the static field. This condition is
known as ion cyclotron resonance (ICR).
Certain biological reactions have been observed to be sensitive to ex-
tremely low frequency alternating magnetic fields at frequencies that follow
ICR conditions within a geomagnetic field (Postow and Swicord, 1996). For
instance, the 40Ca2+ ion, with a q/m ratio of 4.82 x 106 c/kg, has a resonance
frequency of 34.8 Hz in a static field of 50,uT (Liboff and Parkinson, 1991).
The existence of a pure ICR mechanism in biological systems is not a
realistic expectation because of the requirements of a large, unrestricted
path for the ion. Proponents of ion resonance mechanisms do not suggest
that cyclotron resonance is actually occurring in the biological medium, but
rather that some mechanism is present that follows the ICR relationship.
One suggested mechanism involves ion parametric resonance (IPR), as
proposed by Lednev (1991), and later clarified by Blanchard and Blackman
(1994). The IPR model predicts that a resonant condition is created by the
combination of a static and an alternating magnetic field, which influences
transitions of energy states of ions, and that these transitions can affect
biological activity within cells. The critical IPR frequencies coincidentally
occur at multiples of the cyclotron resonance frequency. The IPR theory
predicts that resonance conditions depend on the ratio of flux densities of
the alternating and static magnetic fields, whereas the magnitude of the AC
field does not appear in the ICR formulation (Eq. 9.32). Empirical tests
with cellular preparations show the response of neurite outgrowth can be
affected by application of static and alternating magnetic fields that follow
the IPR formulation (Blackman et al., 1994).
10
High-Voltage and High-Current
Injuries 1
MICHAEL A. CHILBERT
10.1 Introduction
Injuries resulting from electrical accidents can include tissue destruction,
cellular excitation, and trauma secondary to the passage of current.
Thermal injuries in the extremities can lead to amputation because of the
deep nature of the burn. Cell lysis can also destroy tissue if there is a
sufficient electric potential across the cell membrane. The effects of lysis are
sometimes delayed. Cellular excitation of muscle and nerve can lead to
cardiac fibrillation or transient neural dysfunction. Secondary injuries result
from flash burns, falling, or gross contraction of muscles.
Electrical accidents result in nearly 1,100 deaths per year in the United
States, 10% of which are caused by lightning. One-third of these deaths are
caused by voltages below 1,000 V occurring in the home and workplace.
Low voltages account for more than half of the industrial deaths (Dalziel,
1978), High-voltage deaths are typically from industrial accidents. About
2% of low-voltage accidents and 10% of high-voltage accidents are fatal.
The number of accidents resulting in survivable injury are not consis-
tently documented, because victims are not admitted to the same critical-
care areas within a given facility, and those not surviving at the scene are
not admitted. Victims of severe burns are, of course, admitted to burn-care
units, whereas those afflicted only with cardiac maladies are not. Typical
injury reports include only the burn victims. The evaluation of the burn's
extent is very difficult and may require multiple procedures. Electric burn
injuries account for 4% to 6% of all admissions to burn-care facilities
(Hammond and Ward, 1988; Hunt et aI., 1980; Rosenberg and Nelson,
1988). These burns cover an average of 12% of the body surface area
(BSA), but result in limb amputation for 50% to 70% of the cases. The
This research was supported in part by NIH research grant GM34856 and the
1
Department of Veterans Affairs Medical Center Research Center Research Funds,
Milwaukee, Wisconsin.
412

Modes of Injury 413
average BSA resulting in subsequent death is only 34 %, compared to a

much higher BSA percentage for thermal burns.
This chapter emphasizes electrical burn injury and its theoretical basis.
The physical parameters affecting the character of injury and including
impedance considerations and current distribution throughout the body are
also included. Discussions of thermal and nonthermal trauma follow. Light-
ning injuries often have unique characteristics and traumatic sequelae;
these are to be discussed as well.
10.2 Modes of Injury

The following sections provide background for the forms of trauma seen
clinically. Thermal trauma has the greatest consequence to survivors of
electric accidents; loss of limb is often the result. Nonthermal trauma has
the form of lesions in neural and connective tissues. Lightning injuries often
involve lesions exclusive of burns. Other trauma affects the heart rhythm
and can result in fibrillation.
Thermal Injury
Thermal injury is caused by heating of tissue from the passage of current.
The amount of heat generated in the tissue depends on spatial and temporal
patterns of current density and tissue resistivity. The current density and
resistivity are, in turn, affected by the heat generated in the tissue. Conse-
quently, thermal injury involves a feedback process.
Heat generated in the tissue is given in terms of energy density:
(10.1)
where Qj is the thermal energy density (J/cm3), J is current density (A/cm2),
Q is the resistivity (Qcm), and t is the duration of current. In its simplest
form the thermal energy density can be related to the change in tempera-
ture in the following manner:
(10.2)
where p is the tissue density (g/cm3), c is the specific heat of tissue (J/g0e),
and liT is the temperature change (0e). Equation (10.2) assumes that there
are no thermal losses by conduction to blood and adjacent media, or by
convection or radiation into air. In other words, all boundaries of the
specific volume are adiabatic, and there are no internal heat sinks or sources
in the volume. By solving for the change in temperature using Eqs. (10.1)
and (10.2),
liT = Qj = J2 Qt (10.3)
pc pc
414 10. High-Voltage and High-Current Injuries
Thermal injury depends on the duration of exposure and the temperature in

accordance with Figure 11.12 (Henriques, 1947; Henriques and Moritz,
1947). Cutaneous burns occur when the temperature is elevated for a suffi-
cient length of time: 45°C requires more than 3 h, 51°C requires less than
4min, and 70°C requires less than Is for injury. Temperature levels that
cause injury in other tissues are similar, as is the injury rate for a given
temperature. Electrically induced thermal injury of muscle begins at 43°C
with a I-A current in the limb for 15min, and at 46°C the tissue damage is
much greater (Chilbert et aI., 1985b). Electrically induced thermal damage
to nerve has been noted to occur at 48°C after several seconds. Necrosis
limits have been measured for various tissue types in humans and animals
(Heroux, 1992). For example, with 2.75min. of exposure, the temperature
for necrosis was found to be 48.1 °C in human skin, 50.9°C in porcine skin,
and 49.3°C in porcine muscle.
The deposition of heat depends on the current density and resistivity.
Resistivity decreases with increasing temperature, causing further increases
in current density and temperature. This aspect of resistivity change is
rarely considered in burn models because of the resulting nonlinear equa-
tions. Heat is removed from the tissue by conduction into adjacent tissue, by
convection and radiation into the air at the surface, and by blood flow. To
account for conduction and blood flow effects, one can expand Eq. (10.2)
into:
a2 T a2 T a2 T a2 T
pc-2- = k --2 + k -2- + k --2 + WPbCb(Tb - T) + qm + qj (10.4)
dt ax dy az
where T is the tissue temperature, k is the thermal conductivity of tissue, W
is the blood perfusion rate, Ph is the blood density, Cb is the specific heat of
blood, Tb is the initial temperature of blood, qm is the volumetric rate of
metabolic heat generation, and qj is the volumetric rate of electrical heating.
Note that the first three terms of the equation describe the tissue conduc-
tion, and the fourth term describes the effects of vascular convection. Equa-
tion (10.4) is termed the bioheat equation, which was expressed in its initial
form by Pennes (1948) and has been modified extensively for various appli-
cations (Song et aI., 1988). Table 10.1 lists measured values of thermal
TABLE 10.1. Thermal conductivity values (k) for skin,

fat, and muscle.
Conductivity, k x 103
Medium (W/cm°C)
Amorphous 2.09
Skin 3.40
Fat 3.23
Muscle 5.00
Water (40°C) 6.32
Source: Data from Pennes (1948), Olsen et aI. (1985), Song et

al. (1988).
Modes of Injury 415
conductivity for skin, fat, and muscle of human and hog tissues. These
values have been used in thermal models. Values for the thermal conductiv-
ity of water are included for comparison. The thermal conductivity of blood
used by Song et ai. (1988) was S.O X 103 W/cm 0c. Other parameter values
used were Pb = LOS g/cm3 and Cb = 3.8J/g°C. The value for w is 2 to 12(ml
blood)/(ml tissue )/s for dermal perfusion.
Lee and Kolodney (1987b) developed a unidimensional, axisymmetric
model for the heating resulting from a high-voltage electrical contact in the
upper limb. The limb was modeled as a coaxial cylinder having a lO-cm
diameter and included layers to represent bone, muscle, fat, and skin of
1.0-, 3.S-, 0.3-, and 0.2-cm thicknesses, respectively. The bioheat equation
used for this model included the term for electrical heating while neglecting
metabolic heat. Lee and Kolodney show that tissue perfusion is critical for
the removal of heat in the tissue, but does not significantly modify the
heating process during the application of current.
Thermal injury depends on the current path, which is determined by the
contact points on the body (sometimes termed entry and exit sites). Inside
the body, current distribution is determined by tissue resistivity. Early
perceptions of current in the body stated that it would flow equally through
all tissues or that it would flow along the path of least resistance (in the
vessels and nerves). Both are wrong, but are still referred to in clinical
reports. Experimental evaluation of the current path through the body has
shown that the resistivity of each tissue in a given cross-section determines
the current distribution through the cross section (Chilbert et aI., 1983,
1985b, 1988, 1989; Sances et aI., 1981a, 1983).
Tissue resistivity may be anisotropic; that is, the resistivity can vary
with current direction (Sect. 2.1). For example, muscle has a lower re-
sistivity along its fibers than across them, and skin has a high resistivity
across its layers and a lower resistivity along its layers. Some tissues can
be considered isotropic. The implications of anisotropic resistivity in
thermal burns is that the direction of current through a given tissue may
explain selective tissue trauma seen away from the contact site. This selec-
tive tissue destruction is well documented and is seen in muscle tissue near
the bone (Artz, 1974, 1979; Hammond and Ward, 1988; Luce et aI., 1978;
Moncrief and Pruittion, 1971; Pontionen et aI., 1970; Rosenberg and
Nelson, 1988; Sances et aI., 1979; Wang et aI., 1984, 1985, 1987; ZeIt et aI.,
1988).
Selective thermal trauma can also be related to tissue resistivity changes
with temperature and events at the contact site. As the tissue temperature
increases, the resistivity decreases, altering the current distribution and
subsequent temperature changes in the tissue. Contact site events that
change the current distribution are desiccation and arcing. Desiccation of
skin at the contact will cause the total current to decrease rapidly. This
limits the contact time at voltages below 1,000V and the trauma is usually
limited to the contact region. With higher voltages, arcing can occur over
the desiccated tissue and can alter the current path.
Electroporation
When the electric field strength across the cell membrane is sufficiently
large, the cell membrane will rupture and cause cell lysis (Lee and
Kolodney, 1987a; Lee, 1992; Weaver, 1992). The process is called
electroporation, which is an increase in membrane permeability as a result
of development of aqueous pores in the membrane (Lee et aI., 1988).
Electroporation of the cell membrane leads to rupture when the pore size
becomes large enough or when fusion of several pores occurs. Calculation
of the membrane potential uses cable theory, which is discussed in Chapter
4, and depends on the cell length, cell diameter, the internal fluid conductiv-
ity, the membrane conductivity, and the thickness of the cell membrane,
and has been derived for electrical injury situations by Gaylor et al. (1988).
The membrane potential causes the pore size to increase proportionately
until an irreversible state is reached. This occurs when the pore size reaches
a diameter of one-half the thickness of the membrane. The membrane
potential causing rupture is in the range of 800 to 1,000 mV. Reversible
electroporation has been noted to occur below a 200-mV membrane poten-
tial. The incidence of pores has been evidenced by experimentally observed
increases in membrane permeability, but the exact mechanism of electrical
membrane failure has not been determined, although it is thought to be
initiated as molecular defects in the bilayer component of the cellular
membrane (Lee and Kolodney, 1987a; Lee et aI., 1988).
Electroporation can be detected in culture as a sudden reduction of
impedance across the cellular membrane. The membrane voltage required
for electroporation depends on the duration of the stimulus, as illustrated in
Fig.lO.l. In this example, a minimum breakdown potential signaling revers-
ible electroporation is achieved in a giant algal cell for stimulus durations
in excess of about 100 fts; breakdown can occur with shorter stimulus
durations in accordance with a strength-duration (S-D) law not unlike
that observed for neural excitation effects (Chapter 4). Reversible
electroporation has applications in cellular biology research (Neuman et aI.,
1989; Teissie and Rols, 1994), as it can provide a means for inserting
chemicals into the interior of a cell, while otherwise leaving the cell intact.
The role of electroporation in electrical injury is related to the electric
field strength in the biological medium. Studies of muscle cells in culture by
Lee et al. (1988) have shown that an electric field strength between 50 and
300 V/cm can disrupt cells of 1 rom length. This is consistent with values
of the membrane potential of his earlier study (Lee and Kolodney, 1987a).
For smaller cells, such as fibroblasts of lO-ftm diameter, the electric field
strength that causes rupture exceeds 1,000V/cm. The relation of this data to
accident victims is given in Sect. 10.5. For small cells, the membrane poten-
tial depends on the cell length, cell diameter, and membrane thickness
(Gaylor et aI., 1988); for long cells, the membrane potential depends on the
length constant and the electrical field strength (see Chapter 4). Muscle
Modes of Injury 417
3
T= 18°C
Giant Algal cell
~
CD
Cl 2
If
f fI
.l!!
"0
>
c:
fd+ f
;:
0
~
as
~
co
! I
'" !--f-! t
0
0.5 10 100 1000
Stimulus duration (J.IS)
FIGURE 10.1. Strength-duration relationship for reversible cellular membrane

breakdown. Vertical axis indicates potential across cell membrane. (Reprint
from Benz and Zimmerman, © 1980 with permission from Elsevier Science, The
Boulevard, Langford Lane, Kidlington OX5 1GB, UK.)
cells greater than 1 cm in length can be considered electrically long. The

electric field needed for electroporation decreases as temperature in-
creases. Consequently, the tissue trauma caused in a region of elevated
temperature is related to both electroporation and temperature.
Fibrillation
The effects of electricity on the heart are treated in Chapter 6. This chapter
will focus on short-duration, high-voltage, and high-current applications.
Power-line accidents typically result in severe burns, whereas household
voltages result in fibrillation (Sances et aI., 1979). Most deaths at power-line
voltages are attributed to burns or to secondary trauma such as falling.
Severe burns or complications arising from them are the causes of death in
the hospital. Rarely is instantaneous death attributed to fibrillation at high
voltages, yet several case reports of latent cardiac anomalies have been
published (Ahrenholz et aI., 1988; Dixon, 1983; Guinard et aI., 1987;
Hammond and Ward, 1988; Rouse and Dimick, 1978; Wilson et aI., 1988).
However, fibrillation can occur at high current levels with very short appli-
cation times, suggesting that these levels are not constant and may vary in
a similar fashion to the minimum fibrillation level.
Ferris et al. (1936) published a comprehensive study on fibrillation from
currents up to 17 A for a duration of O.3s. Until recently, most studies of
fibrillation have used current levels below lOA to delineate minimum elec-
trical thresholds for the cardiovascular system. Kouwenhoven (1964) re-
ported that fibrillation does not occur above 7.5A when the contact lasts
longer than 1 s. Since 1980, this author and coworkers have been involved
with the study of electrical injuries and current pathways associated with
high voltages and currents (Chilbert et aI., 1983, 1985a, 1985b, 1988; Prieto
et aI., 1985; Sances et aI., 1979, 1981a, 1981b, 1981c, 1983). It was noted that
fibrillation often occurred at current levels above lOA; one study investi-
gated high-current fibrillation in hogs (Chilbert et aI., 1989).
In the high-current fibrillation studies of the author and coworkers, the
current delivery system supplied 1.5 to 200 cycles of 60-Hz currents between
1 and 51A at voltages between 500 and 6,000V. The current was applied
during the preselected cardiac cycle through a triggering circuit synchro-
nized with the cardiac waveform (Chilbert et aI., 1988) as indicted in Fig.
10.2. Nineteen fibrillations occurred in 67 runs performed on 20 animals.
Results show that 18 out of 32 short-duration runs involving the T wave
caused fibrillation; 9 fibrillations out of 14 runs occurred below lOA and 9
out of 18 occurred above lOA. As the current increases, the likelihood of
defibrillation increases (see Chapter 6). This also indicates that the likeli-
hood of fibrillation decreases with current amplitude and has been noted
by others (Daiziel, 1968, 1972; Geddes et aI., 1986). The short-duration
applications used here were always less than one complete cardiac cycle
(excluding the run of 200 cycles).
The work of Ferris et ai. (1936) has shown that, for current durations of
30ms (about 2 cycles at 60Hz) during the sensitive intervals, fibrillation can
be induced by 15-A, 60-Hz currents in sheep. They also noted that the high-
voltage, short-duration shocks do not show a cumulative effect (i.e., in-
crease fibrillation's rate of occurrence), and that normal cardiac rhythm
returned within 5 min in the nonfibrillating runs. Of 913 runs in 132 sheep,
only 100 resulted in fibrillation. They recorded 11 fibrillations out of 16 runs
at 4A, 6 fibrillations out of 17 runs at 12A, and one fibrillation out of 51
runs at 24A. This shows a decrease in the likelihood of fibrillation as the
current level increases; the occurrence at 25 A is about 10 times less than
that at 5 A. This is consistent with the hog data above. For currents of 4 to
14A applied for 30ms during the other intervals of the cardiac cycle, no
fibrillations were recorded; however, 4-A currents applied for 150ms to
260ms did cause occasional fibrillations.
Ferris tabulated body weight and heart size of sheep, hogs, dogs, and
calves. One important aspect of body weight and heart size is the ratio
between the two (Table 10.2). For similar body weights, the hog heart is
closest in size to the human heart, because the body weight and the ratio are
very similar. The thoracic circumference of hogs and humans is also compa-
rable for specimens of the same weight, although the cross-sectional ar-
rangement of organs and other tissues is different. The similarities of ratio
and circumference suggest that the hog is an excellent model for fibrillation
in humans. Also, the results of Ferris et ai. (1936) show that hogs tend to
fibrillate at lower-than-average current levels, thus making the hog a con-
servative model for ventricular fibrillation.
, CRS
T P
~ r'\ ~
~ ~
-
5 CYCLES
current
-
..
.... "'.........".
.".,,"'- 3
.,................
4
5
......'i:<l.................
......... !o.................,................
6
7
." 10
11
- 12
13
...........................""... ............,......,...
15
- - ............
19
20
.............. "'...,.....
.........., ................ 22
......."".................. 23
...
....."'i·.'..'......·
27
.- 35
- --- 40
44
..................,"'"' 47
.""'--
:00,.............., ....""'.........
51
- - fibrillation "".". non -fi b ri II atio n
FIGURE 10.2. Occurrence of fibrillation during specific periods of the cardiac cycle.
Current applications causing fibrillation all are associated with the T-wave period.
Fibrillation occurs less frequently at higher current levels. All currents involve the
cardiac cycle shown if normal cardiac rhythm is extrapolated past the onset of
current. (From Chilbert et aI., © 1989 IEEE.)
TABLE 10.2. Heartfbody ratio of animals and humans.

Average weight Ratio (%)
Species N Heart (g) Body (kg) heart/body
Human * 280 70 0.40
Hog 9 300 79 0.38
Dog 10 170 22 0.77
Sheep 25 270 56 0.48
Calf 10 420 70 0.60
Source: Data from Ferris et al. (1936); Chilbert et al. (1989).
419
10.3 Impedance Considerations and

Current Distribution in the Body
Total Body Impedance
The current that flows through the body in high-voltage and high-current
injuries is determined by the total body impedance and its change with time.
The voltage is typically constant in electrical injuries. The minimum total
impedance typically assumed for low-voltage electrical accident analysis is
500Q (Dalziel, 1978; Taylor 1985), and variations to this value are discussed
in Chapter 2. At high voltages the total impedance of the body is less in
short-duration contacts, while for longer durations the values increase with
time.
Figure 10.3 depicts the current-time relationship for electrical contact at
three low-voltage levels. The current amplitude can be separated into three
phases determined by its rate of change with time. The initial phase occurs
.....c::
~ 2
:::J
()
10 20 30 40 50
Time (5)
FIGURE 10.3. Relationship of current to time for a 5-cm disk in contact with the skin
of a hog at different voltages. Note that an increasing voltage results in a decrease
in the effective application time. Each trace shows the triphasic characteristics for
current, which are the initial phase (1), middle phase (2), and final phase (3), as
described in the text.
Impedance Considerations and Current Distribution in the Body 421
TABLE 10.3. Measured changes with voltage in the hog.

Maximum Minimum
Effective applied total Total
application current resistance energy
Volts n time (s) (rnA) (Q) (J)
75 13 205 721 104 8,773
100 14 92 806 124 6,008
150 15 21 1,087 138 2,789
200 13 11 1,227 163 2,208
250 11 5.6 1,471 170 1,734
Note: Values for a 2.5-cm-diameter disk applied to the

hindlimb with a large plate electrode on the other hindlimb.
as the contact is established (the contact impedance is rapidly reduced); this

is sometimes termed the breakdown of the electrode-skin interface. The
middle phase is the phase of maximal current flow. Increases during the
middle phase can be attributed to heating of the tissue near the electrode
contact and in cross-sections where large current densities occur. The final
phase occurs when the tissue under the electrode becomes highly resistive
because of desiccation or charring of the skin. The rate at which these three
phases occur decreases as the voltage increases (Fig. 10.3). The three phases
are present over a range of application times (Carter and Morely, 1969a,
1969b; Prieto et aI., 1985; Sances et aI., 1981a, 1981b). Application times for
five voltage levels are listed in Table 10.3 and show the decrease in time
with increasing voltage. With increasing voltage, the minimum total imped-
ance increases and the total energy decreases. The changes in impedance
and energy are caused by the rate of charring at the electrode site. The
contact phenomenon of Fig. 10.3 is applicable for low voltages that do not
support arcing.
High voltage and current are usually associated with arcing at the elec-
trode site (Sances et aI., 1979, 1981c). The initial and middle phases of the
current-time plot can occur within milliseconds; the onset of arcing is the
final phase. Voltages above 1,000V are sufficient to arc through air and
desiccated skin. Once arcing is established, the arc length and internal
impedance limit the current. Figure 10.4 shows that current falls with time
when an arc is established, indicating that the arc length increases as tissue is
burned away. The arcing appeared to occur over regions of uncharred tissue,
taking the path of minimum resistance (Sances et aI., 1981c). The tissue
removal by the arc continues until the circuit is interrupted or the limb is
transected. At 7,200V (Fig. 10.4), soft tissue was removed from the limb
within 5 s; the remaining 11 s were required to transectthe bone. At 14,400 V,
limb transection time is about 3 s. Circuit interruption can occur when the
nonconduction path length exceeds the voltage'S ability to maintain an arc
or by a ground-fault interrupter, circuit breaker, or fuse (Sect. 11.7).
25
20
~
15
C
....
0)
.... 1 0
:J
()
2 4 6 8 10 12 14 16
Time (s)
FIGURE 10.4. Current versus duration for limb transection at 7.2kV. Arrow indi-
cates point of soft tissue removal; remaining time is needed to transect the bone.
The contact was between a wire contact on the hindlimb and a large plate under the
other hindlimb. [From Sances et al. (1981b), © 1981 IEEE.]
50
40
~
-c:
....
0)
....
30
:J 20
()
220Q
234Q
10 250Q
470Q
2 4 6 8 10 12 14 16
Voltage (kV)
FIGURE 10.5. Current versus voltage for first contact in the hog. P-H is large plate
electrode-to-hindlimb wire configuration, P-F is large plate electrode-to-forelimb
wire, H-H is hindlimb-to-hindlimb wire contacts, and H-F is hindlimb-to-forelimb
wire contacts. The plate was located under the hindlimb opposite the wire contacts.
Total body impedances are shown for each measurement point and were deter-
mined from the average current for each run. [From Sances et al. (1981b), © 1981
IEEE.]
TABLE 10.4. Total body resistance (Q) for band electrodes located on the hind-
limb and analogously on the forelimb (8-cm disk electrode values d included for
comparison).
38
37/ 24.7
23 24.1
AREA (em 2 ) 30 33 34 35.4
Perimeter (em) 19.5 21.5 22.1
V
6
f-- - f..--
,......-
v- - r-- r-- t--- t----'
AA
-
1 2 3 4 5 6 KILOVOLTS
880 (d) 0.5
490 444 366 268 235 1.5

724 500 543 (d)
558 485 429 300 230 3.0
492 450 379 297 200 166 4.5

475 408 321 258
292 274 6.0
Source: Adapted from Chilbert et al. (1989).
Experimental studies of electrical burn injury have been performed in the

hog, since this animal approximates the human in skin, anatomy, weight,
and cross-sectional dimensions better than other species. The total imped-
ance of the hog at high voltages for different electrode types and locations
is given in Fig. 10.5. This figure shows that a large contact (plate electrode)
on the hindlimb decreases the total impedance and that forelimb-to-
hindlimb impedance is greater that hindlimb-to-hindlimb impedance. This
is further illustrated in Table 10.4, which shows a decreasing impedance
with increasing contact area and perimeter. As the location of the band
electrodes is moved up the limbs, the impedance decreases, which is consis-

tent with observations made previously (Prieto et al., 1985; Sances et al.,
1981b). The impedance also changes with the applied voltage level.
Impedance depends on both applied voltage and its duration (see Fig.
10.10). At low voltages applied for long time periods (those that exceed the
second phase), the impedance increases with voltage while the total energy
decreases (Fig. 1O.6a). The decrease in total energy with increasing voltage
500 'I 'I 'I
400 (a) Long duration contacts -

g
~
c: 300 - -
19
en
'w
Q) 0
0: 200 - 0 0 o
o -
]g 00 0
~ o 0
0
100 - 0 0 -
0 d
10 102 103
Applied Voltage (V)
600
500
• -
g •
Q)
u
c: 400 -
t'l •
'w
Q)
0: 300 I-
• -
]g
~
200 I- (b) Short duration contacts • •
100
102 103
Applied Voltage (V)
FiGURE 10.6. Minimum total body resistances for voltages applied between the
hindlimb and a large plate on the opposite hindquarter. (a) Voltages applied
for long time periods (through the third phase of the current-time waveform).
(b) Voltages applied for short time periods (the first phase of the current-time
waveform).
indicates less tissue heating at higher voltages (Table 10.3). For short
periods of time (t < 1 s), the impedance depends on the initial phase of the
current waveform (Fig. 10.3). The short-duration impedance at low voltages
will therefore decrease as the voltage increases, which is evident in Fig.
10.6b. High-voltage impedance changes are similar to those of low-voltage
changes. Short-duration contacts occurring in the first phase are typically
less than lOOms, and long-duration contacts would be denoted by the onset
of arcing.
Body impedance varies with voltage, duration, and contact parameters.
Increasing the voltage increases the current and decreases the total imped-
ance for short application times. The longer application times increased the
peak current at lower voltages, because the tissue resistivity decreases with
increasing temperature near the electrode sites. The electrode contact site
on the limbs determined the contact area and the perimeter for the band
electrodes (Table 10.4), and this affects the total current flow as well. In
general, increasing the area and perimeter increases the current for a given
voltage; consequently, the largest parts of the limbs had a greater contact
area and perimeter and allowed greater current passage with the band
electrodes. Also, the location of the contact on the body will affect the total
body impedance, since different current paths have different impedances.
Contact Impedance and Segmental Body Impedance

Contact impedance will limit the level of current in the body, its duration,
its path, and the probable extent of trauma. The remaining portion of the
total impedance can be divided into individual impedances for the various
segments of the body. These segmental body impedances can be modified
to account for tissue heating away from the contact points. Tissue heating
decreases impedance, thereby increasing current. By combining informa-
tion from the contact impedance and the segmental body impedances, one
can better model the effects of electrical trauma.
Experimental evaluation shows that, for circular and elliptical contacts,
the contact impedance is related to contact area and perimeter (Prieto et aI.,
1985), where the contact impedance can be expressed as
Rc = KA -1I4p-1I2 (10.5)
where Rc is the contact resistance, K is a constant of resistivity, A is the

electrode contact area, and P is the electrode contact perimeter. For a
circular electrode Rc = 0.3 Klr, where r is the electrode radius. The constant
K depends on the applied voltage, temperature, and skin condition (wet,
dry, abraded, etc.). For dry electrode contact at 100V, Prieto et al. (1985)
derived a value for K of 264 for areas between 0.3 cm2 and 20.3 cm2 and
perimeters between 2 and 28cm. The result of a least-squares linear regres-
sion fit of this equation to the experimental data gives a correlation coeffi-
cient of 0.94. As the electrode size increases in this model, the contact
resistance approaches zero, leaving only the internal impedance to limit

current.
Further analysis of the contact has been performed by using finite-
element analysis on an axisymmetric model in two dimensions. Figure 10.7
shows the boundary and arrangement for the analysis of a circular disk
containing layers for skin, fat, and muscle. The model is based on a 5-cm-
diameter disk with a potential of 100V contacting the tissue, which is a
nonhomogeneous, anisotropic, semi-infinite medium. Model parameters
are listed in Table 10.5. The model includes surface and volume heat trans-
fer effects and temperature-dependent resistivities. The skin and fat layers
have isotropic resistivities; the muscle resistivity is anisotropic with its trans-
verse resistivity defined in the z direction and its longitudinal resistivity
defined in the r direction. Thermal properties are isotropic in all tissues.
The model determines the current density and temperature distribution
in the tissue. Plots of the thermal isoclines resulting after 5s and 9s of
TABLE 10.5. Finite-element analysis parameters for a

disk electrode in contact with a nonhomogeneous,
anisotropic, semi-infinite medium used to model elec-
tric burn injury at the contact site.
Dimensions (em)
Disk radius 2.5
Skin thickness 0.15
Fat thickness 0.35
Dirichlet boundary conditions
Ambient air temperature (0C) 25
Initial tissue temperature (0C) 37
Disk potential (V) 100
Reference potential (V) o
Material properties
Electrical resistivity (Oem)
Skin 280
Fat 375
Transverse muscle 650
Longitudinal muscle 290
Thermal conductivity (Wcm-IOC- I)
Skin 0.037
Fat 0.020
Muscle 0.042
Density (gcm- 3)
Skin 1.00
Fat 0.85
Muscle 1.05
Specific heat (J g-l °e l)
Skin 3.2
Fat 2.3
Muscle 3.8
Heat transfer coefficient: 13.45 X 1O- 6 W cm -2 °C- l
sk'In
fat -
_r / contact electrode muscle I
z I
j
V
/
/
,/
/V
_v
l...--v"""
L--
FIGURE 10.7. Geometric arrangement for an axisymmetric finite-element analysis of

a disk on a three-layered, semiinfinite medium.
contact are shown in Fig. 10.8a and 1O.8b. The thermal gradient emanates
from the edge of the disk and proceeds somewhat concentrically into
deeper tissue. Surface temperatures indicate a peak value at the edge of the
disk (Fig. 10.9), showing the effect at the edge. The electrode edge effect of
increased temperature has been shown experimentally by Sances et ai.
(1981) in saline and in the hog using a thermographic camera. The total
current entering the tissue through the disk electrode is shown in Fig. 10.10.
The model compensates for changes in skin that exceed 100°C by rapidly
increasing the resistivity, resulting in a drop in the current after 9s. This
simulates desiccation of the skin comparable to experimental observations
of Prieto et ai. (1985). The analysis shows that the greatest current density
and temperature increase occurs under the edge of the disk. More than 50%
of the current entering the tissue enters in the outer 15% of the disk radius
or in the outer 28% of the total area. These results indicate the importance
of the circumference to the level of current that will flow into the body. Disk
electrodes at high voltages of similar size to those used previously at house-
hold voltage levels show similar changes due to contact area and edge
length, indicating that the results of the finite-element analysis are valid at
high voltages as well (Chilbert et aI., 1989; Prieto et aI., 1985). Theoretical
studies by Caruso and colleagues (1979) showed that current is concen-
trated at the edge of an electrode to a degree that depends on the conduc-
tivity layers beneath the electrode (see Fig. 2.9).
A 0.0 0.2Sa O.Sa 0.7Sa 1.0a 1.2Sa

skin
r
fat
8 0.0 0.2Sa O.Sa 0.7Sa 1.0a 1.2Sa

skin
r
fat
0.2Sa
muscle
FIGURE 10.8. Results of the finite-element analysis showing thermal isoclines at (a)
5 sand (b) 9 s. Isocline increment is 10 DC per contour. The disk potential was 100 V,
and the model continued until tissue temperatures exceeded 100 DC. Current density
levels under the disk are very similar to the thermal isoclines.
Internal impedance can be represented by segmental resistances of the

body. Each body segment can be modeled by either a constant resistance or
a temperature-varying resistance. Figure 10.11 shows segmental resistance
percentages for the human body based on the internal hand-to-hand values
being 100%. The calculation of segmental impedance is
Rs = Rhh( 1~0 ) (10.6)
where Rs is the segmental resistance, Rhh is the internal hand-to-hand resis-

tance, and S is the segmental percentage in Fig. 10.11 for the desired body
segment. Chapter 2 gives some values for internal resistance in humans, but
the evaluation of the contact impedance is usually quire different in those
studies. The hand-to-hand resistance is used as a reference because most
accidents involve a hand-to-hand contact. Experimental studies in the hog
have shown that the total internal resistance is 370Q, the forelimb is lS0Q,
the hindlimb is 170 Q, and the body impedance is 45 Q. These measure-
ments were made at voltages around lOV and at a constant 60-Hz current
of7mA.
110r--'1---r-1-'1---~1--r-1-'--~--~--~--~--~~
I I I I I
9 s
100 i- 8 s -
7 s
IS 90 i- 6 s -
5 s
4 s -
30~~1--~1~_~1~1~_~1~1--_~1~1~_~1~1~~1--~
10 20 30 40 50 60 70 80 90 100 110
% of Disk Radius
FIGURE 10.9. Temperature along the surface of the skin during the first lOs of
current flow. The graph approximates the transient temperature distribution for an
isopotential disk of 100V contacting a nonhomogeneous, semiinfinite medium. The
peak temperature at the disk's edge is given for each second of contact.
4.0
3.0
~
c: 2.0
......
Q)
::l
U
1.0
2 3 4 5 6 7 8 9 10
Time (min)
FIGURE 10.10. Total current flowing into the tissue from the 2.S-cm disk at 100 V for
a 10-s contact. The current falls off after 9s because the skin temperature would
exceed 100°C, when tissue desiccation would occur.
429
head
neck
biceps
/ / forearm
25
hand
···foot
FIGURE 10.11. Segmental body resistances for the human body. Numbers indicate
the percentage of the total internal hand-to-hand resistance between each of the
designated parts of the body. Figure derived from Figs. 2.39 and 2.40 and from
Biegelmeier (1985).
Tissue resistivity is inversely related to temperature. The slow rise in

current during phase 2 (Fig. 10.3) is likely the result of decreasing resistivity
with increasing temperature. Body segments in contact with the electrodes,
or otherwise subject to temperature change, can be modeled with a variable
resistance value that is temperature dependent. Each segmental body resis-
tance can be modified to change with temperature using the following
equation:
R ts = Rs(l - 0.025LlT) (10.7)
where R ts is the resistance modified by the temperature change LlT. Experi-
ment shows that increasing temperature decreases resistivity at a rate
of 2.5% per °C, which is about 5 Qcm/oC for muscle tissue (Chilbert
et aI., 1983, 1985b). At low voltages, heating occurs in the vicinity of
the contacts, if at all. High-voltage tissue heating occurs around limb joints,
at the contact sites, and in regions of small cross-sectional area. Changes
in the overall resistance will increase the total current flowing through the
body and likewise increase the tissue current density. Because tissues in a
given cross-section have different resistivities, there will be different cur-
rent densities in the different tissues; thus, specific heating in the tissues will
also be different. This leads to nonuniform heating and tissue trauma.
Tissue Current Densities and Current Distribution

The extent of thermal trauma and the likelihood of fibrillation depend on
the current density in tissue and current distribution in the body. Current
densities have been measured in the tissues of the limbs in hogs using a
constant-current source (Chilbert et aI., 1985a, 1985b, 1989; Sances et aI.,
1981a, 1983). Figure 10.12 gives the limb current densities for given applied
currents between to the hind limbs measured in the region of the hindlimb
shown in Fig. 10.13. Table 10.6 lists values of current density and resistivity
at specific applied current levels between the hindlimbs. The listed voltage
values are initial values, because the voltage drops with time when the
current is held constant. Table 10.7 gives tissue resistance, resistivity,
current, current density, and energy density for one current application of
0.9 A with an initial voltage of 400 V. This table shows how the current is
distributed through the cross-section of Fig. 10.13 before tissue trauma and
heating alters the current path. Current density in the limbs is more likely
to change at high current levels than in the body because significant heating
occurs almost exclusively in the limbs and at the contacts.
Current density measured in the body structures is given in Table 10.8,
along with the corresponding resistivities. Linear extrapolation of current
densities (as in Fig. 10.12) in the body for higher levels of current is accu-
rate, because the tissue temperature in the body changes minimally. The
largest current densities are in the back region. The back muscle at the level
of the upper lumbar region has a current density of 0.223mA/cm2 , and the
spinal cord current density is approximately 0.280mA/cm 2 for the same
E 80
<.)
~.
<t
g
.?:-
'iii
c
<l)
20
Cl
E
<l)
l:;;
::>
(.)
10
0.2 0,4 0,6 0,8 1,0 1.2 1,4 1.6 1.8 2,0
Total Applied Current (A)
FIGURE 10.12. Current density versus current in the hindlimb of the hog measured
approximately 7-cm proximal to the ankle joint; electrodes applied across the
hindlimbs. The average line represents the average current density in all the tissues
at that level.
ANTERIQR
LATERAL
skin
copper
tendon electrode
POSTERIOR
b
FIGURE 10.13. (a) Cross section of the hog hindlimb where current densities were
measured. (b) Hindlimb location of measurement region shown with isopotential
percentage lines of the applied voltage for electrodes placed on the hindlimbs.
TABLE 10.6. Tissue current density and resistivity in the limb versus applied current
from hind limb to hindlimb in the hog.
Constant Initial
applied applied
current voltage Artery Nerve Muscle
(rnA) (V) J Q J Q h Qr. QT
10 6 0.32 147 0.26 201 0.18 296 512

30 15 1.18 145 0.86 209 0.50 282 525
100 45 3.4 152 3.0 191 2.0 295 483
300 135 9.8 140 8.7 197 7.0 287 501
600 260 21.0 150 15.2 196 12.0 292 492
1,000 415 35.9 155 27.1 200 19.5 290 650
Bone Bone
Fat marrow cortex
(rnA) (V) J Q J Q J Q
10 6 0.12 375 0.10 550 0.03 1828

30 15 0.43 360 0.27 531 0.09 1880
100 45 1.5 352 1.0 535 0.30 1832
300 135 4.9 377 3.1 547 0.85 1859
600 260 9.3 366 5.3 542 2.0 1876
1,000 415 13.4 386 8.1 525 2.8 1836
Note: J in mA/cm2, Q in Qcm, h = longitudinal current density. QL = longitudinal resistivity,

QT = transverse resistivity of muscle.
Source: From Sances et al. (1983).
432
TABLE 10.7. Typical values measured in the hindlimb cross-section of the hog.
Tissue Tissue Current Tissue Tissue Energy
resistivity resistance density current area density
Tissue (Qcm) (Q) (mA/cm2) (rnA) (cm2 ) (J/cm 3 )
Vessel 155 911.76 51.61 8.77 0.17 412.90
Nerve 200 1,666.67 40.00 4.80 0.12 320.00
Muscle (L) 290 19.24 27.59 415.72 15.07 220.69
(T) 650
Fat 380 31.77 21.05 251.79 11.96 168.42
Bone marrow 550 495.50 14.55 16.15 1.11 116.36
Bone cortex 1,850 898.06 4.32 8.91 2.06 34.59
Tendon 398 196.19 20.09 40.78 2.03 160.70
Dermal layers 432 50.94 18.52 157.04 8.48 148.15
Average/total 363 8.85 22.05 903.95 41.00 176.3
Note: Applied voltage = 400 V between the hindlimbs, current = 0.9 A. Electric field strength
= 8V/cm. Resistivity and current density were measured experimentally. Cross-sectional
tissue resistance = (resistivity) (tissue area)/(length = 1 cm), tissue current = (current density)
(tissue area), tissue energy = (current density)' (resistivity) (time = 1 s). Cross-sectional areas
were determined by digitization of a limb section taken through the measurement region.
Averaged values are resistivity, current density, and energy density. Total values are resis-
tance, current, and area. Total resistance determined by adding conductances of individual
tissue elements.
TABLE 10.8. Current density and resistivity in the hog for lOOmA applied from left
forelimb to right hindlimb.
J Q
n (mA/cm2) S.D. (Qcm) S.D.
Ventral intestine 4 0.071 0.077 511 43.1
Dorsal intestine 4 0.077 0.049 610 96.9
Back muscle:
Neck 1 0.021
Upper thoracic 0.061
Middle thoracic 1 0.141 270(L)
Upper lumbar 4 0.223 0.079 1,006(T) 58.6
Lumbosacral 0.055
Kidney 4 0.097 447 61.2
Liver 4 0.065 570 37.5
Lung 4 0.058 0.033 1,605 42.8
Heart (transverse to axis) 4 0.073 0.044 800
Abdomen, midline (skin, 4 0.065 0.032 547 28.6
fat, muscle layers)
Abdomen, side (skin, 4 0.076 0.028 560 42.0
fat, muscle layers)
Note: Current density in back muscle as measured in the longitudinal direction. L = longitu-
dinal; T = transverse resistivity.
region when a current of 100mA is applied from the forelimb too the
opposite hindlimb. Various current density levels in the spinal cord are
reported in Table 10.9 for three different contact orientations. High current
levels in the spinal cord can lead to latent neural dysfunction (Sances et aI.,
1979). At a level of lOA applied for 30s, the spinal-cord temperature was
elevated only 1.5 DC, showing that temperature in the cord is not an impor-
tant factor in its trauma (Sances et aI., 1983). The cross-sectional area of the
spinal cord is 0.07% of the body cross section, but the cord carries 0.12 to
0.15% of the total current.
Current distribution in the limb changes with time and temperature and
is shown by the change in tissue current densities given in Table 10.10.
Initial values of resistivity and current density are also shown in Table 10.7
for all tissues. Current shifts from artery, nerve, tendon, and dermal layers
to muscle and fat because of increased temperature and tissue degradation.
The decrease of current in the other tissues is the result of an increase in
resistivity or a slower decrease in resistivity. Nerve and vessel resistivities
increase with temperature once tissue damage occurs, whereas muscle resis-
tivity decreases with tissue damage. The ultimate distribution of tissue
trauma is related to the peak temperature in the tissue, which is determined
from the energy density as affected by the resistivity and current distribu-
tion. However, the resistivity changes with temperature, thus altering the
rate of temperature change. For muscle, where the resistivity decreases with
temperature, the tissue damage occurs more selectively, as different parts of
the muscle become damaged at different rates and enhance the current flow
through the damaged regions. This phenomenon is the likely cause for
selective muscle groups that are completely burned being next to groups
having only minor trauma.
TABLE 10.9. Spinal cord current density for 100mA

applied across the limbs.
Transverse Longitudinal
J (mA/cm') J (mA/cm')
Forelimb-forelimb
Cervical 0.0155 0.0247
Thoracic 0.0022 0.0058
Lumbar 0.0011
Forelimb-hindlimb
Cervical 0.0075 0.0708
Thoracic 0.0100 0.299
Lumbar 0.0080 0.257
Hindlimb-hindlimb
Cervical
Thoracic 0.0008 0.0020
Lumbar 0.0057 0.0040
Note: Transverse spinal cord resistivity = 1,970 Qcm, longitu-

dinal resistivity = 214Qcm.
Thermal Trauma 435
TABLE 10.10. Simultaneous current density and temperatures 7 cm proximal to

distal tibia versus time with I-A current applied from hindlimb to hindlimb.
Applied
Time voltage Artery Nerve Muscle
(min) (V) J (rnA/crn2) T (0C) J (rnA/crn2) T (0C) J (rnA/crn2) T (0C)
0 355 30.0 39.0 24.3 38.5 13.3 39.0

2 293 28.7 41.0 22.8 41.5 12.5 40.5
4 265 27.1 43.0 21.9 44.5 12.5 42.5
6 253 26.5 46.0 21.0 47.5 12.2 44.5
8 235 25.5 47.0 20.5 49.5 12.2 46.5
10 220 25.0 48.5 20.0 52.0 12.2 47.5
Fat Bone marrow Bone cortex

(min) (V) J (rnA/crn2) T (0C) J (rnA/crn2) T (0C) J (rnA/crn2) T(°C)
0 355 9.5 38.5 3.0 39.0 1.8 39.0
2 293 10.5 42.0 3.1 40.0 1.9 40.0
4 265 12.8 45.5 3.1 42.5 2.0 42.5
6 253 14.4 49.0 3.2 44.0 2.0 44.0
8 235 14.8 52.0 3.1 46.0 1.9 46.0
10 220 15.0 54.0 3.1 47.0 1.9 47.0
10.4 Thermal Trauma

Thermal burn injury is the most common trauma that occurs at high current
and high voltage levels, although electroporation may also cause tissue
damage. In this section the resultant electrical trauma will be referred to in
terms of thermal burns. This section will discuss the anatomical and
physiological aspects of electrical burns, which occur at the contact site, in
the deeper tissues, around the joints, and in the spinal cord. Thermal
trauma depends on the current density in the tissue more than the resistiv-
ity, as indicated by Eq. (10.1).
Heating at the Contact Site

Most burns are at or near the point of contact. When the injury current is
high, then trauma can also be seen farther away from the contact site. The
discussion here will relate the transient temperature to the severity of burn
and the effects of blood flow on the extent of injury. The superficial burn is
characterized by three regions of tissue destruction and charring at the
electrode contact edge (Fig. 10.14). The first region of tissue destruction,
which is adjacent to the electrode contact, contains desiccated and dena-
tured tissue, indicated by erupted blisters, shrinkage of the tissue, and
charring at the edge of the contact. The second region is the ischemic
region, where the tissue is devoid of blood flow but retains its fluid content.
The outermost region, represented by a darkened ring caused by hemor-
Hemorrhage and
Thrombosis - - -
Ischemia - - -
Dessicated
and Blistered
Charring at the
Contact Edge
FIGURE 10.14. Superficial burn resulting from a circular contact at 100V. There are
three distinct regions of trauma: a desiccated region adjacent to the electrode having
erupted blisters, an ischemic region that maintains its fluid content, and a hemor-
rhagic region darkened by thrombosis. Also shown is a ring of charring that occurs
at the electrode edge.
rhage and thrombosis of the microvasculature, appears several minutes

after current is stopped. These burn regions extend into deeper layers of the
dermis as well (Fig. 10.15) and extend into underlying muscle tissue when
the current is sufficiently high. Desiccation is deeper at the electrode edge
than in the center. Histological evaluation of the dermis shows that the
trauma from electrical burns is very similar in character to nonelectrical
thermal damage. The extent of trauma can be determined by the transient
temperature response of the tissue.
charring at the
/ contact edge ~
hemorrhage
and thrombosis
FIGURE 10.15. Dermal cross section beneath a circular contact at 100V. The
resultant injury is a full-thickness dermal burn. The three regions of Fig. 10.14 are
present. Desiccated tissue extends deeper into the tissue at the electrode edge than
at the center.
Thermal Trauma 437
Henriques (1947) and Henriques and Moritz (1947) have investigated the
thermal tissue response in terms of the time-temperature relationship. As
seen in Fig. 11.12, the time necessary for trauma decreases exponentially
with peak temperature. Skin will burn at 45°C after an hour but at 60°C
only a few seconds are required. Experimentally measured temperatures
are given in Fig. 10.16 for a circular electrode of 5-cm diameter having a
100-V potential and applied through the third phase of the current-time
waveform (Fig. 10.3). The measured temperatures corresponds well with
the finite-element model results of Fig. 10.8. In general, the peak tempera-
ture delay from the time of peak current increased primarily with distance
from the edge of the electrode (Fig. 10.17). A secondary effect increasing
the time delay of peak temperature is the depth from the skin surface, and
regions without blood flow (ischemic regions of Figs. 10.14 and 10.15) also
had longer delays. The secondary effects are the result of heat radiation and
convection at the surface and to slower heat conduction in ischemic tissue.
Tissue perfusion by blood is an important factor in the generation of
electrical burns. Comparing the above in-vitro studies to studies in situ, the
results show significant variations in the total applied energy, contact time,
and rate of tissue cooling. The parameters of voltage, electrode size,
average power, and minimum total resistance were the same. The total
energy applied was 24 % less in situ than in vivo, the contact time was 21 %
less in situ that in vivo, and the rate of tissue cooling was 50 to 75% less in
situ than in vivo. This factor is important to the evolution of electrical burn
trauma, since the data indicate that blood flow helps lessen the severity of
the trauma in short-duration contacts.
electrode
107+ +73 54+
96+ + +
57 43 _
0.5 +
+ 94 9+0 +42
84 77
E
~ 1.0 ;'7 +67 -
a.
L:.
+
0
Q)
1.5 75 -
2.0 I-
+
54
I I I I
2 4 6 8
Radial Distance (cm)
FIGURE 10.16. Peak tissue temperatures measured experimentally beneath a 5-cm-

diameter circular contact at 100V in five hogs. The voltage was applied until the
current ceased from desiccation of the contact site (phase 3 of the current-time
waveform). (Adapted from data of Prieto et aI., 1985.)
110
100
90
0
U
80
~
::l
"§ 70
(J)
0.
E 60
(J)
I-
50
40
30
2 3 4 5 6 7 8 9 10
Time (min)
FIGURE 10.17. Plots of tissue temperatures beneath the electrode at different depths
(cm) with time, 100V contact. Peak temperatures do not occur simultaneously, but
at times related to depth below the electrode edge following the current falloff of
phase 3, as indicated by arrows. Broken line indicates current cessation.
Heating of the Tissues

When currents are high and/or application times are long, trauma extends
beyond the contact site, requiring surgical intervention and possibly ampu-
tation. The trauma affects primarily the limbs and is typically located in
muscle tissue. Peripheral nerves and blood vessels are also affected,
because they are usually adjacent to affected muscle. Bone is usually the
least affected, but adjacent muscle is affected more than distant muscle.
Deep tissue injury depends critically on the current pathway. As noted by
Eq. (10.3), the distribution of current is based on the electric field strength
in a given cross-section and the tissue resistivity.
Heating of tissue and subsequent resistivity changes have been investi-
gated in hogs and dogs (Sances et aI., 1983; Chilbert et aI., 1985b). Table
10.10 shows that current density in fat increases with temperature but
decreases in artery, muscle, and nerve. Tendon and dermal current density
decreases with temperature. The current path will shift due to changes in
the resistivity. This change is the result of a direct temperature dependence
of resistivity and to tissue degradation. Tissue degradation causes increased
resistivity in nerves and vessels while decreasing muscle resistivity. Neural
tissue increases 80 to 100% in resistivity when thermally injured, while
muscle decreases as much as 80%. The changes in muscle resistivity become
Thermal Trauma 439
more important when one considers that a large change in impedance could
be indicative of bum severity.
Studies performed by the author (Chilbert et aI., 1985a 1985b) indicate
that tissue temperature, resistivity, and the severity of trauma can be corre-
lated in muscle tissue. Figure 10.18 shows the measurement sites in the dog
gracilis muscle for temperature and resistivity. Electric bum trauma was
induced by passing 1-A currents at 60Hz until the distal measurement site
reached a temperature of 60°C. Peak temperatures along the muscle de-
creased proximally as the limb cross-section increased (Fig. 10.19). Regions
1 and 2 exceeded 50°C, while the remaining regions remained below 47°C.
Severe trauma was noted in regions 1 and 2, while regions 6 and 7 had
edema with minimal structural changes. Regions 3, 4, and 5 were in a
transition zone between viable and nonviable tissue. The resistivity in-
creased proximally, but all values were less than control values (Fig. 10.20).
The initial resistivities at the onset of the I-A current increased 35%
because of muscular contraction; they then declined with increasing tem-
perature. The resistivities continued to decline even as the tissue cooled,
indicating further progression of tissue injury.
Thermal trauma to peripheral nerves is also related to the maximum
temperature reached. The temperature in a nerve depends on the tempera-
ture of the surrounding tissue. The peak temperature of nerve in Table
10.10 shows that it is only 2°C less than the tissue in which it was measured
(i.e., fat). Neural activity (indicated by averaged evoked response)
gracilis
.....+--muscle
current
·<4f---electrode
FIGURE 10.18. Measurement sites in the gracilis muscle of the hindlimb of a dog.
Temperature and resistivity were measured at each site. Site 1 is the most distal and
site 7 is the most proximal.
60
t current off
0
G" 50
~
::J
iii
Q;
c..
E
Q)
40
I-
30
15 30 45 60
Time (min)
FIGURE 10.19. Temperature increase caused by the flow of a I-A current with time.
Current was applied for 15 min. Temperatures at regions 1 and 2 were elevated above
50°C, while the other regions were between 40 and 46 °C and are grouped together.
decreases with temperature as neurons are damaged. Degradation of the

peripheral nerve response with a thermal injury (not electrically induced),
caused by slowly increasing the temperature, shows the development of
trauma (Fig. 10.21). Important features of the evoked response indicate
that certain neurons are more sensitive to increased temperature, noted by
a larger decrease at 45°C and at 50°C. Which neurons are affected at these
temperature levels in not presently known. For rapid heating of the nerve
the temperature of trauma is above 50°C, where recovery of the evoked
response is usually rapid once the temperature reaches that of the body.
However, heating of the nerve with electric current causes permanent
damage at lower temperatures. The evoked response is always immediately
affected by current, at levels above 3.5 mAlcm2 along the nerve, causing
blocking of the response from 5 to 20 min. Once temperatures exceed 45°C,
permanent reductions in the evoked response may occur, and above 50°C
the response never returns. This suggests that the current may affect
neurons directly, exclusive of temperature, possibly by electroporation.
Limited studies in the spinal cord are consistent with peripheral nerve
information, but the temperature threshold for electrically induced trauma
is lower.
Thermal trauma to bone is seldom seen except at very high voltages
(Sances et aI., 1981b). The high resistivity of cortical bone prevents signifi-
Thermal Trauma 441
cant direct heating and protects the marrow as well. Table 10.10 shows that
bone has the lowest temperature in cross-section. Muscle tissue near the
bone is often burned, while muscle tissue away from the bone is not
affected. Some clinicians have erroneously attributed this phenomenon
to heating of the bone, but it is the result of higher current densities in
the muscle near the bone (Sances et ai., 1983) because heating is more
dependent on current than resistance.
In cross-section, the joints have a large quantity of bone compared to
other more conductive tissues. This creates very high current densities in
the more conductive tissue, consequently causing severe trauma at the joint
while leaving tissues by the long bones viable (Sances et ai., 1983; Zeit et ai.,
1988). Also, the joint capsules have been known to disrupt explosively from
the buildup of steam (Sances et aI., 1981b, 1983). However, if the victim
were to flex the joint during electrical contact, the current will have an
alternate path through the skin surfaces above and below the joint, thus
preserving the joint. This has been noted clinically around the elbow, where
burn marks appear on the skin proximally and distally.
140
120
~ 100
> 7
.~
en 80 6
(I)
c: 5
~
0 60 4
3
40 2
20
5 10 15 20 25 30
Time (min)
FIGURE 10.20. Changes in resistivity measured during and following the application
of a I-A current. Resistivities of muscle were measured in the transverse direction
and perpendicular to the direction of current flow and are given in percent of control
values. Initial increase of resistivity is due to the contraction of the muscle by
stimulation from the current. Initial resistivity is for contracted muscle at time =
omin. Muscle resistivity continues to decrease with decreasing temperature after
current is turned off at t = 15min in region 1.
100
90
80 T1lK1
Ql
"t:l
70 LV~
.~
a. 60
E
ex:
ec 50
0
<.)
40
a
~
0
30
20
10
a
35 40 45 50 55 60
Temperature (0C)
FIGURE 10.21. Changes in the evoked response of the tibial nerve in the hog with
temperature. The reductions in the peak amplitudes were permanent. The first
component of the evoked response is denoted by A, and the second is denoted by
B. Numbers on curves indicate time in minutes after the temperature increase was
started. Amplitude is given in percent of normal evoked response amplitude.
10.5 Nonthermal Trauma

Lesions in tissue not caused thermally are likely caused by electroporation
(Lee et aI., 1988; Weaver, 1992; Gaylor et aI., 1992). Typically, the electric
field strength needed to rupture muscle cells or cause nerve damage must
exceed 100V/cm. Also, it has been noted that elevated temperature
increases the likelihood of electroporation. Studies in the peripheral nerve
of hogs show that no significant alterations in the evoked response occur
at current densities up to 167mA/cm2. This current density level results in
an electric field strength of 33.2V/cm, which is below the minimum level of
100V/cm cited by Gaylor et aI. (1989). To attain such field strengths in the
limb, the applied voltage would have to exceed 5,000V. For spinal cord
injury to occur, the minimum applied voltage would need to be between
35kV and 50kV for a hand-to-foot path. These values were predicted from
measurements in the hog given in Tables 10.4 and 10.9 and in Fig. 10.3.
Lightning Injuries 443
From Table 10.9 one notes that a maximum current density of 0.3 mA/cm2
results when a current of 100mA is applied from the forelimb to the
hindlimb and the spinal cord has a resistivity of 214 Qcm. Since the electric
field strength should be above 100V/cm, the current density in the cord
needed to cause irreversible electroporation would have to be 500mA/cm2 •
To obtain this current density in the spinal cord, an applied current of
167 A between forelimb and hindlimb is needed. To estimate the voltage
level for 167 A, refer to Table 10.4: At electrode location 4, nearly 20A will
flow at 6kV. A simple extrapolation of the voltage-current ratio (6 kV-20 A)
shows that 50kV is needed to generate 167 A in the body. However, resis-
tance decreases with increasing voltage, therefore using the values of7.2kV
and 25A from Fig. 10.3 and extrapolating to the 167-A level, the applied
voltage needed is 35kV. The value for humans may be lower, since reports
in the clinical literature have cited spinal cord dysfunction at 25kV
(Kanitkar and Roberts, 1988).
10.6 Lightning Injuries

Lightning accidents occur infrequently relative to other electrical incidents,
but often enough to be well documented. According to Biegelmeier (1986),
one-third of all lightning injuries are fatal. The immense power of lightning
is shown by the effects of its millions of volts and 5,000 to 200,000 A (Sances
et aI., 1979; Strasser et aI., 1977). Reviews of lightning injury have been
given by Taussig (1968) and Silversides (1964). Persons struck by lightning
who were clinically dead have been revived, and maladies seen afterward
eventually disappear in most cases (Taussig, 1968).
The afflictions commonly encountered with lightning are mostly neuro-
logic. Burns can occur at the entry points and may be accompanied by
internal lesions. The burns often have a splashlike or arborlike appearance,
assumed to be caused by current tracking over the body (Artz, 1979).
Skeletal fractures occur when the person falls or is thrown by muscular
contractions. Effects on the nervous system can occur immediately, can be
delayed for several hours, or can occur days after.
Immediate effects of lightning strike depend on the severity of the
incident. The victims usually undergo a total arrest of all body functions,
which slowly return over a given length of time (Taussig, 1968). Fatalities
can occur when the victim has a long respiratory and cardiac recovery
period and no assistance is administered such as CPR, (Strasser et al.,
1977; Taussig, 1968). Abdominal lesions are also notably fatal, especially
when internal hemorrhage is present (Artz, 1979). For the less serious
accidents, the heartbeat is present with respiration intact or soon returning,
leaving the victim in an unconscious state. Upon arousal, the victim may act
disoriented and may have paresis of some or all the limbs (Critchley,
1934; Silversides, 1964). Other, less common deficits are visual, auditory,
and speech (Critchley, 1934). Often amnesia occurs with lightning insults
(Strasser et aI., 1977; Taussig, 1968). Hypertension is normally observed,
with anxiety and, at times, neurotic behavior. These symptoms usually
vanish within a week (Strasser et aI., 1977).
Secondary effects appear within the first few days after the accident.
Except for an occasional change in the electrocardiogram or trauma to
internal organs, all secondary effects are neurologically oriented, consisting
of paralysis (usually of the legs), muscle pain throughout the body,
photophobia from the intense light, and various autonomic disturbances.
These effects normally vanish within a week. Latent effects (seen after 5
days) usually are resolved within a month after the accident. Neurologic
symptoms are thought to be associated mainly with the progression of
vascular disease or dysfunction. Neurologic symptoms in this stage are also
common to high-voltage power-line injuries, which will be discussed later
(Silversides, 1964).
Soft tissue lesions are found more frequently with lightning accidents than
in high-voltage accidents (Massello, 1988; Sharma and Smith, 1978). Neuro-
logic lesions often occur where there is an interface between regions of
different resistivities. In lightning accidents one commonly sees splitting of
the cortical layers of the brain and petechial and subarachnoid hemorrhage
(Critchley, 1934; Silversides, 1964). Several explanations of the lesions
have been suggested, but all lack supportive evidence. One hypothesis
suggests that the lesions are caused by the heating effects, which produce
pockets of gas or steam. This is not generally accepted as a good explanation,
because the heat generated is probably insufficient to cause boiling of the
tissue. Fluid electrolysis has also been suggested; however, the time involved
and current necessary may exclude this as a possibility. The most acceptable
theory is electroporation causing cellular lysis at or near the tissue inter-
faces. The lesions have been seen more often at tissue interfaces, where
charge accumulation can occur (Critchley, 1934; Strasser et aI., 1977).
10.7 Clinical Observations

Of all electrical injuries, most occur to those working for electric utility
companies. In one report, 95% worked for utility companies, of which 50%
or more were linemen (Butler and Gant, 1978). The voltage levels for most
of these injuries exceed 1,000V and are at the standard commercial fre-
quency. Reviews of the general aspects of electrical injury have been pre-
sented by many (Dixon, 1983; Sances, 1979; Skoog, 1970; Wu, 1979; Lee et
aI., 1992, 1994). Clinical reviews have shown common observations among
patients (Haberal, 1986; Hammond and Ward, 1988; Luce et aI., 1978;
Sharma and Smith, 1978; Solem et aI., 1977). Some authors have focused
reviews on the neurologic sequelae noted in electric accidents (Critchley,
1934; Silversides, 1964; Strasser et aI., 1977). Others have reviewed the
Clinical Observations 445
aspects of electric bums (Artz, 1974; Butler and Gant, 1978; Rosenberg and
Nelson, 1988; Rouse and Dimick, 1978).
The various forms of trauma seen with electrical injury are the same as
those seen in thermal bums with the addition of others unique to electrical
injury. Some aspects of electrical injuries are similar to crush injuries.
Clinical reports and reviews present observations in the general areas of
bums, lesions, neurologic effects, and cardiovascular effects. The reports
and reviews usually focus on one of the areas, while the patients may
present symptoms from all areas.
Electrical Burns
The pattern of electrical bums can vary greatly from those seen in thermal
injuries (Artz, 1979). Electrical bums follow the current path, whereas
thermal bums start at the surface from one location and radiate into the
surrounding tissue (Janzekovic, 1975; Luce et aI., 1978; Ponten et aI., 1970;
Sances et aI., 1979). Deep tissue necrosis is often evident. Although tissue
may appear viable, it may be damaged and require secondary procedures.
The transformation of visibly viable tissue to necrotic tissue has been
termed progressive necrosis, but experimental studies indicate that the
visible viable tissue is actually necrotic from onset of the trauma (ZeIt et aI.,
1988). Artz (1979) has reviewed and presented the various aspects of elec-
trical injury, especially those associated with electric bums. Also, reference
is made to internal lesions and renal sequelae. The following is a summary
of bum wound observations:
Contact-site wounds usually signify deep tissue destruction locally. The
contact sites can be defined in terms of the primary site, which is the contact
with the energized source, and the secondary site, which is the contact with
the ground or neutral source. The primary wound is usually charred and
depressed. The secondary wound is dry, depressed, and has the appearance
of the current exploding outward. Massive swelling is evidence of extensive
tissue damage caused by heat and electroporation.
Skin injury varies from small circular spots to large areas of charring.
Adjacent to the charred tissue is a whitish-yellow ischemic region. Sur-
rounding the ischemic skin is an area darkened by vascular hemorrhage
and thrombosis. All three regions are relatively cold and without sensory
perception.
Vessel damage may extend beyond the general area of injury. Thrombosis
has been noted away from the bum injury. Typically, extreme vascular
spasms, thrombosis, and necrosis of vessel walls are observed. In non-
thrombotic, damaged vascularization, hemorrhaging may result and lead
to serious complications.
Muscular trauma is caused by the direct heating of current or by occlusion
of the arterioles supplying the muscle. Damage to the muscle is usually
uneven, affecting groups of fibers while not affecting adjacent areas.
Uneven damage is characteristic away from the contact sites and may not
be noticed initially (ZeIt et aI., 1988).
Burns caused by arcing imply high temperatures at the contact site and
extensive deep tissue destruction. Traumatic limb amputation has been
documented (Sances et aI., 1979). Cursory burns are caused when arcing
ignites the victim's clothing.
Renal failure is more common in electrical injuries than in thermal burns.
Renal damage is caused by direct electric involvement of the kidneys and/
or renal vessels, or more often by the abnormal breakdown of protein from
other injured tissues. Devitalized muscle in electrical injury causes renal
complications similar to those seen in severe crush injury (Artz, 1974).
There is also a greater occurrence of hemoglobinurea and hematinuria in
the electrically injured patient. One critical complication is acute tubular
necrosis, probably caused by myoglobin breakdown.
Oral burns are most often seen in children and cause severe burns to the
lips, tongue, and dentition. This form of electrical burn has been reported
extensively particularly in recent years, because of the special treatment
needed (Barker and Chiaviello, 1989; Donly and Nowak, 1988; Palin et aI.,
1987; Sandove et aI., 1988; Silverglade, 1983). Of particular concern is tissue
contraction of the lips and cheeks, which is prevented by special splints
(Sandove et aI., 1988).
Most pediatric electrical injuries admitted to the hospital occur to the
mouth (35-60%) and most victims of oral burns are under the age of 4
(Baker and Chiaviello, 1989; Palin et aI., 1987). Oral burns account for
almost all severe burns. Electrical trauma in children usually results from
defective equipment. Nearly all injuries occurring to children are prevent-
able, especially with routine equipment maintenance.
Tissue Lesions
Tissue lesions are evident at the tissue interfaces and are commonly associ-
ated with neural dysfunction and abdominal complications. Autopsy upon
electrical accident victims often reveals submucosal hemorrhages dispersed
throughout the gastrointestinal tract. Abdominal lesions are associated
with a high mortality rate. They are difficult to detect and treat. Delayed
fatality is usually associated with abdominal lesions of the gastrointestinal
tract (Artz, 1979). Typically the patient is comatose, and vital signs degrade
over several days until death occurs.
Lesions have the most significance in neural tissues. Once a nerve has
been severed, its function seldom returns. Lesions may also be responsible
for latent neural dysfunctions that appear up to 3 years after the injury
(Sances et aI., 1979; Silversides, 1964). When lesions are noted in the spinal
cord, they are seldom complete transections. Clinical signs of these lesions
are spastic paresis with little or no sensory deficit. The lesions usually are
not suspected until the onset of patient ambulation (Baxter, 1970). These
afflictions, like many associated with the nervous system, are not well
understood.
Neurologic Sequelae
Neurologic sequelae are manifest at all levels of the nervous system and are
either permanent or transient. Permanent effects are the result of thermal
trauma or lesions, while transient disorders can disappear within days or
last for months (Hooshmand et aI., 1989).
Damage to peripheral nerve can be caused by electroporation (lesions)
or excessive heating of the tissue. The ulnar, radial, and femoral nerves
have the highest incidence of injury due to heating. Occasionally lesions are
formed in the peripheral nerves, causing sensory or motor deficits. Some
cases have been reported as progressive with time, the victim slowly losing
sensation in the limbs and other clinical abnormalities developing
(Kinnunen et aI., 1988). Injury to the peripheral nerves, as stated above, is
permanent, and recovery is minimal. Peripheral nerve damage is usually
associated with adjacent tissue injury.
Transient disorders often seen are neurovascular, neuromuscular, and
sensory. The neurovascular disorders include vascular constriction and
spasmodic reactions, both of which reduce blood flow and vascular dilation
(Christiansen et aI., 1980; Hooshmand et aI., 1989). Vascular spasms cause
a reduction of blood flow in an area, usually associated with tingling, numb-
ness, and, at times, paralysis. Vasodilation is believed to be the cause of
fainting spells sometimes encountered after electrical injuries. Neuro-
muscular disorders are paresis, paralysis, hypertension, and muscular pain.
Current having a magnitude great enough to affect the spinal cord or
nerve roots has a transthoracic path, most commonly hand to hand, or hand
to foot. Head-to-extremity paths, although not very common, will also
affect the spinal cord. With hand-to-hand contact, current is greatest in the
cervical region (Table 10.9). Hand-to-opposite-foot current paths mostly
affect the heart and thoracic spinal cord. Head-to-extremity current paths
affect the upper brain centers and cervical spinal cord (Butler and Gant,
1978; Sances et aI., 1979; Solem et aI., 1977).
Permanent damage to the spinal cord may be caused by vertebral dis-
placement secondary to falling, by muscular contractions, or by lesions in
the cord or nerve roots. Skeletal fractures in electrical injury are normally
attributed to falling or being thrown by muscular contraction. Lesions in the
spinal cord account for many of the permanent disabilities of motor func-
tion. Lesions are seen mainly in cases of power-line contact above 11 ,000 V
(Butler and Gant, 1978). The magnitude of the lesions in the spinal cord
is not readily apparent until the onset of ambulation (Baxter, 1970; Solem
et aI., 1977).
Transient effects of current on the spinal cord appear at all stages of
recovery: immediate, secondary, or latent. The immediate effects are loss of
consciousness, respiratory arrest, and neural effects on the heart rate.

Unconsciousness may cease immediately after current flow has stopped or
linger for a day or more (Silversides, 1964). Respiratory arrest occurs from
the sustained contraction of intercostal muscles and interruption of the
neural respiratory centers. Sustained contractions of the intercostal muscles
last only while the current is applied. Spontaneous breathing has been
delayed for several hours in some victims, but usually returns within 20min.
Secondary effects of electric injury involving the spinal cord are temporary
paralysis ofthe limbs, vascular spasms, and muscle pains (Silversides, 1964).
Immediate effects of electrical accidents involving the head may include
convulsions, coma, cerebral edema, hysteria, amnesia, tinnitus, deafness,
and visual disorders. Convulsions are induced in electrical injury from the
passage of current through the cerebral cortex, analogous to those induced
in electroconvulsion therapy. Coma is observed in patients who have been
in cardiopulmonary arrest for relatively long periods of time. Clinical signs
of electric coma are dilated, unresponsive pupils, along with no reflex
response. Cerebral edema is seen only in the severest cases (very high
voltages and lightning), typified by a softening of the tissue (Critchley,
1934). Hysteria is mentioned throughout the literature; its severity is
minimal. Specific hysteric symptoms are agitation, confusion, amnesia, and
transitory auditory and visual dysfunction (Critchley, 1934; Silversides,
1964). Amnesia is the most common symptom. Often the cranial nerves are
directly affected; tinnitus and deafness are observed in victims, particularly
when the victim does not lose consciousness. Permanent deficits in cranial
nerves are associated with lesions along the cranial nerves.
Visual disorders include blindness, blurred vision, photophobia, and
cataracts (Silversides, 1964). Blindness is caused as a result of current
effects on the optic nerve or interference with the photoreceptors in the
retina (AI-Rabiaed et aI., 1987). Permanent blindness is caused by lesions
along the optic tract or separation of the retina from the choroid. Blurred
vision results when the innervation of the lens musculature and iris have
been affected. This condition is typified by dilated pupils. Photophobia
occurs when the retina is exposed to intense light from an arc or lightning
flash. Cataracts have been known to develop after lightning strikes and very
high-voltage contacts with the head (Moriarty and Char, 1987).
Secondary conditions associated with the cerebrum involve the prolong-
ing and delayed development of the immediate symptoms discussed above.
Secondary effects are mainly hysterical complications. In addition to those
stated above are speech loss, disorientation, and narcolepsy (Critchley,
1934; Silversides, 1964).
Latent effects includes psychosis, hemiplegia, aphasia, epilepsy,
choreoathetosis, and other hysterical conditions (Critchley, 1934; Silver-
sides, 1964). Psychosis is rare and may be related to preaccident manifesta-
tions. Hemiplegia and aphasia (not as common) are seen in conjunction
with cerebrovascular disturbances. Prior existence of vascular disease cor-
relates well with hemiplegia and aphasia resulting from electrical accidents.
Posthemiplegic Parkinson's disease has been known to develop, either the
unilateral or the asymmetric bilateral form (Silversides, 1964). The initial
convulsion in a developed epileptic condition may have started during the
current flowing in the body. If the current path includes the head, the
formation of an epileptic focus is likely. Actual lesions are rare, leaving a
diffuse focus as the cause. Victims who developed epilepsy respond well to
drug therapy, which usually is withdrawn over a period of time.
Cardiovascular Effects
Electrical accidents have also produced latent cardiovascular disorders that
are usually transient. Some of these affects may be associated with fibrilla-
tion and defibrillation of the victim (Wilson et aI., 1988) or with severe
burns. Because these effects are usually less life-threatening than the
complications of burn, they are seldom reported. However, the occurrence
of cardiovascular complications is frequent (Guinard et aI., 1987). High
current levels flowing through the chest can cause cardiac arrest, which is
then followed by a somewhat normal cardiac rhythm. Most cardiovascular
effects follow high-voltage electrical injuries. Latent effects include atrial
fibrillation, premature ventricular contractions, bradycardia, tachicardia,
ventrical and atrial ectopic foci, conduction branch block, and nonspecific
S-T interval and T-wave changes (Butler and Gant, 1977; Jones et aI., 1983;
Skoog, 1970; Solem et aI., 1977).
The causes of latent cardiac effects and delayed arrhythmias is unknown.
Some suggested causes include the residual effects of fibrillation, enzymal
reactions with the cardiac tissue, neural dysfunction, or coronary artery
spasms. The fibrillation of the heart or a sustained myocardial contraction
by the current may alter the metabolic activity of the heart and result in
altered cardiac rhythm (Solem et aI., 1977). Certain enzymes are released
from necrotic muscle tissue, such as various forms of creatine kinase, which
can alter the function of the cardiac cells or mitochondria and again alter
the cardiac rhythm (Ahrenedolz et aI., 1988). Because various forms of
transient neural anomalies have been noted in other areas of the nervous
system, neural alterations in the inervation of the heart may also occur
(Skoog, 1970). Similarly, the vascular spasms noted in the periphery may
also occur in the coronary arteries and lead to altered cardiac rhythm (Luce
and Gottlieb, 1984; Skoog, 1970). Fortunately, these disturbances are
usually transient and disappear within a week (Jones et aI., 1983; Luce and
Gottlieb, 1984).
Latent vascular effects, such as spasms or peripheral coldness, have
been attributed to neural disturbances (Kinnunen et aI., 1988). Similarly,
transient neural disturbances have been attributed to vascular spasms
(Silversides, 1974; Skoog, 1970). Some latent effects are attributed to loss of
endothelial cells and elasticity in the vessel wall, which leads to vascular
impairment (Wang and Zoh, 1983). Also, the vasculature may be impaired
by arachidonic acid metabolites, such as thromboxane, that cause vasocon-
striction. The latent effects are often a result of the initial trauma, but
appear as healing occurs.
10.8 Clinical Treatment

Treatment of electrical injury begins at the scene of the accident with
removal of the victim from the electrical source and the observations of
vital signs. A lack of pulse or respiration requires immediate cardio-
pulmonary resuscitation. One should note specific details of the accident,
such as the contact voltage, whether the victim fell, if there are fractures,
and if there are burns evident at the contact sites. Rapid transportation of
the victim to an adequate trauma center is essential (Dixon, 1983).
The potential for considerable fluid loss coincident with deep tissue
damage is great, therefore fluid replacement must be started as soon as
possible. Intravenous isotonic fluids are administered to maintain the urine
output at approximately 100mllh. Traditional burn formulas cannot be used
to calculate fluid replacement requirements. These formulas underestimate
the required fluid volume because they are based only on the skin surface
area involved; fluid replacement principles for a crush injury are more
appropriate (Artz, 1979; Dixon, 1983). If myoglobinuria is present, sodium
bicarbonate is administered to alkalinize the urine, which inhibits pigment
precipitation. Also, an osmotic diuretic, usually mannitol, can be used to
flush the renal tubules (Baxter, 1970; Dixon, 1983, Hunt et aI., 1980;
Moncrief and Pruitt, 1971).
The electrocardiogram should be monitored, and echocardiography may
be useful in assessing cardiac damage. Monitoring the hemoglobin and
hematocrit will indicate the degree of hemolysis, and arterial blood-gas
measurements are used to determine the acid-base state of the victim
(Artz, 1974; Dixon, 1983).
The external appearance of the electrical burn wound is not indicative of
the extent of underlying tissue damage. The electrical burn may be masked
by flame or flash burns (Moncrief and Pruitt, 1971). Initially, the wounds are
thoroughly cleansed. If the initial examination reveals a charred limb, an
absent or diminished pulse in a limb, a loss of sensory or motor nerve
function in a limb, or evidence of considerable limb swelling, then an
escharotomy and fasciotomy should be performed. Fasciotomies have both
a therapeutic and diagnostic function, as they may restore circulation and
permit the surgeon to inspect the underlying musculature visually (Baxter,
1970; Wang 1985). Amputation of at least one extremity is frequently
required following high-voltage injuries. If a limb is clearly necrotic, then
amputation should be performed immediately (Parshley et aI., 1985).
Amputations are generally done in the first week postinjury, after the
Clinical Treatment 451
patient's condition has stabilized. Arteriograms may be helpful in determin-

ing the initial level of amputation (Hunt, 1979).
Although there is general agreement that obviously necrotic tissue
should be debrided as soon as possible, there is still controversy over the
treatment of tissue that is questionably viable (Barnard and Bostwick, 1976;
Baxter, 1970; Rouse and Dimick, 1978). One approach involves early explo-
ration and debridement, with repeated surgical procedures after 24 to 72h;
until all nonviable tissue has been identified and removed (Artz, 1974; Hunt
et aI., 1980; Luce and Gottlieb, 1984; Parshley et aI., 1985). A more conser-
vative approach may be used, where the debridement procedure is delayed
until a definite delineation between viable and nonviable tissue becomes
evident, usually after 8 to 10 days. Proponents of the aggressive approach
claim earlier wound closure and less incidence of infection, while propo-
nents of the conservative approach feel that unnecessary multiple surgical
procedures are avoided, the local blood supply is not altered by surgical
trauma, and viable tissue is less likely to be removed. ZeIt et aI. (1988)
showed experimentally that all tissue injury is histologically evident at the
onset of injury; progressive necrosis does not occur.
Several methods have been advanced to determine the viability of ques-
tionable tissue so that excision of the necrotic tissue can be conducted as
early as possible. Arteriography has been used to assess vascular damage,
but this method does not consistently visualize the smaller arterial and
arteriolar vessels in the muscles (Hunt et aI., 1980). Thus, only occlusion
of the larger vessels can be reliably determined with arteriography. The
technetium-99 m pyrophosphate scan has been used to identify the extent
of cardiac and skeletal muscle damage, including deep areas of muscle
necrosis that may be missed by visual inspection during a fasciotomy
(Hunt, 1979). Other studies have suggested that electrical impedance or
nuclear magnetic resonance spectroscopy can be useful in determining
tissue necrosis (Chilbert et aI., 1985a, 1985b).
Tissue resistivities were compared with histologic changes as well as
results of 31PNMR spectroscopy. Abnormalities in anatomic as well as
metabolic measures were shown to correspond well to resistivity changes.
Measurements of impedance at several frequencies were made to compute
the phase-plane plot of muscle tissue (Fig. 10.22). Impedance is unique for
each tissue as well as its condition, reflecting traumatic injury, viability,
edema, and so on (Ackmann and Seitz, 1984). Likewise, changes in the
phase angle are indicative of alterations in tissue, but are more proportion-
ately related to the number of cells present. The relaxation frequency
occurs at the point of maximum reactance. These boundary changes
(caused by cell lysis) alter the relaxation frequency; loss of cells increases
the relaxation frequency. With electrical burn injury, the amount of cell
lysis is coincident with the temperature rise in the tissue. Figure 10.20 shows
how muscle resistivity changes with temperature and has been correlated to
trauma (Chilbert et aI., 1985b). The level of trauma, as indicated by
20 40 60 80 100 120
Resistance (0)
FIGURE 10.22. Impedance and relaxation frequency changes seen in muscle tissue
with electric burn. The proximal tissue impedance curve corresponds to tissue that
is only mildly edematous, the transition tissue impedance curve corresponds to
tissue showing greater edema and some loss of cell structure, and the distal tissue
impedance curve corresponds to tissue that shows gross destruction of cell structure
and extensive damage.
31PNMR spectroscopy and histology, shows a 75% decrease in resistivity

with severely burned tissue, whereas tissue only slightly affected by the
electricity showed a decrease of 25% or less. Using impedance techniques
to determine tissue viability allows separation of compartmental changes
(edema) and boundary changes (cell lysis) so that tissue vitality is better
determined. Figure 10.22 indicates both the impedance change and relax-
ation frequency change seen in electric burns. Once the probability of tissue
survival is determined for a given change in impedance and relaxation
frequency, clinical measurements can be performed and appropriate exci-
sion of nonviable tissue can be made. This will eventually lead to reduced
morbidity in the electric burn patient by minimizing surgical procedures
and chances for infection.
The bacterial population of the wound must be controlled to prevent
infection, especially clostridial myositis, which may develop as a result
of inadequate excision of necrotic tissue (Artz, 1979; Hunt et aI., 1980).
Topical antibacterial agents are used rather than systemic agents, because
the blood flow is often compromised in the region of the wound. Tem-
porary wound closure can be achieved with porcine xenografts, autografts,
or homografts, whereas permanent wound closure is usually achieved with
an autograft. The use of a free muscle flap, utilizing the latissimus dorsi,
Clinical Treatment 453
for immediate coverage of deep electrical injury has been advocated.

However, if the debridement prior to grafting is inadequate, then muscle
necrosis or infection may be masked by the vascularized muscle flap.
A viable cutaneous homograft may be a useful tool to test the readiness
of a wound for surgical closure or definitive autografting. Vein grafts,
using autotransplantation of the saphenous vein to replace the radial and
ulnar arteries, have been used to restore circulation to the hand and prevent
limb necrosis (Wang et aI., 1984, 1985).
Electrical injuries involving only the scalp may be managed with the same
procedures used for soft-tissue electrical burns of other parts of the body
(Worthen, 1982). Contrary to earlier procedures for treating an electrical
burn of the scalp and skull, a full-thickness scalp flap should be used to
cover the exposed skull as soon as possible, regardless of the depth of
damage to the skull. The vascularized cover promotes skull regeneration,
as ingrowth and revascularization occur on both the pericranial and
meaningeal surfaces (Worthen, 1982). Alternatively, split-thickness skin
grafts can be applied to the skull after the cortical bone is removed to
expose cancellous bone, although later resurfacing with rotational scalp
flaps is required to restore the hair. A graft consisting of autogenous greater
omentum covered with autogenous split skin has been used to cover a large
irregular area of exposed skull following a high-voltage electrical injury.
Extensive destruction of the face requires a multistage reconstructive
procedure, which may extend over a period of several years.
Advances in the understanding of cell damage because of electroporation
have suggested possible therapies for reversal of cell damage. It has been
observed that following electroporation of cells in vitro, membrane repair
can be promoted by the application of surfactant compounds within a
couple of hours of the injury (Lee et aI., 1994). Whether this would be
a practical therapy in vivo remains to be demonstrated.
11
Standards and Rationale
11.1 Introduction
The previous chapters provide an account of electrical forces that can exert
measurable influences on biological systems. Although the bioelectric
mechanisms treated in this book can be used in a controlled fashion to exert
a beneficial effect, the same forces, when presented in an uncontrolled
fashion, can sometimes be detrimental. For that reason, various agencies
have developed protective standards and guidelines.
There are many such agencies whose purposes are often directed to
specialized environments, applications, or user groups. This chapter re-
views standards and guidelines concerning electromagnetic radiation expo-
sure to the general populace, to individuals in occupational settings, and to
patients undergoing magnetic resonance imaging. We also review standards
and protective measures developed to protect individuals from electric
shock in consumer products. The rationale for these standards and related
bioelectric mechanisms will be discussed.
11.2 Electromagnetic Field Exposure Standards

Standards a/the IEEE and ANSI
In the United States, the most widely applied standards for exposure to
electromagnetic fields above 3 kHz are those of the Institute of Electrical
and Electronic Engineers (IEEE), which were adopted by the American
National Standards Institute (ANSI) in 1992. ANSI is a private, nonprofit
organization that does not write specifications. Rather, this organization
provides an organizational structure for the development, review, and pub-
lication of standards, such as the standards on electromagnetic radiation
known as IEEE/ANSI Standard C9S.1 (IEEE, 1992). ANSI standards are
considered advisory, as the agency has itself no enforcement or regulatory
powers. Nevertheless, many of the ANSI standards, and the IEEE C9S.1
454

Electromagnetic Field Exposure Standards 455
electromagnetic standard in particular, are adopted as official regulations

by many authorities. The technical work and development of the IEEE
electromagnetic standards are carried out by the IEEE Standards Coordi-
nating Committee-28 on Non-Ionizing Radiation (SCC-28). The current
standard was issued by the IEEE in 1991 and adopted by ANSI in 1992
(IEEE, 1992) as a revision of a previous standard that existed in 1982
(ANSI, 1982). The 1982 standard covered the frequency range from
300kHz to 100GHz; the 1991 standard covered the frequency coverage
from 3 kHz to 300 GHz. At the time of its publication, the membership of
the committee responsible for the 1991 standard consisted of more than 125
members from research (72%), industry (10%), consultants for industry
(3%), government administration (4%), and independent consultants and
the general public (11 %). The disciplines of the members included the
physical and biophysical sciences (33%), life sciences (43%), medicine
(13%), and others, such as law, medical history, safety, and so on (11 %)
(see Petersen, 1995).
Rationale for IEEEIANSI C95.1 Standards

The standards before 1982 limited incident power density to 10mW/cm2
across the applicable frequency band based on the assumption that biologi-
cal effects were related to a thermal mechanism in the exposed organism
(Petersen, 1991). Better dosimetric information was available for the 1982
standards, and it was determined that the data on biological effects corre-
lated better with the specific absorption rate (SAR), than with the incident
power density. The SAR metric is stated in watts per kilogram of tissue
(see Sect. 11.5).
The IEEE committee considered the SAR metric superior because the
absorbed power averaged over the whole body for a given incident power
density was found to vary with frequency, as seen in Fig. 11.1 (Durney et aI.,
1985). This figure shows the incident power density consistent with a SAR
value of 0.4 W/kg averaged over the whole body. A vertically polarized
(electric field vector orientation) propagating field is assumed, which is the
worst case polarization with respect to absorbed energy. The figure shows
that the incident power density at a constant value of SAR displays a
resonance at frequencies above 3 MHz, and the frequency at which the
resonance effect is greatest depends on the size of the subject. The figure
shows resonance curves for human body weights ranging from 10 to 70kg-
a range that would include small children and adults.
After applying pre screening criteria, the database used in the 1982 ANSI
standard consisted of 321 papers; an additional 60 papers were included in
the 1991 standard (IEEE, 1992). These papers could be classified into the
following categories: environmental factors, behavior and physiology, im-
munology, teratology, central nervous system, cataracts, genetics, human
studies, thermoregulation, biorhythms, endocrinology, development,
456 11. Standards and Rationale
ANSI (Sep 82)

200
ht = 1.75 m, wt = 70 kg
6' 100 ht = 1.38 m, wt = 32 kg
E
~ 50 ht = 0.74 m, wt = 10 kg
~ Man on ground plane
.s 20
z.
'w
c 10
Q)
0 5
(j)
;: 3
0
Il..
Q)
Ol
~ 0.5
Q)
J(
0.2
0.1
10 10 2
Frequency (MHz)
FIGURE 11.1. Power densities that limit human whole-body SAR to 0.4 W/kg
compared to ANSI standard of 1982. (From Durney et al. 1985.)
evoked auditory response, hematology, and cardiovascular effects. The

category associated with the lowest thresholds of a possible detrimental
effect was that of behavioral disruption of complex tasks (in rats) corre-
sponding to an SAR value of 4 W/kg. Whether observed laboratory effects
were the result of a thermal mechanism or some other mechanism was not
determined in the C9S.1 rationale. In terms of the human metabolism,
4 W/kg is expended by moderate activity such as housecleaning or
driving a truck, and falls well within the normal range of thermal regula-
tion. The committee applied a "safety factor" of ten, which brought the
whole-body SAR limit to 0.4 W/kg.
The solid boundary in Fig. 11.1 is drawn to encompass the family of
curves for various body sizes. By so limiting the allowable incident power
density, one ensures that a SAR value averaged over the body does not
exceed 0.4 WIkg in any individual. This boundary was chosen as the basis
for exposure limits in the 1982 IEEEI ANSI radiation standards. The expo-
sure at frequencies below 3 MHz is limited to an incident power density of
100mW/cm2, whereas the absorption curves in Fig. 11.1 indicate that the
corresponding field rises without limit as the frequency falls below 3 MHz.
A cap was placed on the incident power density in order to limit the
possibility of induced shock, or spark discharges in high intensity fields (see
Sect. 11.4).
We can relate electric and magnetic field intensities to the incident power
density using the relationships for a propagating electromagnetic wave:
z = Ii = 3770. (11.1)
H
S = E2 = H2Z (11.2)
Z
where E is the electric field (Vim), H is the magnetic field (Aim), S is the
power density (W/m2), and Z is the impedance of the propagation medium
(0.). In "free space" (a vacuum), Z = 3770. which is the assumption used in
the radiation standards; the impedance of air is essentially the same as that
of free space. The relationships expressed in Eqs. (11.1) and (11.2) apply to
a propagating wave in the "far field" many wavelengths distant from the
source of energy.
In many applications, standards are quoted with respect to the magnetic
flux density, B (units: Tesla, T), rather than magnetic field intensity, H
(units: Amperes per meter, Nm). These two quantities are related by
B = fiH (11.3)
where fi is the permeability of the medium. In free space or air, fi = 4n X
1O- 7 H/m, and we have the relationship that lNm corresponds to 1.257 fiT.
Electric and Magnetic Field Criteria for

IEEEIANSI C95.1
Figure 11.2 illustrates the limits in C95.1 over the frequency range 3kHz to
lOGHz. Separate limits are shown for controlled and uncontrolled environ-
ments which comprise a so-called "two-tier" standard. Although not shown
in the figure, the standards extend the high-frequency cap of lOmW/cm2 to
300GHz for both controlled and uncontrolled environments. Controlled
environments are locations where people are aware of the potential for
exposure as a result of employment, by other cognizant persons, or where
exposure is incidental as a result of transient passage through areas in which
the exposure may be between the controlled and uncontrolled environment
limits. Uncontrolled environments are locations where individuals have no
knowledge or control of their exposure, such as in living quarters or work
places where there is no expectation of exposure.
Before 1983, exposure standards in the Unted States did not distinguish
between exposure of the general public and those in occupational or con-
trolled environments (Osepchuk, 1994). In 1983, the Massachusetts Public
Health Department adopted a two-tier policy in which exposures to the
public were one-fifth of the levels stated in the then-existing ANSI C95.1 of
1982. The averaging time was 30 minutes for the lower tier and 6 minutes
for the upper tier. Consequently, the two tiers are equivalent for exposure
614 Vim 10mW/cm2

\
163 Aim
E
~
"0
Qj 10
±
0 H-field
E 2
~ 10 mW/cm
"0
Qj controlled environment \
~ uncontrolled environment
,
,
(fJ \
::!: \
a: \
\
.1 \
0.073 Aim
.01 .1 10 100 1000 10000

Frequency (MHz)
FIGURE 11.2. Maximum permissible exposure to electric and magnetic fields accord-
ing to IEEE C95.1 (1991).
durations of 6 minutes or less. The lower limit was intended to protect the
general public who might have chronic exposure when living near EMF
sources. The factor one-fifth was derived from the presumed ratio of maxi-
mum occupational weekly exposure (40hr/wk) to that for the chronically
exposed public (168hr/wk). The two-tier Massachusetts limits were essen-
tially adopted by the National Council on Radiation Protection and
Measurements (NCRP, 1986).
One criterion applied to the 1991 standard was the resonance curve for
SAR = 0.4 (Fig. 11.1). In conformance with existing NCRP practice, the
IEEE committee included an additional safety factor in the minimum reso-
nance region for uncontrolled environments, resulting in a minimum SAR
of 0.08W/kg. Based on studies of absorption of electromagnetic energy in
humans, it was concluded that the electric field component coupled energy
into the biological medium more efficiently than did the magnetic field
(Gandhi, 1988). Consequently, the electric field limits were selected accord-
ing to Eq. (11.2), with incident power density limited by the solid line of Fig.
11.1; the magnetic field limits were selected to permit greater exposure at
frequencies below 100 MHz, in recognition of the less efficient coupling of
the magnetic field in this frequency regime. For instance, in a controlled
environment, the E- and H-fields in the minimum plateau region are consis-
tent with an incident power density of 1 m WIcm2(= 10 W1m2), namely E =
61.4 Vim and H = 0.163 Aim, as calculated by Eq. (11.2). In an uncontrolled
environment, the allowed power density in the resonance region is O.2mWI
cm2 • At frequencies above lSGHz, the standards have a cap of lOmW/cm2
for both controlled and uncontrolled environments. By observing both the
electric and magnetic field limits, one ensures that an average SAR value of
0.4 W/kg will not be exceeded at any frequency, whether the exposure is in
the near or far field of the source. Note that the SAR limits contained in
C9S.1 apply only at frequencies between 0.1 MHz and 6GHz.
Although average whole-body SAR limits may be so limited, individual
regions within the body will exhibit absorption rates above and below
the average rates. Implicit in these limits is the assumption that peak
SAR within limited regions of the body may exceed the average. Accord-
ingly, IEEE C9S.1 provides that SAR limits may be relaxed in any 1 g
of tissue to 8 W Ikg, except for the hands, wrists, and feet, where the
SAR limit is limited to 20W/kg as averaged over any 109 of tissue in
the controlled environment and one-fifth of that value in uncontrolled
environments.
The averaging time applicable to IEEE C9S.1 is 6 minutes for frequencies
below 15 GHz; at higher frequencies, the averaging time goes down to a low
of 10 seconds in controlled environments. For pulsed fields, IEEE C9S.1
specifies additional criteria which would limit the peak exposure (see Sect.
11.3). Averaging time rules are specified in uncontrolled environments for
frequencies in the resonance region (30-100MHz) in which the averaging
time is 6min. at and below 30MHz, and 30min. at 100MHz and above.
Considerations involved in averaging time are addressed by Osepchuk
(1989).
Induced Current and Electric Shock Considerations

Based solely on absorbed power, one would conclude that permissible fields
can increase without limit in inverse proportion to frequency below the
resonance region. However, in order to avoid problems of induced electric
shock at low frequencies, IEEE C9S.1 imposes additional constraints on the
maximum permissible fields, and also on induced current density within the
body as discussed below. Thermal perception of electric currents is dis-
cussed in Sect. 7.5.
As seen in Fig. 11.2, the permissible E-field is capped at 614 Vim below
3 MHz. That cap was imposed in recognition of the possibility of spark
discharges to exposed individuals. Note that C9S.1 standards do not pre-
clude the possibility of spark discharges under some conditions (see Sect.
11.4). The magnetic field below 100kHz is capped at H = 163 Aim, which is
equivalent to B = 0.20SmT. Compare that value with Fig. 9.17, which shows
magnetic thresholds calculated for whole body exposure of a large person.
TABLE 11.1. Maximum induced current limits from

IEEE/ANSI C95.1.
0.003-0.1 MHz 0.1-100 MHz
Controlled environment
Both feet 2,000/ 200
Each foot 1,000/ 100
Contact 1,000/ 100
Uncontrolled environment
Both feet 900/ 90
Each foot 450/ 45
Contact 450/ 45
Limits are stated in mA-rms, with/in MHz. Averaging time =

1 s. Foot limits assume a free-standing individual (no metallic
contacts). Contact limits are for touching a metallic object.
Even if we use the worst case assumption of the parameter Ie in Fig. 9.17, we
see that the C95.1 magnetic field limit is well below the threshold for
peripheral nerve excitation.
The C95.1 standard also imposes limitations on induced currents as noted
in Table 11.1, which were intended to limit the severity of induced electric
shock or burns. The intent of the Standards Committee was to limit the
possibility of perceptible contact current with a grasping contact. However,
these limits would allow perceptible current with a touch contact under
some conditions. Consider, for instance, the limits on contact current.
At a frequency of 10kHz, the standard would allow a contact current of
lOrnA in the controlled environment, and 4.5mA in the uncontrolled
environment. At a frequency of 100kHz, the standards would allow 100mA
in the controlled environment and 45 rnA in the uncontrolled environment.
Comparing these values with the reaction thresholds in Fig. 7.12, we con-
clude that the standards below 100kHz would allow perceptible shock
in the uncontrolled environment and painful shock in the controlled envi-
ronment, particularly for small individuals, if subjects were to access the
current through a touch contact with an energized conductor. Above
100kHz, the standards would permit sensations of heating that could be
perceptible with a touch contact in the uncontrolled environment and pain-
ful in the controlled environment. For small subjects or children, even
stronger sensations might occur across the range of frequencies shown in
Table 11.1 (see Sect. 7.10). The perceived reactions would be reduced
relative to a touch contact if the subject were to contact the energized
conductor with a large area, such with by grasping with the whole hand
(refer to Fig. 7.15, and Chatterjee et aI., 1986), as was assumed in the C95.1
standard.
Other Standards on Electromagnetic Exposure

Another group with national stature in the United States is the National
Council on Radiation Protection (NCRP). This group was chartered by
Congress in 1928 to develop information, and issue exposure guidelines on
all forms of radiation. NCRP issued guidelines on radio frequency electro-
magnetic fields in 1986 (NCRP, 1986). The standing committee members
and consultants on the radio frequency standards consisted of 16 experts at
the time. NCRP adopted the then-existing ANSI standards of 1982 for
occupational settings and applied an additional acceptability factor of 5 for
the general public. Subsequently, IEEE/ANSI C95.1 issued in 1991 echoed
the NCRP two-tier philosophy with "controlled" and "uncontrolled" envi-
ronment limits and applied the additional safety factor of 5 to the uncon-
trolled environment. The NCRP reasoned that the lower limits should be
applied to the general public who may consist of particularly vulnerable
individuals and others who may be unaware of, or have no option about
their exposure. The additional safety factor of 5 was arrived at by rounding
the ratio of a 40-hour work week to the total number of hours in a week
(40/168 = 0.2). The guidelines of the NCRP are more conservative than
those of the IEEE/ANSI at the higher frequencies. Above 1,500 MHz,
NCRP limits top off at a power density of 0.2 and 5mW/cm2 for general
and occupational popUlations respectively. The IEEE/ANSI limits reach a
ceiling of lOmW/cm2 for both populations.
Electromagnetic exposure standards and guidelines have been published
by other world-wide agencies covering the general public, occupational,
and military populations. Of particular interest are the standards of the
National Radiation Protection Board (NRPB) in the United Kingdom
(NRPB, 1993; McKinlay et aI., 1993). These have been recently published,
their rationale is well documented, and the standards cover a wide fre-
quency range-from zero to 300GHz. With respect to static fields, the
NRPB recommendations are designed to prevent acute direct effects, such
as vertigo and nausea. Restrictions on ELF electric and magnetic fields are
intended to avoid acute effects on the central nervous system. Restrictions
on radio-frequency and microwave fields were intended to avoid adverse
responses as a result of increased heat load and elevated tissue temperature.
The NRPB specifies "basic restrictions" as listed in Table 11.2-these are
the fundamental dosimetric quantities of limitation in the standards. For
frequencies from 1 to 10GHz, the basic restrictions specify current density
and SAR limits within the human body. From 10 to 300GHz, the restric-
tions specify incident power density, becuase the depth of penetration of
energy at these high frequencies was assumed to be small. SAR limits are
specified with respect to the volume of tissue for averaging.
From the basic restrictions, the NRPB derives "investigation levels,"
which are environmental measureables for investigating whether compli-
TABLE 11.2. Basic restrictions on exposure to electromagnetic fields as specified by

the NRPB.
Frequency range
(Hz) Basic restriction Averaging area Comments
0-1 200mT 24-h average
2T maximum, body
5T maximum, limbs
100mNm2 Current density
1-10 100/jmNm2 Current density
10-Hi' 10mNm2 Current density
1<Y-105 j/100mA/m2 Current density
105_107 O.4W/kg,lOg body SAR, 15 min avg.
lOW/kg 109 SAR, 6min. avg., head & fetus
lOW/kg 100g SAR, 6min. avg., neck & trunk
20W/kg 100g SAR, 6min. avg., limbs
f7100mNm 2 Current density, head, neck, trunk
107_1010 O.4W/kg body SAR, 15 min. avg.
lOW/kg 109 SAR, 6min. avg., head & fetus
lOW/kg 100g SAR, 6min. avg., neck & trunk
20W/kg 100g SAR, limbs
1010_3 X 1011 100W/m2 Power density, max. on body
Note: j specified in Hz.

Source: McKinlay et al. (1993).
ance with the basic restrictions is achieved. If the measured quantity is

below the investigation level, one assumes that compliance with the basic
restrictions is achieved. If the measured quantity is above the investigation
level, it does not necessarily imply that the basic restrictions are exceeded.
According to the NRPB, factors that might be considered in such an assess-
ment include the efficiency of the coupling of the person to the field, the
spatial distribution of the field across the volume of space occupied by the
person, and the duration of exposure (NRPB, 1993). The NRPB investiga-
tion levels are similar to the IEEE/ANSI C9S.1 standards (Fig. 11.2) in that
they recognize the body resonance phenomenon, and the associated SAR
and current density values are in a similar range.
An international agency also specifying EMF guidelines is the Interna-
tional Commission on Non-Ionizing Radiation Protection (ICNIRP), which
was estalished in 1992 as a separate entity under the International Radia-
tion Protection Association (IRPA). This commission, which has as many
as 18 representatives from 11 countries (depending on year of commission),
publishes EMF guidelines in association with the World Health Association
(WHO). Separate publications cover static fields (ICNIRP, 1994), 100 kHz
to 300GHz (IRPA, 1988), and SO/60Hz (IRPA, 1990). The most recent
ICNIRP pUblication (1998), which was published as this book was in press,
covers the frequency range from OHz to 300GHz.
General reviews of international standards have been provided by

Gandhi (1990), Polk and Postow (1996), and Klauenberg and colleagues
(1995). More recent standards have been published by the Netherlands
(1997). In general these standards recognize the resonance phenomenon
discussed above and provide for minimum exposure at frequencies within
the resonance region. The minimum fields allowed within the resonance
region typically are within a factor of two of the IEEEIANSI standard. The
standards differ much more, however, at frequencies below 1 MHz. For
instance, in the frequency range 10 to 100kHz, the E-field limits (in Vim)
for continuous exposure by the general public, in the order greatest to least
are: 1,500 (Germany), 1,000 (NRPB, United Kingdom); 614 (IEEE, United
States); 614 (IRPA); 280 (Canada); 87 (Netherlands); 25 (Soviet Union).
The maximum magnetic field limits (in Aim) for the same frequency range
are: 350 (Germany); 163 (IEEE, United States); 64 (NRPB, United King-
dom); 64 (IRPA); 4 (Netherlands); 1.8 (Canada). An equally wide disparity
exists among other standards at frequencies below 10kHz, as noted in the
next section.
Electromagnetic Exposure Limits at Low Frequencies

Much public discussion has occurred in past years concerning electromag-
netic exposure at very low frequencies and particularly at power frequen-
cies (60Hz in the North America, and typically 50Hz elsewhere). This
attention has been largely driven by concerns about public health issues
associated with power transmission and distribution lines-a subject that
remains controversial. There are presently no IEEE standards for frequen-
cies below 3kHz, although the IEEE C95.1 committee is presently consid-
ering standards in that frequency regime. However, other standards do
specify exposure limits at very low frequencies, and these will be reviewed
here. The reader is also directed to Sect. 11.7 for an analysis of established
bioelectric mechanisms as a basis for low frequency standards.
Figures 11.3 and 11.4 illustrate very low frequency magnetic and electric
field limits that are specified by national and international agencies. One
agency is the American Conference of Governmental Industrial Hygienists
in the United States (ACGIH, 1996), which is a private professional group.
The limits specified by ACGIH are intended to apply to professionals in
industrial and occupational settings, and not necessarily the general public.
The limits are intended to protect against hazards in the workplace. The
ACGIH states that biological effects for E-fields below the specified limits
have been demonstrated in the laboratory, but these reactions were not
thought to be hazardous. In the case of the B-field limits, an exposure to the
arms and legs could be increased by a factor of 5 in the frequency regime 1
to 300Hz; for the hands, a factor of 10 could be applied. A second agency
is the European Agency for Electrotechnical Standards (CENELEC, 1995).
The CENELEC standard shown in Figs. 11.3 and 11.4 is considered a
1000
(a) CENELEC (workers)
(b) CENELEC (public)
(c) NRPS
100
(d) ACGIH
(e) Mechanisms-based
10
c
.1
:
~..:.,.~ ..~...~ ..~ .. ..~.
a
.01
.1 10 100 1000 10000
Frequency (Hz)
FIGURE 11.3. B-field limits at very low frequencies: whole body exposure, extended
durations (e.g., whole work day). Curves (a)-(d): existing standards. Curve (e):
hypothetical mechanisms-based standard for short-term exposure in uncontrolled
environments.
"prestandard," subject to further review. These limits are intended to pre-

vent adverse short-term effects in humans caused by current densities that
exceed lOmA/m2-a criterion derived from laboratory demonstration of
magnetophosphenes, possible nervous system effects, and bone healing
effects. A third agency represented in the figures is the National Radiologi-
cal Protection Board in the United Kingdom (NRPB, 1993; McKinlay et aI.,
1993). The stated NRPB rationale of the low-frequency standards is to limit
induced current density in the head, neck, and trunk as indicated in Table
11.2 so as to avoid effects on the central nervous system functions, such as
control of movement, posture, memory, reasoning, and visual processes.
The most recent guidelines of ICNIRP (1998) cover the frequency range
OHz to 3000Hz.
The magnetic field limits shown in Fig. 11.3 are all below the threshold for
phosphenes, which at the most sensitive frequency (20Hz) is 10mT (Fig.
9.20); the limits are also well below thresholds for excitable tissue stimula-
tion (Fig. 9.18). Thus, one would not expect acute reactions for exposures
constrained by the indicated B-field standards.
Figure 11.3 includes curve (e), which is a hypothetical standard for short-
term exposure in uncontrolled environments. Curve (e) is based on estab-
lished mechanisms for reactions to short-term magnetic field exposure as
developed in Sect. 11.7.
The E-field limits shown in Fig. 11.4, however, are sufficiently great that
they would not preclude significant reactions from induced shock, spark
discharges, and corona. For instance, with fields in the range 20 to 30kV/m,
corona discharges could easily be produced from body surfaces, and severe
spark discharges could easily occur with ordinary activity in which a person
touches grounded objects (see Sects. 9.2-9.4). It is difficult to see how one
could function in such a high field environment without taking special
precautions to avoid these effects. In recognition of the difficulties in high
E-field environments, ACGIH recommends that workers avoid situations
where spark discharges may be produced in fields above 5kV/m and recom-
mends the use of protective clothing in fields above 15kV/m.
Considering the significant public debate about health effects from power
frequency fields, it is useful to focus on EMF limits at 50/60Hz, which are
listed in Table 11.3. Table 11.3 also includes limits of the International
Commission on Non-ionizing Radiation Protection (ICNIRP, 1998). In the
a
.....................................................................................-..
c, d ,,
,
..........................................
,,
b ' \,
10
,,
,,
(a) CENELEC (workers) ,,
(b) CENELEC (public) ,,
(c) NRPB ,,
(d) ACGIH
.. -
\~ ~':'
\ ........ ~.... .
.1
.1 10 100 1000 10000
Frequency (Hz)
FIGURE 11.4. E-field limits at very low frequencies: whole body exposure, extended
durations (e.g., whole work day).
TABLE 11.3. Other standards evaluated at power frequency.

E-field (kV/m) B-field (mT)
Agency 50Hz 60Hz 50Hz 60Hz
ACGIH (1996)
whole body 25 25 1.2 1
arms & legs 25 25 6 5
hands & feet 25 25 12 10
ICNIRP(c) (1998)
workers 10 8.33 0.5 0.42
general public 5 4.17 0.1 0.083
NRPB (1993) 12 10 1.6 1.33
CENELEC (1995)
workers, whole body(b) 30 25 1.6 1.33
workers, limbs 30 25 25 20.8
gen. public, whole body 10 10 0.64 0.53
gen. public, limbs 10 10 10 8.33
Notes: (a) Limits are stated as rms quantities; (b) exposure duration (hr) :s 80/E.
(c) Exceedance of ICNIRP limits requires analysis to test for compliance with current
density restrictions.
frequency range below 1 kHz, the ICNIRP magnetic field limits are
intended to provide an adequate safety margin below an induced current
density of 100 mNm2 , which was taken as a threshold for acute changes
in central nervous system excitability or visual evoked potentials. Safety
factors of 10 and 50 were provided for occupational groups and the
general public, respectively, thus providing basic restrictions of 10 and
2mNm2.
Additional power frequency guidelines have been developed for applica-
tion to electric power transmission and distribution lines, as shown in Table
11.4 (after Goellner et aI., 1993). Note that the B-field units in Table 11.4
are in.uT, rather than in mT, as in Table 11.3. While there are no applicable
federal standards in the United States, a number of states and local regula-
tory bodies have specified limits of exposure to the general public from
power lines. Table 11.4 also lists power line environmental limits adopted in
other countries. The power frequency magnetic field limits of state and local
agencies in the United States are well below corresponding limits in other
countries, and well below the international limits listed in Table 11.3. One
should distinguish between an environmental limit and a standard. An
exposure standard generally specifies the area of the body being exposed,
the allowed level, the duration, as well as other exposure conditions. An
environmental limit or a product performance standard would generally
not specify such conditions, although they might include considerations
other than human exposure (e.g., product interference).
Additional product performance guidelines on low frequency EMF emis-

sions have been specified by the Swedish federal government for applica-
tion to video display devices, such as computer monitors, and television sets
(Mild and Sandstrom, 1994; MWN, 1995). Existing Swedish guidelines in
the frequency range 5Hz to 2kHz specify an E-field of 25V/m measured at
a distance of 50cm in front of the display, and a B-field of 250nT measured
50cm all around the display. In the frequency range 2 to 400kHz, the
corresponding values are 2.5V/m and 25nT. Proposed guidelines specify
field values in categories, and the manufacturer would be required to label
the equipment as to its category of compliance (MWN, 1995). According to
Mild and Sandstrom, the guidelines were developed without clear proof of
health risks, but as a "prudent avoidance" strategy that was felt to be
readily achieved by manufacturers.
The exposure limit discussed above encompass an enormous range. For
instance, at 60Hz, the various guidelines and standards specify E-field limits
ranging from lOV/m to 30kV/m, and B-field limits ranging from O.2,uT to
21 mT. Although part of this range can be attributed to distinctions between
standards, performance limits, and environmental limits, the wide range
TABLE 11.4. Power frequency exposure standards for transmission and distribution
lines.
E-field (kV/m) B-field (uT)
Agency In ROW Edge ROW In ROW Edge ROW
US State/local
California (Irvine) 0.4
Fiorida(a)
:s;230kV 8 3 15
500kV, single 10 2 20
500 kV, double 10 2 25
Minnesota 8
Montana(b) 1
New Jersey 3
New York PSC 1.6 20
North Dakota 9
Oregon 9 0.4
Tennessee 0.4
Outside US(d)
Australia(e)
occupational 2 500
general public 10 2 100
Germany 5000
Italy
extended exposure 5 100
limited exposure 10 1000
Notes: (a) More stringent standards apply to Lake Tarpon line; (b) 2:69kV; (c) limits are
stated as rms quantities; (d) 50 Hz; and (e) whole day exposure;
Source: Goellner et al. (1993).
also reflects the controversy and lack of agreement concerning chronic-

exposure EMF health effects, especially at frequencies below a few kHz.
Most EMF standards recognize constraints imposed by acute human reac-
tions to EMF exposure, which have been treated in the previous chapters of
this book. The mechanisms responsible for these acute effects are widely
recognized, reasonably well understood, and experimentally verified. On
the other hand, most standards do not acknowledge hazards for exposures
below the limits where acute reactions are known to exist; exceptions to this
statement apparently apply to some transmission line standards, and the
Swedish VDT emission standards, which limit emissions to very low values.
The mechanisms that might be responsible for these presumed weak field
effects are not well-established. Although a thorough treatment of weak-
field effects is beyond the intended scope of this book, we will treat this
subject in brief in Sect. 11.6.
Static Electric and Magnetic Fields

Strong static magnetic fields can cause biological reactions through their
influence on moving charges, as described in Sect. 9.11. Examples would
include blood flow, and the movement of the biological system itself, such
as with eye or body movement. Table 11.5 summarizes limits on static fields
as recommended by several agencies. The limits in the upper part of the
table apply to extended exposure in occupational groups, or by the general
public. ICNIRP guidelines for static magnetic field exposure state that
detrimental effects do not occur below 2 T according to current knowledge.
The ICNIRP exposure guidelines in Table 11.5 reflect a safety factor of
10 for occupational exposure, and 50 for the general public. For the
CENELEC occupational limits, the duration of exposure is specified as t :s
1121E, where t is the duration (hours) in any 8-hour period, and E is the E-
field in kV/m. For instance, at the upper limit of 42kV/m, an exposure
duration of 2.67 hour per 8-hour period would be allowed. The NRPB
reduces limits by a factor of 10 for chronic (24hr) exposure because of the
lack of chronic human exposure data, rather than specific adverse findings
related to chronic exposure. The lower part of Table 11.5 applies to patients
undergoing magnetic resonance imaging. It is assumed that for those indi-
viduals, the duration of exposure would be relatively brief, and not often
repeated.
11.3 Pulsed Electromagnetic Fields

It is possible to stimulate peripheral nerves and even the heart with a
sufficiently strong time-varying magnetic field. For instance, we have seen
that peripheral nerves can be excited with a magnetic stimulus of about
40T/s oriented normal to the sagittal or frontal cross section of a large
Pulsed Electromagnetic Fields 469
TABLE 11.5. Static field criteria.

Max. Max
E-field B-field
Agency (kV/m) (T) Ref.
General Exposure Criteria
NRPB
max. body 25 2.0 NRPB,1993;
limbs only 5.0 McKinlay, 1993
24-h avg. 0.2
ICNIRP(b) ICNIRP, 1994
workers-avg. in work day 0.2
workers-maximum 2.0
workers-limbs 5.0
gen. public-continuous 0.04
CENELEC CENELEC, 1995
workers-whole body 42 2
gen. public 14 0.04
workers-limbs 5
gen. public-limbs 0.1
MRI Exposure Criteria
IEC 2.0' IEC,1995
NRPB NRPB,1991
trunk/head 2.5'
limbs 4.0
IRPA IRPA,1991
trunk/head 2.0
limbs 5.0
FDA 2.0 FDA, 1992
Notes: (a) Limits may be extended to 4 T under specially controlled conditions; (b) Warnings
expressed below 0.5 mT for individuals with implanted pacemakers. See also ICNIRP (1998).
person, provided the duration of dB/dt is sufficiently long (Table 9.6).

Although such fields would be considered extremely large as compared
with environmental exposures, even in occupational settings, such fields are
attainable in magnetic resonance imaging machines. When magnetic reso-
nance imaging technology was first introduced, such fields were not con-
sidered attainable. However, subsequent advances in MRI echo planar
technology have drastically increased the limits of attainable fields to the
point that future development will be limited not by attainable technology,
but by the need to avoid nerve and heart excitation.
Pulsed Field Limits in Magnetic Resonance Imaging

The new generation of fast imaging echo planar MRI devices involve sig-
nificantly greater switched gradient fields than used previously. To under-
stand the conditions required for peripheral nerve stimulation with this
advanced technology, the U.S. Food and Drug Administration (FDA) com-
missioned two studies that established theoretical predictions of nerve and

heart excitation by MRI switched gradient fields (Reilly, 1989, 1991). These
studies formed the basis of advisory guidelines on MRI switched gradient
fields adopted by the FDA in the United States (FDA, 1992) and subse-
quently by the International Electrotechnical Commission (lEe) in Europe
(IEC, 1995). Subsequent revisions to the FDA guidelines (FDA, 1995)
identifies previous FDA limits as "levels of concern," above which the
manufacturer must establish through human volunteer studies that painful
stimulation will not occur. A decade ago the limits expressed in the FDA
and IEC guidelines would have imposed no real limit on MRI technology.
Today, these limits impose significant constraints on operating parameters
for the new generation of MRI machines.
The studies cited above were theoretical ones that combined the nonlin-
ear models of myelinated nerve, described in Chapter 4, with magnetic
induction models, as described in Chapter 9. At the time the first FDA-
sponsored study was completed, there was little experimental verification
for these predictions, although focal nerve excitation by small coil systems
had already been demonstrated. Subsequently, various experimental stud-
ies with large coil systems (reviewed in Chapter 9) demonstrated that
thresholds for both nerve and the heart are consistent with theoretical
predictions.
Subsequent to their adoption by the FDA, these guidelines were adopted
by the National Radiological Protection Board in the United Kingdom
(NRPB, 1991; Saunders, 1991), and later by the International
Electrotechnical Commission (IEC, 1993, 1995). Figure 11.5 illustrates the
IEC Guidelines, on which are superimposed several reaction curves. The
IEC curves (a) and (d) separate three operating regions. The lowest region
is considered an uncontrolled exposure zone, in which only routine patient
monitoring is required. In the middle region, designated the first controlled
operating zone, deliberate action and medical supervision are required. In
the uppermost region, the second controlled operating zone, specific secu-
rity measures are required to prevent unauthorized operation in this zone,
and patient exposure is permitted only under a human studies protocol
approved according to local requirements.
Curve (b) represents expected thresholds for peripheral nerve excitation
based on the studies described in Chapter 9, in which the magnetic field is
perpendicular to the sagittal or frontal cross-section of the patient; for a
field oriented along the longitudinal dimension of the body, the nerve
excitation threshold would be about 50% greater. Curve (c), representing
the threshold of discomfort or pain, is elevated above the perception curve
by a factor of two in accordance with similar multiples obtained from
human electrocutaneous stimulation (Table 7.3). More recent experimental
data with human subjects excited by experimental MRI devices suggest that
painful reactions may be somewhat closer to the perception curve-elevated
above it by a factor of about 1.3 (Budinger et ai., 1991; Bourland et ai.,
D IEC Second Controlled

Operating Zone
IEC First Controlled
4000 r"lln'"T1rrT1rr-r-T'"rTmmr-,""i Operating Zone
"" • IEC Uncontrolled
", ,~,
"" 1:1 Zone
" "0.
1000 ~" "~~
",0"" ',~
"....~. "o~
c0'
,,~o " 'f-C};
',~', ~.1'
,~. ,'0
,,~~. ,,~.~
~
-0
",0" ",""O~
,~ ;0 '- 0
, " ",
,~ .....
100 " -"",,-"-- - --

"
"
10
1 10 100 1000 10,000
Duration of Change of Magnetic Flux Density ij.ls)
FIGURE 11.5. lEe criteria and anticipated reactions to pulsed magnetic field-
whole-body exposure of large adult. (a) Upper limit of lEe uncontrolled exposure
zone; (b) threshold of nerve stimulation; (c) threshold of discomfort or pain; (d)
upper limit of lEe first controlled operating zone; (e) one-percentile threshold for
cardiac excitation; (f) 50-percentile threshold for cardiac excitation. Limit criteria
from lEe (1995); reaction threshold from Reilly (1992).
1997). Curves (e) and (f) are one- and 50-percentile thresholds for cardiac
excitation (see Chapter 9). The upper limits on curves (a) and (d) were
intended to limit whole-body heating effects.
The curves appearing in Fig. 11.5 are asymptotic expressions of the
strength-duration (S-D) relationships for neural and cardiac excitation by
rectangular, monophasic dB/dt pulses (as in Fig. 4.2). The S-D time con-
stant, T" which defines the corner of the lower asymptote, is taken as 120,us
for nerve stimulation as determined by a myelinated nerve model for a
terminated or sharply bent axon (see Sect. 4.5); a time constant of 3ms has
been chosen for cardiac stimulation, which is considered applicable to
stimulation over a large area of the myocardium (see Sect. 6.3). Although
these curves specifically apply to rectangular monophasic dBldt pulses, they
also represent conservative lower limits to excitation thresholds for pulsed
or continuous sinusoidal stimuli, provided that one interprets the horizontal
axis as the duration of a half-cycle of the sinusoidal waveform (Reilly,
1989).
The lower plateaus of the curves in Fig. 11.5 represent the minimum
excitation thresholds for long duration stimuli (tp > lOie)' For short dura-
tion stimuli, (tp < 0.1 ie) theoretical thresholds are inversely proportional to
duration even into the sub microsecond range. Indeed, experimental data
supports this interpretation for durations down to a small fraction of a
microsecond (see Chapter 7). However, curves (a) and (d) are limited to
upper plateaus in order to avoid unacceptable tissue heating associated with
eddy currents from high dB/dt values. The placement of the upper plateau
depends on the temperature rise one is willing to accept, and the duty factor
one assumes for the pulsed stimulus.
The standards discussed here do not consider the possibility of patient
implants. For a treatment of that subject, see (Reilly and Diamant, 1997;
Buechler et ai., 1997).
Extrapolation of MRI Exposure Limits to

EMF Standards
The maximum permissible exposure limits of IEEE/ANSI C95.1 (Fig. 11.2)
are stated as rms values applicable to continuous, sinusoidal radio fre-
quency fields. A difficulty arises for pulsed sinusoidal waveforms of low
duty factor, for which the allowable peak field may rise above the nerve
excitation threshold, yet still satisfy the rms limit of the standard. If Be
represents the allowable peak field for continuous exposure at a particular
frequency, then the allowable peak field Bo consistent with a constant rms
value, for an arbitrary duty factor df is given by:
B Be (11.4)
o = d l/2
f
where df is the fractional duration of the on time during the repetition
period (assuming an integer number of half-cycles of the sinusoidal varia-
tion). Equation (11.4) is a criterion based on constant energy.
To avoid nerve stimulation from pulsed fields, it is necessary to impose a
constraint on the peak value of the allowable field. Since pulsed field limits
have already been specified for MRI exposure, we proposed the application
of MRI criteria to IEEE/ANSI standards (Reilly, 1998).
As seen in Fig. 11.5, MRI exposure criteria are stated in terms of the
duration of dB/dt. In order to make a connection with sinusoidal fields, we
need to define the relationship between the peak field and its time deriva-
tive. For a pulsed field consisting of a linear increase over the pulse duration
(trapezoidal pulse), the relationship is:
Bo = Botp (pulsed trapezoidal field) (11.5)
where B)s the value of dB/dt during the pulse, and ip is the phase duration,
defined as the pulse duration for a rectangular pulse, or the half-cycle time
for a sine wave (see insert of Fig. 4.15). For a sinusoidal field, the relation-
ship is
1 .
Bo = - Bot p (sinusoidal field) (11.6)
7r
In this case, Eo is the peak of dB/dt during the sinusoidal cycle.

Figure 11.6 plots the nerve excitation limit of Fig. 11.5 (curve (a) in both
figures), which is read on the left vertical axis of Fig. 11.6. Curves (b) and (c)
plotted on Fig. 11.6 show the corresponding peak flux density according to
Eqs. (11.5) and (11.6), and are read on the right vertical axis. Curve (c) may
be interpreted as a limit to the peak value of the flux density for sinusoidal
fields.
Figure 11.7 shows curve (a) of Fig. 11.6 expressed on a format consistent
with the magnetic field standards of C95.1; also plotted is the limit imposed
by C95.1 on the peak allowable field for a 100% duty factor waveform (d! =
1.0). Peak flux density, Bm is expressed on the left vertical axis of Fig. 11.7
in units of mT; the right vertical axis gives the corresponding peak magnetic
field strength, Hm in Aim, in conformance with the units of C95.1. The
horizontal axis has been expressed in frequency units using the relationship
104 102
(i) Bo=
{80 tp (Pulse)
.
.§.O tp (sine-wave)
n
-- -tp
t:.
0
~ 1t
C}~
-co Q>~6 10 I='
<Ii' 10 3
>
~
"~.
~i'
.s0
> co
.~
-0
.;!-
'Ci5
Q) c
:§ 0
Q)
Z' x
:J
'iii
c 102 =- (c) u:
Q)
-0 -"
ell
x
:J
Q)
0...
;;:::
-"
ell
Q) Bo=20T/s
0...
10 10- 1
10 102 103 104
Phase duration tp (l1s)
FIGURE 11.6. Magnetic field exposure criteria-adapted from MRI exposure limits
for switched gradient fields. (a) = lEe criteria on B (MRI); (b) = curve (a) limits on
B for pulsed dBldt; (c) = curve (a) limits on B for continuous sine dB/dt. (From
Reilly, 1998.)
10S
102c---~--,--------,--------,--------,------~
E
I I
~
-flr 0
::r:
t='
.§.
ID
0
10
.rvvv-
dB/dt waveforms
10 £
c;,
&::
~
U5
~
rn
"0
CD
&::
<D iI
"0 0
><
::::J
Nerve
Stimulation 103
:a;
;;:::: &::
..I<:
1.0 Limit 01
co
co
<D ~
0.. ..I<:
co
C9S.1 @ 100% d.f. <D
0..
0.1L--------L--------~------~------~--~--~
102
10 100 1K 10K 100K 1M
1/(2 tp) = f (Hz)
FIGURE 11.7. Magnetic field exposure criteria that avoids peripheral nerve stimula-
tion (including safety margin) for pulsed sinusoidal fields. Nerve stimulation limit has
safety factor of 2X below threshold for pulsed or CW sine wave. (From Reilly, 1998.)
1
f=- (11.7)
2tp
For frequencies above 3kHz, Bo = 0.76mT, or equivalently, Ho = 605 AIm
according to the MRI-based limits. The peak value of Ho from C9S.1 with
df = 1.0 is 231A/m. It can be seen from Eq. (11.4) that the IEEE/ANSI
rms limits would exceed the peak MRI-based limits for df ::5 0.15.
One might question whether the avoidance of peripheral nerve stimula-
tion with exposure to pulsed magnetic fields is a proper criterion for stan-
dards that are intended to avoid hazardous conditions, not merely
perceptual ones. Although a reaction at the threshold of perception is
unlikely to be hazardous, most would agree that unpleasant or painful
stimuli could be hazardous under many conditions. Since the threshold
multiple from perception to painful reactions is about 1.4 (Sect. 9.7), the
multiple implicit it the proposed peak limits is about 2.7 for a large subject
with frontal exposure of the torso and 4.4 with longitudinal exposure (refer
to Table 9.6). For comparison, an acceptability margin of 10 or more has
been incorporated into the SAR limits that underlie existing C9S.1 stan-
dards. Note, however, that since SAR is an energy metric, it is proportional
to the square of the field (Eqs. 11.2; 11.9). Therefore, a margin of 10 in SAR
corresponds to a margin of 3.16 in the associated field.
Limits of Applicability
For stimulation by sinusoidal currents, thresholds of excitation rise approxi-
mately in proportion to frequency beyond a critical frequency (a few kHz
for neurosensory or neuromuscular stimulation), provided that the stimulus
consists of multiple sinusoidal cycles (see Sect. 4.6). This fact is responsible
for the lower plateau of nerve excitation threshold beyond a few kHz when
expressed in magnetic flux density units as in Fig. 11.7-as we increase
frequency, the threshold increases in proportion to frequency, as does
dBldt (and consequently, current density). The two effects compensate one
another, resulting in a constant magnetic flux threshold of excitation versus
frequency.
As we increase frequency, the current density for nerve excitation
increases proportionately and can become so great that electrical excita-
tion thresholds will exceed requirements for perception due to heating.
For sinusoidal currents that are at least a significant fraction of a second
in duration, the cross-over between electrical and thermal perception
thresholds occurs at about 100kHz (Sect. 7.5). The thermal cross-over
accounts for the diminishing limits beyond 100kHz in the C9S.1 standard,
as seen by the lower curve in Fig. 11.7-the indicated limit ensures that
the current density within the medium is maintained to a constant value
that does not exceed thermal perception thresholds. However, for pulsed
sinusoidal waveforms, it is possible to electrically excite sensory neurons
without thermally exciting thermal receptors. This statement can be under-
stood in terms of excitation models presented in previous chapters.
For instance, we have seen that in Fig. 4.19 that the perception threshold
to a high-frequency sinusoidal stimulus reaches a plateau at durations
beyond a millisecond or so. Consequently, by reducing the duration of a
continuous sinusoidal stimulus to a millisecond, for example, one signifi-
cantly reduces its rms value (i.e., heat producing capacity) without signifi-
cantly reducing the sensory perception threshold associated with that
stimulus.
The question arises: how high in frequency can we extend the nerve
stimulation plateau shown in Fig. 11.7? As noted in Sect. 7.5, human per-
ception experiments with relatively long duration sinusoidal currents dem-
onstrate the proportionality between electrical thresholds and frequency to
100 kHz, at which point thermal perception thresholds dominate, suggesting
a frequency limit to at least 100kHz. Experiments with anesthetized rats,
which allowed significant tissue heating, demonstrated neuromuscular
stimulation with approximate frequency-proportional thresholds to
1 MHz-the highest frequency tested (Fig. 4.22). From this observation, it
would appear reasonable to extend the plateau of Fig. 11.7 to at least
1 MHz. And from human sensory experiments with brief conducted current
pulses, one could infer experimental correspondence to perhaps 10 MHz.
This statement is based on the fact that human electro cutaneous perception
thresholds are inversely proportional to the duration of brief monophasic

currents to about O.l,us-the shortest tested duration (see Sect. 7.3).
11.4 Consideration of Spark Discharges

in EMF Standards
We have seen in Chapters 2 and 9 that spark discharges may occur when a
person contacts metallic objects in high-intensity electric fields. Spark dis-
charges induced by alternating fields may be felt and can be unpleasant or
even painful if strong enough. Besides causing sensory effects, spark dis-
charge can cause local skin erosion or burns because of the very small
effective area of contact, particularly when conveyed to dry skin in an area
where the corneal layer is thin. In those areas the energy of the spark is
primarily dissipated in a very small volume of tissue. Section 2.6 discusses
corneal degradation as a result of spark discharges.
To support a spark discharge to dry, intact skin, an object charged by a
static or ELF field must attain a critical voltage of about SOOV relative to
the intact, dry skin, as noted in Sect. 2.6. That voltage represents the
minimum potential necessary to initiate a spark to the skin and is approxi-
mately the potential below which a spark to intact skin will extinguish. A
plateau voltage as low as 330V has been measured for spark discharges to
the skin of the forearm on which the corneal layer has been stripped away,
suggesting that 330V is also a minimum spark initiation potential for moist
or damaged skin.
The preponderance of published research concerning effects of spark
discharges on humans has been conducted with static charges and at 60Hz.
Lacking specific laboratory data applicable to frequencies above 60Hz, it is
difficult to state the amount of skin erosion or damage that would occur with
discharges from higher frequency fields. Unless the speed of approach
between the person and the energized conductor is high in relation to the
oscillation frequency, a spark will initiate at the peak of the sinusoidal
waveform, regardless of frequency. For frequencies substantially above
60Hz, the number and severity of sparks to the skin during a discharge
episode will depend on the frequency, and the discharge time constant (see
Fig. 2.34). Also, the potential required for the breakdown of air decreases at
frequencies substantially above 60Hz. This occurs when the transit time of
positive ions or electrons is less than the half-cycle time of the oscillating
field (Craggs, 1978). At radio frequencies, the reduction of breakdown
potential across small metallic electrode gaps «1 mm) in air is about 1S% to
20%. At frequencies of several GHz, more substantial reductions can occur.
Table 11.6 lists the peak E-field needed to achieve a spark discharge from
various objects using a spark initiation criterion of 330 and SOOV; values are
calculated using the parameters listed in Table 9.1. Table 11.6 applies to
situations where the discharging object has dimensions much smaller than
Consideration of Spark Discharges in EMF Standards 477
TABLE 11.6. Electric field parameters for spark discharge induction.

Required E
CO VJE Va = 500V 330V
Charged object (pF) (m) (Vim) (Vim)
Person (h = 1.8m) 150 0.48 1042 638
Auto (mid size) 200 0.21 2,380 1,571
Auto (large) 2,000 0.16 3,333 2,062
School bus 3,700 0.26 1,923 1,269
Commercial bus 2,900 0.45 1,111 733
Fence wire 720 1.00 500 330
(h = 1m, L = 100m)
Table applies to quasi-static condition with vertically polarized field.

Discharging object is several object dimensions from the source of the field. Field wavelength
is much larger than major dimension of object.
the wavelength of the field and is several object dimensions distant from the
source of the electric field. The last column of Table 11.6 lists the minimum
undisturbed E-field (i.e., the field in the absence of the discharging object)
needed to support a spark discharge. Except for the fence wire, the mini-
mum E-field required for spark discharges exceed the low-frequency E-
field limits of IEEE/ANSI C9S.1, except as time-averaging permits higher
peak fields with pulsed EMF.
It is important to recognize the potential hazards from burns to personnel
who contact metallic objects in fields produced by radio frequency (r-f)
transmitters. An example would be on an aircraft carrier, where a commu-
nications antenna may be placed near an aircraft that is standing on a
metallic deck. A voltage may be induced on the aircraft relative to the deck
that is sufficient to cause a skin burn when an individual touches the air-
craft. Note that there may be significantly different voltages developed at
various points on the aircraft if the wavelength of the energizing field is less
than the dimensions of the induction object (a typical situation). The U.S.
Navy uses a voltage criterion of 140V-rms in r-f fields to define a potentially
hazardous situation that could cause a person pain, visible skin damage, or
could cause an involuntary reaction (NA VSEA, 1982). The navy recognizes
that because of the many variables involved, it is not uncommon to encoun-
ter significantly higher voltages that do not result in a burn problem. In tests
conducted near a 1kW transmitting antenna operating in the S to 10MHz
range, I experienced very small skin lesions on the fingertip and knuckle
when touching a fighter aircraft having an open circuit voltage of 140V-rms
at the point of contact. The contacts were attended by a stinging sensation,
and lesions appeared as localized white spots on the corneum.
The navy voltage criterion amounts to a peak voltage of 198V, which is
substantially below the spark discharge criterion of 330 to SOOV mentioned
above for ELF potentials. If we reduce ELF breakdown potentials by 20%
for application to r-f, one would infer a peak breakdown potential in the
range 233 to 354 V based on the spark discharge data-still well above the
navy burn criterion. It is possible that the r-f burns observed at 140V-rms
do not involve air breakdown, but rather corneal breakdown when direct
contact is actually made. The result might be essentially the same from the
point of view of skin damage-a very small area of current conduction
would be involved in either case. While an open-circuit voltage criterion of
140V-rms might be protective of r-f burns for most cases, it is likely to be
overly conservative in situations where the current conducted by an indi-
vidual is small. A criterion combining open-circuit voltage and conducted
current is probably needed to adequately address r-f burn hazards.
11.5 Absorbed Energy and Thermal Considerations

in EMF Standards
Specific Absorption Rate (SAR) Limits
Exposure standards in IE EEl ANSI C95.1 are related to the specific absorp-
tion rate (SAR), which is the energy absorbed per unit mass of biological
tissue. SAR may be related to the internal current density by
J2 J2(2
SAR = - = - (11.8)
op p
or, in terms of the internal electric field by
oE2 E?
SAR= - ' =- (11.9)
p (2P
where E; is the rms in-situ electric field, J is rms current density, 0 is tissue
conductivity, (2 = 110 is the tissue resistivity, and p is tissue density. Tradi-
tionally, SAR is specified in units of Watts per kilogram.
We can relate SAR to tissue temperature rise using the heat equation (see
Eq. 10.3), with the following result
l1T SAR
(11.10)
pc
where 11.T is the temperature rise (oq, t is the duration of exposure, p is
tissue density (gm- 3 ), c is tissue specific heat (Jg-1C- 1), and SAR is in units
(Wig). If we express SAR in W/kg, and use c = 3.8 and p = 1.05 X 106 for
muscle tissue (see Table 10.5), the relationship is
(11.11)
Equation (11.11) is a worst-case expression in which there are no thermal
losses from the tissue, thereby implying that the temperature would rise
indefinitely as long as the energy input were provided. In reality, thermal
Absorbed Energy and Thermal Considerations in EMF Standards 479
losses will occur locally through conduction to adjacent tissue and vascular
cooling, and will occur over the body through thermal radiation from exter-
nal body surfaces, respiration, and evaporative processes. As an illustration
of temperature rise, Eq. (11.11) states that with SAR = 1 W/kg, a tempera-
ture rise of 1 °C would require 4,000s (1.1 hr) of continuous exposure under
worst-case conditions.
The value of SAR expressed by Eqs. (11.8) and (11.9) will vary with
frequency for a given in-situ E-field due to the variation of tissue conductiv-
ity (a) with frequency. For instance, at frequencies of 103 ,108 , and 109 Hz, a
(parallel) for muscle tissue is approximately 0.5, 0.9, and 1.4S/m respec-
tively (see Table 2.1). Consequently, if we calculate the value of Ei associ-
ated with SAR = 1 W/kg in muscle tissue, we obtain values of 45, 35, and
27V/m at frequencies of 103 , 108 , and 109 Hz respectively.
Endogenous Field Limits to EMF Interactions

Biological systems experience endogenous electrical forces that provide an
ultimate limitation on the minimum induced electrical force that can have a
measurable biological effect. One of these forces is a consequence of the
thermal agitation of charges within the biological medium. Other endog-
enous electrical forces include "lIf noise" from cell membrane activity,
"shot noise" associated with the discrete nature of electronic charge, and
fields from electrical activity of nerve and muscle (Barnes, 1996). The
following development demonstrates some of the considerations concern-
ing thermal limitations to bioelectric interactions (after Weaver and
Astumian, 1990). Whereas thermal noise represents one basic limit on
bioelectric interactions, the other sources of noise can only add further
limitations. In fact, lIf noise (so-called for its frequency spectrum) may be
the dominant source of noise at cellular membranes for frequencies below
160Hz (Barnes, 1996).
The mean-squared voltage generated by random thermal fluctuations
(so-called Johnson-Nyquist noise) is given by
(11.12)
where R is the resistance across which the voltage is measured, k is
Boltzman's constant (1.8 X 10- 23 JK- 1), T is absolute temperature, and I'l.f
is the noise bandwidth.
In the case of electromagnetic field interactions, it is widely believed that
the site of interaction is at the cellular membrane, where we will consider
the application of Eq. (11.12). The noise bandwidth of the membrane is
related to its time constant i m, by I'l.f = 1I(4im), where im = RC, R is the
membrane resistance, and C is its capacitance. For a circular cell of radius
r and membrane thickness d, C = 47rc o crr 2Id, where Co is the dielectric
permittivity of free space (8.85 X 1O- 12 Fm- 1), and Cr is the relative permit-
tivity of the membrane. We therefore can write Eq. (11.12) as
(11.13)
If we evaluate Eq. (11.13) for a circular cell with T = 310K, d = 5nm, r =

lO.um, and E, = 2, and taking the square-root, we obtain V kT = 2.S X 1O-5 V,
which is the rms noise voltage present across the cellular membrane.
If an induced field it to have a cellular effect, it must not be swamped
by endogenous thermal noise. Consider therefore the intracellular E-field
necessary to create a transmembrane voltage equal to the noise voltage
(i.e., signal-to-noise ratio = 1). The maximum voltage Vm developed across
the membrane of a spherical cell in response to an intracellular field E; is
Vm = 1.5 E;r, where E; is the in-situ or intracellular E-Field (see Sect. 4.3).
Therefore, the minimum in-situ E-field that would just equal thermal
noise (Vm = Vu) in the example spherical cell is E; = 1.9 Vim. For an
elongated cell of length L, the maximum induced membrane voltage is
V m = EiLl2. Following a procedure similar to the above example but for
an elongated cell with L = 150.um and r = 25.um (e.g., a fibroblast or
muscle cell), the in-situ field that would just equal thermal noise is
calculated as Ei = O.OSV/m. These examples would be restricted to
frequencies below a few kHz, because of the bandwidth limitation of the
membrane.
Others have calculated thermal noise limitations using different assump-
tions about the site of the effect. For instance, a value Ei = 4mVim for unity
signal-to-noise ratio was calculated by Polk (1993) by assuming that the
pertinent mechanism of action is the redistribution of charge at the
counterion layer of the cell surface, rather than the transmembrane voltage.
We expect that a measurable cellular response would require a signal-to-
noise ratio greater than unity-information processing systems typically
require signal-to-noise voltage ratios of at least 4 for meaningful signal
detection and false alarm probabilities. Consequently, it is likely that a
similar multiple applies to the noise-equivalent fields cited above for a
credible biological effect.
Consider the thermal noise limits in the above examples in comparison to
induced in-situ fields associated with typical environmental fields at 60Hz.
For example, consider a magnetic field of lO.uT, and an electric field of
5kV/m-values that are at the upper limits of the fields that may be en-
countered beneath high voltage transmission lines. Using the principles
and models described in Sect. 9.5, we calculate that the magnetic field
would induce a maximum E-field of O.6mV/m at the periphery of the
torso of a large adult. And using the principles described in Sect. 9.3, the
electric field is calculated to induce 15mV/m in the ankles of that
person, and much smaller fields elsewhere in the body if grounded through
the feet.
It can be seen in the above examples that even relatively strong ELF
environmental fields induce in-situ fields below, or at best marginally above
Consideration of EMF Interaction Mechanisms in Standards Setting 481
the most conservative thermal noise limits discussed above. Many experi-
menters report biological effects in which ELF in-situ fields are below
lOmV/m, and in some cases are as low as lO,uV/m (i.e., less than presumed
thermal noise limits) (Weaver and Astumian, 1990; Polk, 1993; Tenforde,
1993).
It has been argued that thermal noise limitations cannot be overcome by
any known biophysical mechanism, and therefore that biological effects
from weak environmental fields are impossible (Adair, 1991). What then
are we to make of the many reports of experimentally determined biologi-
cal effects that result from induced fields below calculated thermal noise
limits? Should we dismiss all such reports as violating thermal limits, and
ipso-facto flawed as some have suggested? Or is it possible that other
biophysical mechanisms might be responsible for overcoming thermallimi-
tations? In order to address these questions, the next section discusses
established and proposed mechanisms that may be involved in biophysical
reactions.
11.6 Consideration of EMF Interaction Mechanisms

in Standards Setting
Categories of Mechanisms of Bioelectric Interaction
When setting limits on electromagnetic exposure, it is often necessary to
make assumptions about underlying mechanisms. With a mechanism
model, one can specify appropriate dosimetric measures and parametric
relationships, such as frequency or temporal sensitivity factors. For
instance, if we were to assume that the biological action responds to tissue
heating, then we would conclude that the energy deposition is the appro-
priate dosimetric parameter, which can be measured by SAR units.
Furthermore, parametric relationships involving the frequency of the
electromagnetic energy, and its temporal and spatial distribution can be
adequately modeled. On the other hand, when dealing with membrane
polarization phenomena, such as nerve excitation, it is the in-situ electric
field that produces the effect. Furthermore, the spatial and temporal struc-
ture of electrical forces that influence excitable membrane effects will be
quite different than those that determine a thermal effect.
Numerous mechanisms have been advanced to explain a variety of bio-
electric phenomena. With application to standards criteria, it is necessary to
differentiate between established and proposed mechanisms. We define an
established mechanism as one for which interaction with a person is well
established, and for which thresholds of reaction are understood. On the
other hand, a proposed mechanism is one which is not sufficiently under-
stood to define the threshold of interaction in a living person, or where
experimental results have not yet been established with confidence.
The classification of a mechanism as "proposed" does not necessarily

mean that biological activity connected with that mechanism is in doubt,
but, of course, in some cases it may mean just that. It is possible that a
mechanism might be well established at a cellular level, for instance, but
that its application to the whole intact person is not presently understood.
Or a mechanism might be well established in a nonhuman species, but its
application to humans may be uncertain. From the point of view of standard
setting, we would classify such a mechanism as "proposed." Only estab-
lished phenomena applicable to living humans should form the basis for
standards. On the other hand, developing knowledge on proposed mecha-
nisms should be carefully monitored to ascertain whether reclassification to
the "established" class is warranted.
Tables 11.7 and 11.8 provide a listing of various mechanisms that have
been advanced to explain electric and electromagnetic interactions with
biological systems. Table 11.7 lists those mechanisms that can presently be
characterized as established; Table 11.8 lists proposed mechanisms. Two
columns indicate whether the effect is considered to be associated with the
in-situ electric field or magnetic field. In Table 11.7, an attempt has been
made to provide a numerical estimate of the lowest in-situ field that is
expected to have a measurable biological effect under most favorable con-
ditions; we will apply the term "rheobase" to these minimum thresholds.
The column labeled "condition" provides a brief indication of conditions
TABLE 11.7. Established mechanisms of human interaction with E&M fields.

In-situ In-situ
E-field B-field Section
Mechanism (Vim) (T) Condition reference
1. Synapse activity alteration 0.05 NA [= 20Hz 3.7,9.8
by membrane polarization (phosphenes)
2. Peripheral nerve excitation via 6 NA t> Ims 3.3,4.4
membrane depolarization
3. Muscle cell excitation by membrane
depolarization (skeletal) 6 NA t> IOms 8.3
(cardiac) 12 NA t> 10ms 6.2-6.5
4. Electroporation (reversible) 50 NA t> O.lms 10.2
(irreversible) 300 NA t> 0.1 ms
5. Resistive (joule) heating
(e.g., SAR = 1 Wlkg in muscle NA t '" continuous 11.5
for I'J.t '" Ihr, I'J.T '" 1 0C)
[= 103 Hz 45
[= 108 Hz 35
[= 109 Hz 27
6. Audio effects via thermoelestic expansion 300 NA [= O.1-IOGhz 9.8
(Peak SAR '" 100Wlkg in brain) t'" Ims
7. Magneto hydrodynamic effect NA 1.5 ["'0 9.11
References are to book sections.

TABLE 11.8. Proposed mechanisms of human interaction with E&M fields

In-situ In-situ
E-field B-field
Mechanism (Vim) (T) Condition Reference
1. Soliton mechanisms ,j NA Lawrence & Adey, '82
through cell membrane
proteins
2. Spatial/temporal ,j NA Litovitz et aI., '94
cellular integration
3. Stochastic resonance ,j NA Krugilikov &
Derfinger, '94
4. Temperature mediated ..j NA mm waves Alekseev et aI., '94, '97
alteration of membrane
ion transport
5. Plasmon resonance ,j NA f= 109_1011Hz Fisun, '93
6. Radon decay product ,j NA Ea = 1kV/m Henshaw et aI., '96
attractors
7. Rectification by cellular ,j NA Astumian et aI., '95
membranes
8. Ion resonance NA ,j f(elm, Bd/BaJ Liboff, '85; Lednev, '91;
f= lO-lOOHz Blanchard &
Blackman, 94
9. Ca ++ oscillations NA ,j Liboff, '93
10. Nuclear magnetic NA ,j Blackman et aI., '88
resonance
11. Radical Pair Mechanism NA ,j Steiner & Ulrich, '89;
Grissom, '95
12. Magnetite interactions NA ,j Kirschvink, '89
Notes: Ea: ambient electric field; elm: charge-to-mass ratio of resonant ion; Bd/Bac: steadyl
alternating magnetic field.
which are necessary to attain the rheobase threshold. The last column refers
to book sections where the mechanism is treated in some detail.
Table 11.8 lists proposed mechanisms of human interaction with electric
and magnetic fields. In this table, the action of the mechanism is classified as
being mediated by the in-situ electric or magnetic field, as indicated by a
check in the appropriate column. A numerical value for human reaction is
not provided in Table 11.8 because of the lack of sufficient knowledge
concerning the mechanism. However, an attempt has been made to indicate
exposure conditions where the mechanism is thought to be active, where
such information can be determined. The last column of Table 11.8 refers
the reader to a few citations where the mechanism is defined. The listing is
not meant to be a comprehensive bibliography, as that would require an
inordinate amount of space.
The following paragraphs give a brief description of each mechanism.
The references that are included in the descriptions are intended to provide
minimal background on the particular mechanism.
Established Mechanisms for Human Bioelectric

Response
The first four mechanisms listed in Table 11.7 are membrane polarization
effects, which are produced by an in-situ electric field that exerts a polariz-
ing voltage across the nerve membrane (See Sect. 4.3). For elongated cells,
such as nerve cells, the most favorable orientation for membrane polariza-
tion effects is when the E-field is aligned with the long axis of the cell, in
which case the maximum polarization will occur at the terminus of the cell.
Items (5) and (6) are thermal mechanisms. Item (7) is a magnetic field
effect. The following paragraphs provide a modicum of detail for Table 11.7
Synapse Interactions (Sects. 3.7 and 9.8)

Post-synaptic membrane potentials vary with a complex of excitatory
and inhibitory cells that synapse on a given cell. Small changes in the
presynaptic resting potential can be greatly multiplied in the postsynaptic
membrane and make an effective synapse inoperable, or a weak synapse
highly effective in producing a postsynaptic action potential (see Sect.
3.6).
An example of this effect is attributed to the phenomenon of electro- and
magneto-phosphenes, which are the visual effects that are produced when
electric currents or magnetic fields are applied to the head (see Sect. 9.8).
Researchers have concluded that phosphenes are generated through
modification of synaptic potentials in the receptors or neurons of the
retina, rather than direct stimulation of post retinal pathways or the visual
cortex.
Using the observed magneto-phosphene thresholds at the most sensitive
frequency (20Hz), the induced E-field at the location of the retina is calcu-
lated to be about 50mV/m-a value that is a factor of about 100 below
the rheobase nerve excitation thresholds (refer to Sect. 9.8). It does not
necessarily follow that such low thresholds would apply to other neural
synapses, because of the highly specialized configuration of neurons in the
retina.
Peripheral Nerve Excitation Through Membrane Depolarization

(Sects. 3.3 and 4.4)
Nerve excitation is initiated by depolarization of the neural membrane,
which activates voltage-gated ion channels, thereby producing a propagat-
ing action potential (see Chapter 3). In contrast to the graded response of
synaptic activity mentioned above, nerve excitation is a threshold response,
sometimes referred to as an all-or-nothing phenomenon. Nerve excitation,
characterized by a propagating action potential, is triggered when the mem-
brane is depolarized by about 15-20mV. The external field necessary for
nerve excitation depends on the duration of the excitation. A time constant
in the range 100 to 200.us typically applies to S-D curves for nerve excita-
tion. A minimum threshold of about 6V/m is found for the largest myeli-
nated nerves (=20.um diameter), with a long duration (=2ms) monophasic
E-field pulse (see Chapter 4).
Muscle Cell Excitation Through Membrane Depolarization

(Sects. 6.2 to 6.5 and 8.3)
Although electrical excitation of skeletal muscle typically occurs through
excitation of motor neurons, it is possible to directly stimulate muscle cells
(see Chapter 8). Strength-duration time constants of muscle cells are typi-
cally in the range of 1 to 10ms, which is a factor of 10 greater than the time
constants for nerve excitation. Table 11.7 lists a rheobase threshold for
skeletal muscle and cardiac muscle cells based on experimental and theo-
retical data. The rheobase field for muscle excitation is similar to that for
nerve stimulation i.e., 6V/m. The median rheobase for cardiac excitation is
about 12V/m; at the one-percentile level, the cardiac rheobase is about 6VI
m (see Chapter 6). In order to achieve these rheobase thresholds, it is
necessary to stimulate the tissue with a monophasic pulse having a duration
greater than about lOms.
Electroporation (Sect. 10.2)

The electric field normally developed across a cellular membrane is very
large. With a membrane potential of 0.1 V and a membrane thickness of
10- 7 m, the electric field would be 106 V/m across the membrane. If the cell
is sufficiently hyperpolarized relative to its normal resting potential, the
intense electric field developed across the membrane can produce pores
(see Chapter 10). Hyperpolarization to a membrane potential around
200mV will produce temporary pores, which revert to a normal condition
with the cessation of the hyperpolarizing field. When the membrane poten-
tial is raised to about 800mV, the individual pores become enlarged or can
fuse and become irreversible. The electric field values listed in Table 11.7
apply to electric field pulses of sufficient duration (t > 0.1 ms) that can lead
to reversible or irreversible electroporation of muscle cells.
Resistive Tissue Heating (Sect. 11.5)

Tissue heating results from the passage of electric current through resistive
material. The appropriate measure of heating is energy deposition, which
can be quantified by the SAR measure (see Sect. 11.5). For example, with
SAR = 1 W/kg in muscle tissue, it would require 4,000s (1.1hr) of continu-
ous exposure to achieve a temperature rise of 1°C. The in-situ electric field
associated with this value of SAR in muscle tissue ranges from 45 to 27V/m
for frequencies from 1<r to 109 Hz, as noted in Table 11.7.
Audio Effects Through Thermoelastic Expansion (Sect. 9.8)

Auditory sensations can be perceived by a person whose head is exposed
to microwave energy. The effect is explained by the development of a
thermoelastic wave that results from a brief temperature rise because of the
absorption of pulsed microwave energy within the head. The brief tempera-
ture rise causes an expansion of the absorbing tissue that launches an
acoustic wave within the skull and is perceived mechanically through the
normal auditory mechanism of the ear. The threshold stimulus is character-
ized by energy density in the vicinity of 40,uJ/cm2. The corresponding SAR
value will depend on the pulse width. The threshold E-field listed in Table
11.7 corresponds to a value of SAR = lOOW/kg in the brain, which is the
lowest threshold attributed to microwave hearing. Most data on microwave
hearing have been collected at 2.45 GHz; data at other frequencies are
scarce.
Magnetohydrodynamic Effect (Sect. 9.11)

When ions flow in a direction orthogonal to a magnetic field, a voltage is
produced in a mutually orthogonal direction (e.g., velocity in x-direction,
magnetic field in y-direction, induced E-field in z-direction). For example,
assuming a flow rate of 0.6 rnIs in the human aorta, the authors calculate that
a potential of 15mV would be produced in a field of 1 T, as noted in the
table. Magnetohydrodynamic effects result in vertigo or taste sensations at
1.5T (noted in Table 11.7). In addition to the magnetohydrodynamic gen-
eration of electrical potentials mentioned above, there will also exist a drag
force on a conducting fluid that is flowing within a magnetic field. For a field
of 5T/s, the pressure within human vasculature will be affected by less
than 1 %.
Proposed Mechanisms for Human Bioelectric Response

Numerous mechanisms have been proposed to account for human reactions
to low-level electromagnetic exposure. The placement of some of these
mechanisms in the "proposed" category does not necessarily imply that the
particular mechanism is in doubt, but rather that there presently does not
exist a quantifiable and verified theory through which one may specify
reaction thresholds in the human organism. It must be anticipated, how-
ever, that some of these proposals will be shelved, or simply ignored for lack
of supporting development. Listed below are several mechanisms from
Table 11.8 that have been proposed to account for human reactions to low
level electromagnetic radiation. The frequency thought to result in measur-
able biological activity ranges from the ELF to the microwave regimes.
Soliton Mechanisms (Lawrence & Adey, 1982)

According to this theory, a nonlinear wave (soliton) is created in proteins
associated with membrane channels, which propagates along the protein
and through the membrane, thereby supplying energy for chemical events
within the cell. These processes are thought to be affected by the electric
field external to the cell.
Spatial/Temporal Cellular Integration (Litovitz et aI., 1994a, 1994b)

Various experimenters have reported bioelectric effects from exposures
that produce in-situ E-fields that are below the endogenous field within the
biological medium. One possible explanation is that spatial and temporal
integration processes exist that can differentially enhance spatially or
temporally coherent applied fields with respect to incoherent endogenous
thermal noise fields.
Stochastic Resonance (Krugilikov & Dertinger, 1994)

A weak signal may be amplified by system noise itself in a bistable or
multistable system. The cellular membrane is thought to provide the non-
linear characteristics necessary for such enhancement in biological systems.
Temperature Mediated Alteration of Membrane Ionic Transport

(Alekseev et aI., 1994, 1997)
The authors have determined that in-situ exposure of pacemaker neurons
to millimeter waves (e.g., f = 75 GHz, SAR = 4.2 mW /kg) cause significant
changes in firing rates and spike amplitude. The effect was attributed to a
brief, transient temperature rise of 0.0025°C/s at the neural membrane,
resulting in altered membrane ionic transport. The authors found that with
millimeter radiation, local SAR hot spots on cellular membranes can exceed
1,000W/kg at lOmW/cm2 incident power density.
Plasmon Resonance Mechanisms (Fisun, 1993)

This mechanism refers to sheets of surface charge (plasmons) on cellular
membranes, which display resonance properties in response to an applied
field. The advocates of this mechanism speculate that the associated redis-
tribution of surface charges can alter transmembrane potential, thereby
significantly affecting membrane ion transport, conformation of intra-
membrane proteins and lipid bilayers, with potentially pathological
consequences. According to theoretical models, resonant plasmon activity
would apply to fields in the frequency range 109 to 1011 Hz.
Radon Decay Product Attractors (Henshaw et aI., 1996)

Henshaw and colleagues have found that radon decay products are
attracted to common sources of electric fields, such as power lines and
electric appliances. They observed that the electric fields produced around
power sources can enhance the deposition of radon decay products on
environmental surfaces. They speculate that an external electric field as low
as 1 kV/m might enhance the deposition of aerosols containing decay prod-

ucts in the air passages of the mouth, neck, and chest. They suggest that
these observations may provide a link between ELF fields and human
cancer.
Rectification by Cellular Membranes (Astumian et aI., 1995)

Proteins in cellular membranes are said to rectify oscillating electric fields,
thereby allowing a DC response to accumulate from an AC field. The
development of a DC membrane voltage bias in excitable tissue can be
demonstrated in response to a moderately strong sinusoidal membrane
voltage in the vicinity of 1mV (see Sects. 4.6 and 6.5). However, this action
as an explanation for biological response to very weak fields has not been
established.
Ion Resonance (Liboff, 1985; Lednev, 1991;

Blanchard & Blackman, 1994)
Ion resonance models suggest that a magnetic field will resonate with ionic
motion at a frequency that depends on the charge-to-mass ratio of particu-
lar resonating ions, as well as the copresence of a static magnetic field, such
as the earth's field (see Sect. 9.11). So-called "paramagnetic" formulations
state that the resonant frequency also depends on the ratio of magnitudes of
the steady and alternating magnetic field components.
Ca++ Oscillations (Liboff, 1993)

ELF electric fields surrounding a cell are said to alter the frequency of Ca ++
oscillations in the cellular cytosol (liquid medium of the cytoplasm). These
oscillations are believed to convey information to the cell based on the
frequency of oscillation of Ca ++ ions.
Nuclear Magnetic Resonance (Blackman et aI., 1988)

This mechanism is suggested to affect 45CA ++ efflux from brain tissue. The
theory suggests that a nonzero nuclear spin is required to have an effect,
whereas most biological isotopes have zero spin. Therefore, a NMR
explanation for coupling by weak ELF may be questionable.
Radical Pair Mechanism (Steiner & Ulrich, 1989; Walleczek, 1994;

Grissom, 1995)
A magnetic field can modify the electron valence spin states through quan-
tum mechanical mechanisms during free radical formation, thereby altering
radical-dependent reactions. Such states are independent of random
thermal interactions. Consequently, in principle, there should not exist a
thermal noise limit governing such magnetic field interactions.
ELF Magnetic Field Standards Derived from Established Mechanisms 489
Magnetite Interactions (Kirschvink, 1989, 1996)

Magnetite (ferromagnetic compounds) are found in migratory birds, and
other animals, including mammals. It is found in trace amounts in the
human brain, and possibly other tissue. It has been speculated that these
trace compounds may provide a link for magnetic field interactions with
brain tissue. At microwave frequencies in the range 0.5 to 10GHz, the
mechanism of interaction is said to be through the process of ferromagnetic
resonance, which results in enhanced absorption of electromagnetic energy
by cells.
11.7 ELF Magnetic Field Standards Derived from

Established Mechanisms
We have seen in Sect. 11.2 that the IEEE/ANSI C9S.l standards do not
currently specify limits below 3kHz, although the responsible IEEE com-
mittee is presently at work on that problem. This section will consider
exposure limits in this frequency regime based on the excitation mecha-
nisms that are discussed above.
Threshold Criteria
Figure 11.8 illustrates asymptotic limits to S-D curves for pulsed magnetic
fields derived from principles and measurements discussed in previous
chapters. The rheobase dBldt and the S-D time constant for each curve is
given in Table 11.9. Curve (a) is based on excitation of a 10-.um nerve fiber
in the cortex of the brain, using a rheobase E-field of 12.3 Vim (Table 4.2),
along with the assumption that the brain can be represented by a sphere of
ll-cm diameter. In this case, the threshold dBldt is calculated according to
Eq. (9.16). The assumed diameter of IO.um is a rather large one within the
distribution of fibers to be found in the human brain, although a small
fraction of brain neurons may exceed this diameter (Sect. 3.3). Curve (b)
applies to the median cardiac excitation threshold and is obtained by mul-
tiplying the one-percentile threshold in Table 9.6 by a factor of2. Curve (c)
represents the threshold of a 20-.um peripheral nerve fiber as listed in Table
9.6. Curve (d) is a median threshold limit for phosphenes, which has been
derived from the sinusoidal thresholds shown in Fig. (9.20), using the
principles described below.
The pulsed field thresholds of Fig. 11.8 can be converted to sinusoidal
thresholds by applying principles and assumptions that have developed
previously. One assumption is that at the excitation threshold, the peak of
a pulsed stimulus is equivalent to the peak of a sinusoidal stimulus having a
duration of many cycles (Sect. 4.6 and Fig. 4.19). Another assumption is that
the S-D time constant, i" and the strength-frequency constant, fe, are
104 ....
103
.. .
....... :~".,,'.
.- .. - .. - .. - .. - .. - .. - .. - .. - ..
.. ~
<al 10-.m fibe,. bm;c
(j)
t:- ..
:eco . .. .••.•.•...••.•.•. (b) 50% cardiac excitation
.................................................................................
"'C
102
. ,----- ___ ~l~~~~@~~ ________ _
"'C peripheral nerve
(5
r. 10
rn
~
r.
f-- (d) phosphenes
.1 ~~~~~~~~~~~~~~~~~~~-u~
.01 .1 10
phase duration, tp (ms)
FIGURE 11.8. Strength-duration curves for stimulation by monophasic dBldt pulses,

whole body exposure, short-term reactions, large adult.
related by Ie = 1/(27:e ). A further relationship between the peak flux density,

B, and the peak dBldt is defined for a sinusoidal field in Eq. (11.6). A final
relationship is that the rms value of a sinusoid is its peak divided by "';2.
Figure 11.9, obtained by applying these relationships to Fig. 11.8, illustrates
magnetic field thresholds for sinusoidal fields. It is apparent that phos-
phenes define the lowest thresholds at frequencies below 430 Hz and that
peripheral nerve stimulation defines the lowest thresholds above that
frequency.
Limit Criteria
The thresholds depicted in Fig. 11.9 are intended to represent approximate
median values among a population of healthy individuals. To derive protec-
TABLE 11.9. Pulsed magnetic field threshold

parameters.
Eo <,
Reaction (T/s-pk) (ms)
lO-,um brain neuron excitation 448 0.12
50% cardiac excitation 98.0 3.0
20-,um peripheral nerve excitation 37.8 0.12
phosphenes 1.78 25
Eo = rheobase dB/dt; peak values listed.

ELF Magnetic Field Standards Derived from Established Mechanisms 491
r=- 102
.S-
III
:?3-
·w 10 . ..
c
Q) ...
. ,------------
""0
X
:::J
;;:::
C9S.1
.1
.01
.1 10
frequency (Hz)
FIGURE 11.9. Thresholds for short-term reactions to sinusoidal magnetic fields-

whole body exposure, large adult. Curves (a)-(d): human reaction thresholds; (e):
derived limit criteria with acceptance factor.
tive standards from thresholds, it is customary to apply an acceptability

factor to account for particularly sensitive individuals, those in a pathologi-
cal state, and for uncertainties in the methodology of determining thresh-
olds. Although we do not know the statistical distribution of thresholds
for the reactions shown in Fig. 11.9, it would not be unreasonable to
assume a log-normal distribution, similar to threshold distribution for
electrocutaneous sensitivity (Table 7.10), or cardiac excitation (Table 6.12).
Accordingly, a reasonable guess is that magnetic thresholds at the one-
percentile level would be a factor of perhaps 2 or 3 below the median for
healthy individuals. Further allowances should be made for individuals in a
pathological state, which can lead to greater population variance (e.g., Fig.
6.17). Considering these factors, an acceptability factor of 10 will be applied
to the thresholds of Fig. 11.9.
Note that the acceptability factor of 10 applied here is a conservative one
when applied to a field magnitude, rather than an SAR metric as in IEEE
C9S.1. This is true because the SAR value is proportional to the square of
the field (Eqs. 11.2, 11.9). Consequently, a factor of 10 applied to SAR is
equivalent to a factor of ~1O in the associated electric or magnetic field
under free-field propagation conditions.
If we apply a factor of 10 to the lowest thresholds in Fig. 11.9, we obtain
the acceptance curve (e) shown in the figure. At frequencies above 3 kHz,
TABLE 11.10. Sinusoidal magnetic field threshold

parameters.
Bo [,
Reaction (mT-rms) (Hz)
10-,um brain neuron excitation 12.1 4,170
50% cardiac excitation 66.2 167
20-,um peripheral nerve excitation 1.02 4,170
phosphenes 10.0 20
the acceptance curve has been merged with the C95.1 curve by extrapolat-
ing the C95.1 low frequency plateau on a slope inversely proportional to
frequency below 3kHz. With this procedure, the C95.1 plateau is a factor of
5.0 below curve (c) in the region 3 to 100kHz; below 3kHz, it is a factor 10
below curve (d). The acceptance curve below 0.2Hz has been capped at
100mT-rms (141 mT-peak). This limit is a factor of 10 below a peak field of
1.41 T, in consideration of human reactions (vertigo, taste sensations, car-
diac rate changes) reported at similar intensities of a static field (Sect. 9.10).
Table 11.11 provides a numerical description of curve (e).
The purpose of a standard is to protect against a detrimental effect, not
just a perceptible one. Consider first the peripheral nerve excitation thresh-
old curve (c) from that perspective. Although nerve stimulation is not a
disturbing experience at the threshold of perception, unpleasant or painful
sensations are experienced with magnetic stimuli that exceed the peripheral
nerve perception threshold by only 40% or so (Sect. 9.7). Consequently,
curve (e) at frequencies above 615Hz is approximately a factor of 7 below
a detrimental effect. Curve (d) is based on phosphene perception. While
this phenomenon has not been reported to be disturbing in a laboratory
setting, it is not clear whether this would be the case for affected individuals
in an uncontrolled environment. Consequently, induction of phosphenes
is assumed to be a situation that should be avoided in an uncontrolled
environment.
TABLE 11.11. Magnetic field exposure limits based on

mechanisms of short-term reactions.
Frequency range Bo H
(Hz) (mT-rms) (A/m-rms)
<0.2 100 7.95 X 104
0.2-20 201[ 1.59 x lOW
20--615 1 795
615-3 x 103 6151[ 4.89 x lO'l[
3 x 103_105 0.205 163
105_106 2.05 x 1041[ 1.63 x 10'1[
Criteria above 3kHz based on existing IEEE/ANSI C95.l.

Criteria below 3 kHz based on threshold reactions, with accept-
ability factor of 10.
Standards in Consumer Products and Installations 493
The mechanisms-based limits are compared in Fig. 11.3 with existing

standards on magnetic field exposure in the frequency 0.1 Hz to 10 kHz. The
mechanisms-based curve (e) provides conservative limits with respect to
the other standards in the frequency range 1 to 20Hz. At higher frequen-
cies, existing standards are equal to, or below curve (e).
Other Considerations for ELF Standards

The acceptance curve shown in Fig. 11.9 is derived from established
bioelectric mechanisms, as defined in Sect. 11.6. A mechanisms-based
approach is not the only basis to be used in setting electromagnetic field
acceptability standards. Other information to be considered includes epide-
miological data, and laboratory studies that are deemed reliable, but for
which a mechanistic explanation may not presently be established. Such
information must be judged as to its reliability, application to an intact
human, and evaluated hazard risk. Consequently, curve (e) should be con-
sidered as an upper limit on acceptable exposure to magnetic fields in
uncontrolled environments.
For particular applications in controlled situations, it may be acceptable
to allow greater exposure than indicated by curve (e). For instance, in
magnetic resonance imaging procedures, higher exposure levels are permit-
ted by existing guidelines (Sect. 11.3). Furthermore, with exposure to a
limited region of the anatomy, greater exposure may be tolerated. For
instance, with exposure to the head only, thresholds of peripheral nerve
excitation [curve (c) in Fig. 11.9] would be elevated considerably, and
acceptance levels in the frequency regime above 500 Hz could be increased
accordingly.
Note that the acceptance limits shown in Fig. 11.9 are hypothetical ones.
These have been included in this chapter to demonstrate a methodology for
deriving limits based on established mechanisms in the ELF regime. One
can also see that existing ELF standards (Fig. 11.3) are similar to the
mechanisms-based standards over a broad portion of the frequency range
depicted. The methodology and the derived limits for curve (e) resemble
those applied in European standards (Bernhardt, 1985; 1988).
11.8 Standards in Consumer Products and Installations

Standards Setting Agencies
In addition to the electromagnetic exposure standards discussed previously,
standards and performance guidelines have also been developed by various
agencies to protect against electric shock hazards in electrical products and
installations. For instance, the National Electric Safety Code, developed by
the IEEE and published as an ANSI Standard (IEEE, 1990), contains rules
and provisions for the installation of supply and communication lines,

equipment, and associated work practices employed by a utility for electric
supply, communication, railway, or similar activity. The IEEE also pub-
lishes as an ANSI standard a guide for the safe grounding practices in AC
substation design (IEEE, 1986). The National Electrical Code, published by
the National Fire Protection Association (NFPA, 1990) provides guidance
for the installation of electrical wiring in homes and businesses. These codes
are advisory ones from private or professional organizations without
enforcement powers. Nevertheless, the codes are widely applied, and many
enforcement agencies in North America require conformance to these
rules.
Various agencies are involved in electrical safety standards for consumer
products. For instance, trade organizations in the United States that de-
velop advisory criteria for product safety are the National Electrical Manu-
facturers Association (NEMA) and Underwriters Laboratories (UL). UL is
widely recognized since many consumer products carry the UL seal of
approval, which means that the product has been manufactured to conform
to UL safety standards. The Consumers Product Safety Commission
(CSPC) was established by Congress in 1972 under the Consumer Product
Safety Act. The main activity of the CPSC is to monitor safety problems in
consumer products, to disseminate such information, and to work with
other agencies in developing safety standards, although the agency has
written a few mandatory standards itself. Outside of North America, the
International Electrotechnical Commission (1EC) develops electrical safety
standards for European products.
Other agencies that issue guidelines and standards for consumer products
are the Center for Devices and Radiological Health (CDRH) of the Food
and Drug Administration (FDA) under Public Law 90-602 of the Radiation
Control for Health and Safety Act of 1968, and the IEEE Standards Coor-
dinating Committee-34. There is a verbal understanding that FDA does not
wish to issue more performance standards or regulations on consumer
products like color TV or cellular phones. Instead, they support SCC-34.
Current projects under SCC-34 are cellular phones, and a marine radar-
the latter is just in the planning stage.!
Safety Criteria for Consumer Products

Table 11.12 lists criteria of UL and the IEC for electrical products. In this
table, the UL data are intended as "limits" (i.e., recommended maximum
currents); the IEC criteria are described as "minimum thresholds" (i.e.,
minimum values where a physiological effect is expected to occur). The
difference between the two is that a "limit" may include an additional safety
1 Information from John Osepchuk of Full Spectrum Consulting, August, 1997.

TABLE 11.12. Conducted current criteria for consumer

products
IEe(') UL(b)
Effect (rnA-rrns) (rnA-rrns)
Startle 0.5 0.5(')
0.75(d)
Let-go 10 5
Ventricular fibrillation 35 20
Notes: (a) IEe thresholds for 15-100 Hz, per publication IEe
479; (b) UL limits for 60Hz; (c) applies to portable appliances;
(d) applies to fixed appliances.
factor applied to a "threshold." The startle current criteria apply to leakage

current from portable appliances. The intent is to protect against an electri-
cal startle reaction that may lead to injuries when engaging in inherently
dangerous activities, such as climbing a ladder, handling hot liquids, or
handling power cutting tools. The experimental basis for the UL startle
criteria is described in Sect. 7.11. The values listed under "let-go" are
intended to protect against grip tetanus with continuous SO/60Hz currents,
as described in Sects. 8.5 and 8.6. The UL limit of 5 rnA applies to an
especially sensitive small child at approximately the 0.5% percentile rank of
a distribution of let-go thresholds applying to children.
In some cases, such as with an electric fencer, the energizing voltage may
be pulsed. In such cases, grip tetanus may not be a credible hazard. Al-
though it is possible to eliminate a let-go hazard with a pulsed current, one
might still retain a cardiac fibrillation hazard if the current during the on-
time were great enough. Consequently, UL and IEC have separate ven-
tricular fibrillation criteria as shown in Table 11.12. The listed values apply
to exposure durations of one or more seconds for a small child. The weight
of a child that is just ambulatory was estimated by UL to be 8.4 kg (18.51b)
based on anthropomorphic data (Skuggevig, 1992). Applying that weight to
Dalziel's regression formula (Eq. 6.5), one obtains a median fibrillation
current of 59.4 rnA for an 8.4-kg child. And from the data of Table 6.12, the
fibrillating current at the 0.5 percentile rank is 25.5 rnA. Thus, we see that at
the 20mA UL limit, the probability of delivering a fibrillating current to a
very small child would be less than 0.5 %.
The UL limit has been developed using a body weight scaling relation-
ship. On the other hand, the European standards do not recognize a body
weight law, and instead use data on dogs to provide a conservative model
for human safety applications (see Sect. 6.7). The IEC threshold of Table
11.12 is intended as a conservatively low fibrillation value for dogs and does
not include an additional safety factor that may be applied to a safety
standard.
Hazard Criteria Vesus Current Duration and Frequency

The relationship between a hazardous current and the duration of exposure
is important in many safety applications, such as determining the speed of
operation of a current interrupting device. Figure 11.10 (adapted from
Skuggevig, 1992) illustrates UL and lEe ventricular fibrillation criteria as a
function of the duration of current. The figure shows duration from 0.1 ms
to lOs, although the criteria may be extended by another decade to the
right, and the UL curve may be extended along the same slope two addi-
tional decades to the left. The rationale for these curves is given in Sect. 6.5,
in connection with Fig. 6.12. The vertical axis represents the rms value of
pulsed SO/60Hz current during its on-time. For durations beyond 18 to
20ms, waveforms at the power frequency would include one or more indi-
vidual cycles of oscillation. For durations less than a single cycle, the pulsed
waveform could represent a single monophasic or biphasic event. For a
monophasic pulse (the worst case for brief stimuli), the rms value of the
pulse is the same as its peak value.
Figure 11.11 illustrates hazard criteria of lEe and UL as a function of
frequency for ventricular fibrillation, let-go, and startle reaction. The lEe
curves are intended as minimum reaction thresholds, and the UL curves are
recommended as maximum limits. These curves have a minimum plateau at
10 3
;;(
.s
"E
~
:;
()
10 2
Duration (5)
FIGURE 11.10. Ventricular Fibrillation criteria of UL and lEe. For durations above
approx. 20ms, criteria apply to SO/60Hz AC waveforms. Below 20ms, criteria apply
to biphasic or monophasic nonrepetitive currents.
Ventricular
Fibrillation
100
,
« 10 --------
5
Startle Reaction
lEG Threshold - - - -
UL Limit
0.1
10
Frequency (HZ)
FIGURE 11.11. Safety criteria of IEC and UL as a function of frequency. (After IEC,
1984; UL, 1988.)
low frequencies, which include 50 and 60Hz power frequencies. These

curve shapes are based on neurosensory, neuromuscular, and cardiac
muscle reactions to alternating current as described in previous chapters
(cf., Figs. 6.4, 6.8, 7.13, and 8.10)
If hazard criteria were based solely on nerve and muscle excitation, then
the hazard curves of Fig. 11.11 would rise without limit in proportion to
frequency above 1kHz or so. However, we have seen in Sect. 7.5 that tissue
heating will limit the maximum current that can be tolerated. The tempera-
ture rise that will cause tissue damage depends on the duration of the
thermal event. Figure 11.12 provides a strength-duration curve for cutane-
ous tissue damage. For instance, a temperature of 45°C must be maintained
for approximately 2 hours in order to sustain tissue damage to the porcine
skin, which is considered a good thermal model for human skin. It is no
coincidence that 45 °C is the approximate temperature at which a thermal
stimulus to the human skin will be judged as painful. Although it would
take prolonged exposure at that temperature to actually damage skin, a
painful reaction would serve as a healthy warning of an potentially noxious
event.
The heating potential of an electrical stimulus is largely determined by
the rms current, almost independent of frequency. For a small finger touch
70
65
~ 60 / Epidermal necrosis (2nd degree burn)

::J
'§
C1>
c.
E
2 55
i
::J
(j) 50
Damage threshold (1st degree burn)
45
10 100 1000 10,000

Exposure time (s)
I I (
2 4 10 20 30 60 2 4 10 20 30 60 2 4 8
, '---v-------'
Sec;nds Min~tes Hours
FIGURE 11.12. Surface temperature thresholds for human or porcine skin injury.
(From Moritz & Henriques, 1947.)
contact of 25mm2, a perceptible thermal reaction is reached at 37mA-rms,

and a painful thermal reaction is reached at about 45mA-rms over the
frequency range 100Hz to 3 MHz (see Chatterjee et al., 1986 and Fig. 7.12).
Greater thresholds would apply to larger areas of contact. As noted in
Chapter 10, the temperature rise is approximately proportional to the prod-
uct of the duration of exposure, current density-squared, and tissue resistiv-
ity. Although the duration of current was not measured in Chatterjee's
tests, one can surmise a duration of about 0.5 to 1 s from the experimental
protocol.
In order to protect against thermal burns, UL recommended a limit of
25mA rms (UL, 1981); more recently, this recommended limit has been
increased to 70mA rms-a value consistent with that recommended by the
IEC for application to electrical equipment for measurement, control, and
laboratory use (IEC, 1990a). A more conservative limit of 70mA peak has
been adopted by the IEC for application to switching of radio frequency
energy (IEC, 1986).
Limits for Capacitor Discharges

Voltage limits on capacitor discharges have been developed by UL as
shown in Fig. 11.13 (UL, 1987). These limits have been derived from the UL
ventricular fibrillation (VF) limits shown on Fig. 11.10, which can be inter-
preted in terms of the voltage limits for capacitor discharges by noting that
the charge, Q, for a monophasic stimulus is Q = It, where I is the peak
current (Amperes), t is the duration of a square wave pulse (seconds) or the
time constant of a exponential decay, and Q is in Coulombs. For instance,
at t = 1O- 4 s, the UL limit curve in Fig. 11.10 is equivalent to a charge of
400,uC. The discharge time constant for a capacitor discharge to a pure
resistance, R, is given by t = RC. In deriving capacitor discharge limits, UL
assumes R = 500Q, except for low-voltage discharges, where the resistance
is assumed to be higher. The stored charge in Fig. 11.10 is given by Q = CV.
For instance, at C = O.I,uF, the UL limit curve is equivalent to a charge of
290,uC, and a time constant of 50,us. We contrast the VF limits to perception
and pain thresholds for capacitor discharges as noted in Sects. 7.3 and 7.4.
For instance, at low capacitance values, one can perceive a capacitor dis-
charge by touching a charged electrode at about 0.25,uC (Fig. 7.4), and the
pain threshold would be at about 0.9,uC (Table 7.2).
10 4
>
(])
OJ
~
~
~
10 3 -
.8
·13
ca
0.
ca
()
10 2
10
Capacitance, C (IlF)
FIGURE 11.13. Capacitor discharge criteria. UL limits based on ventricular

fibrillation.
Ground Fault Current Interrupter

The ground-fault circuit interrupter (GFCI) is probably one of the most
significant devices to be developed in the field of electric safety. It was
introduced in the United States in the early 1960s and has since been
required by the National Electrical Code to be installed in certain branch
circuits and as part of certain products. In Europe, the equivalent device is
called a residual-current device (RCD). The RCD operates on the same
principle as the GFCI but was developed independently and with a few
differences that will be explained presently.
The two-conductor GFCI (and RCD) continuously compare the magni-
tude of current in the energized conductor to that in the grounded conduc-
tor. If the current in the two conductors differs by more than a specified
amount, the device rapidly trips and opens the circuit. Typically, a GFCI
trips in less than 25 ms. The presumption is that the current flowing through
an unintended path to ground might be through a person's body as illus-
trated in Fig. 11.14. If the person's body were to touch parts that are
connected across the line rather than between a live part and ground, the
resulting body current would flow through the same pathway as normal
electric load current, and no protection against electric shock would be
provided. However, when a person simultaneously touches a live part and
ground, the device would sense differential current, and can be effective
against most electric shock scenarios.
A GFCI acts independently from other safety mechanisms that might
also be used. When it is used to protect a product that is equipped with an
equipment grounding conductor or that is double insulated, each system
contributes to protect the user against electric shock-one system does not
interfere with or depend on the other. Being current sensitive, it protects
the user regardless of the body impedance. Most GFCIs in the United
States are rated at 5 rnA differential current and are intended to prevent
inability to let go as well as ventricular fibrillation. In Europe, an RCD with
a rating as high as 30mA is considered suitable for protection of people
against electric shock. At this rating, the protection provided by the RCD
would be against ventricular fibrillation, but not against inability to let go.
An advantage of the 30-mA-rated RCD in Europe is that it can trip
satisfactorily using only the power available in the fault current itself; it
does not need auxiliary power to open the contacts. As such, it is inherently
invulnerable to being inoperative if the grounded line conductor is inad-
vertently open. The GFCIs in the United States take power from the line at
the input to the device to open the contacts when the device is tripped. The
power available from 5mA at 120V is not sufficient by itself to cause the
rapid tripping required in this application. If the GFCI is not a type that is
2 Adapted from Skuggevig (1992).

Ground-Fault Circuit-Interrupter (GFCI)

Appliance
,-----,
I I
-'i':-;-+--,...-z.!:;;:.*:
I
A~--f---..I
Grounded Conductor
FIGURE 11.14. Ground-fault circuit interrupter (or residual-current device) sensing

body current by comparing the magnitudes of the currents in the conductors serving
the appliance. If the difference exceeds approximately 5 rnA, the ground-fault
circuit interrupter opens the circuit. (From Skuggevig, 1992.)
likely to be located where the continuity of the grounded conductor is

sufficiently reliable, the device is required by UL to protect regardless of
the continuity of the grounded conductor. Many listed GFCls satisfy this
requirement by including circuitry that will not energize the output unless
there is voltage across the input terminals.
If a GFCI is used to protect against electric shock from a product that
itself generates high voltage, such as with an internal autotransformer, it
might not trip fast enough to prevent ventricular fibrillation because of the
higher body current. For example, GFCls typically contain mechanical
contacts that interrupt the circuit when the device trips. These contacts
require a certain minimum time to separate-on the order of 15 to 20ms. If
the body impedance is as low as 500 Q, and the high-voltage source is
capable of applying 1,000 volts across the body, then the body current
would be 2A. According to the current limit as a function of duration in
Fig. 11.10, the duration of body current of this magnitude should be
limited to 264,us, which is beyond the tripping-speed capability of the
typical GFCI.
References
Abdeen, M.A, and M.A Stuchly (1994). Modeling of magnetic field stimulation of
bent neurons. IEEE Trans. Biomed. Eng. 41(11): 1092-1099.
Accornero, N., G. Bini, G.L. Lenzi, and M. Manfredi (1977). Selective activation of
peripheral nerve fibre groups of different diameter by triangular shaped stimulus
pulses. J. Physiol. 273: 539-560.
ACGIH (1996). 1996 TLVs and BEls. American Conference Governmental and
Industrial Hygienists, Cincinnati, OH.
Ackmann, J.J., and M.A. Seiti (1984). Methods of complex impedance measure-
ments in biologic tissue. eRC Crit. Rev. Bioeng. 11(4): 281-31l.
Adair, R.K. (1991). Constraints on biological effects of weak extremely-low-
frequency electromagnetic fields. Physical Rev. A 43(2): 1039-1048.
Adrian, D. (1977). Auditory and visual sensations stimulated by low-frequency
currents. Radio Sci. 12(65 S): 243-250.
Agnew, W.F., and D.B. McCreery, eds. (1990). Neural Prostheses: Fundamental
Studies. Prentice-Hall, Englewood Cliffs, NJ.
Agnew, W.F., D.B. McCreery, T.G.H. Yuen, and L.A. Bullara (1989). Histolo-
gic and physiologic evaluation of electrically stimulated peripheral nerve:
Considerations for the selection of parameters. Annals Biomed. Eng. 17: 39-
60.
Ahrenholz, D.H., W. Schubert, and L.D. Solem (1988). Creatine kinase as a prog-
nostic indicator in electrical injury. Surgery 104(4): 741-747.
Al-Rabiah, S.M., D.B. Archer, R. Millar, AD. Collins, and W.F. Shepherd (1987).
Electrical injury of the eye. Int. Ophthamol. 11(1): 31-40.
Albuquerque, E.x., and S. Thesleff (1968). A comparative study of membrane
properties of innervated and chronically denervated fast and slow skeletal
muscles of the rat. Acta. Physiol. Scand. 73: 471-480.
Alekseev, S.L, and M.e. Ziskin (1994). Millimeter microwave effect on ion trans-
port across lipid bilayer membranes. Bioelectromagnetics 16: 124-131.
Alekseev, S.L, M.e. Ziskin, N.V. Kochetkova, and M.A. Bolshakov (1997). Milli-
meter waves thermally alter the firing rate of the Lymnaea pacemaker neurons.
Bioelectromagnetics 18: 89-98.
Allessie, M.A., F.I.M. Banke, and F.J.G. Schopman (1973). Circus movement in
rabbit atrial muscle as a mechanism of tachycardia. Circ. Res. 33: 54-62.
Allessie, M.A., F.LM. Banke, and F.J.G. Schopman (1976). Circus movement in
rabbit atrial muscle as a mechanism of tachycardia. II. The role of nonuniform
502
References 503
recovery of excitability in the occurrence of unidirectional block, as studied with

multiple microelectrodes. Circ. Res. 39: 168-177.
Allessie, M.A, F.I.M. Bonke, and F.J.G. Schopman (1977). Circus movement in
rabbit atrial muscle as a mechanism of tachycardia. III. The "leading circle"
concept: A new model of circus movement in cardiac tissue without the involve-
ment of an anatomical obstacle. Circ. Res. 41: 9-18.
Alon, G., J. Allin, and G.F. Inbar (1983). Optimization of pulse duration and pulse
charge during transcutaneous electrical nerve stimulation. Austral. 1. Physio-
therapy 29(6): 195-20l.
Altman, K.W., and R Plonsey (1989). Excitation in a model for time-dependent
electrical nerve bundle stimulation. Proc. 11th IEEE-EMBS Conf: 975-976.
Amassian, V.E., RO. Cracco, and P.J. Maccabee (1988). Basic mechanisms of
magnetic coil excitation of nervous system in humans and monkeys and their
applications. Proc. 10th Ann. Conf IEEE-EMBS: 10-17.
Amassian, V.E., J.B. Cracco, RO. Cracco, L. Eberle, P.J. Maccabee, and A Rudell
(1989a). Suppression of visual perception by magnetic coil stimulation of human
occipital cortex. EEG Clin. Neurophysiol. 74: 458-462.
Amassian, V.E., RO. Cracco, and P.J. Maccabee (1989b). A sense of movement
elicited in paralyzed distal arm by focal magnetic coil stimulation of human motor
cortex. Brain Res. 479: 355-360.
Anderson, AB., and W.A Munson (1951). Electrical stimulation of nerves in the
skin at audio frequencies. 1. Acoust. Soc. Am. 23(2): 155-159.
Anderson, N.H. (1970). Functional measurement and psychophysical judgment.
Psycho I. Rev. 77: 153-170.
ANSI (1982). Safety Levels with Respect to Human Exposure to Radio Frequency
Electromagnetic Fields, 300kHz to 100GHz. Document ANSI C95.1-1982,
published by The Institute of Electrical and Electronics Engineers, New York.
ANSI (1986). Leakage current for appliances. ANSI C1Ol.1-1986, American
National Standards Institute.
ANSI (1987). Procedures for the development and coordination of American
national standards. American National Standards Institute.
Antoni, H. (1979). What is measured by the so-called threshold for fibrillation?
Prog. Pharmacol. (Stuttgart) 2(4): 5-12.
Antoni, H. (1985). Pathophysiological basis of ventricular fibrillation. In J.F.
Bridges, G.L. Ford, LA Sherman, and M. Vainberg (eds.), Electrical Shock Safety
Criteria, Pergamon, New York: 33-43.
Antoni, H. (1996). Electrophysiology of the heart at the single cell level and cardiac
rhythmogenesis. In R Greger and U. Windhorst (eds.), Comprehensive Human
Physiology. Springer-Verlag, Berlin: 1825-1842.
Antoni, H., J. Toppler, and R Krause (1970). Polarization effects of sinusoidal 50-
cycle alternating current on membrane potential of mammalian cardiac fibers.
Pflugers Arch. 314: 274-29l.
Antzelevitch, c., and G.K. Moe (1981). Electrotonically mediated delayed con-
duction and reentry in relation to "slow responses" in mammalian ventricular
conducting tissue. Circ. Res. 49: 1129-1139.
Artz, c.P. (1974). Changing concepts of electrical injury. Am. 1. Surg. 128(5): 600-
602.
Artz, c.P. (1979). Electrical injury. In C.P. Artz, J.A Moncrief, and B.A Pruitt,
(eds.), Burns: A Team Approach, W.B. Saunders, Philadelphia.
504 References
ASAE (1993). Dimensions of livestock and poultry, ASAE Standards D32l.2,

American Soc. Ag. Eng., st. Joseph, MI: 511-517.
Ashton, W.D. (1972). The Logit Transformation. Hafner, New York.
Astumian, RD. et al. (1995). Rectification and signal averaging of weak electric
fields by biological cells. Proc. Nat!. Acad. Sci. 92: 3740-3743.
Athey (1992). Current FDA guidance for MR patient exposure and consideration
for the future. In RL. Magin, RL. Liburdy, and B. Persson (eds.), Biological
Effects and Safety Aspects of Nuclear Magnetic Resonance Imaging, New York
Academy of Sciences, New York: 242-257.
Atkinson, W.H. (1982). A general equation for sensory magnitude. Perception &
Psychophys. 31: 26--40.
Attneave, F. (1962). Perception and related areas. In S. Koch (ed.), Psychology: A
Study of a Science, McGraw-Hill, New York: 619-659.
Babkoff, R. (1976). Magnitude estimation of short electrocutaneous pulses. Psycho!.
Res. 39: 39--49.
Baeten, C.G.M.I., B.P. Geerdes, E.M.M. Adang, et al. (1995). Anal dynamic
graciloplasty in the treatment of intractable fecal incontinence. N. Engl. 1. Med.
332: 1600-1605.
Bajzek, T.J., and RJ. Jaeger (1987). Characterization and control of muscle
response to electrical stimulation. Annals Biomed. Eng. 15: 485-50l.
Baker, M.D., and C. Chiaviello (1989). Household electrical injuries in children;
Epidemiology and identification of avoidable hazards. Am. 1. Dis. Child. 143(1):
59-62.
Banks, R.S., and T. Vinh (1984). An assessment of the 5-mA 60-Hz contact current
safety level. IEEE Trans. Pwr. Sys. PAS-103(12): 3608-3614.
Barker, AT. (1994). Magnetic nerve stimulation principles, advantages, and
disadvantages. In S. Veno (ed.), Biomagnetic Stimulation, Plenum, New York: 9-
28.
Barker, AT., I.L. Freeston, R Jalinous, P.A Merton, and H.B. Morton (1985).
Magnetic stimulation of the human brain. 1. Physiol. 369: 3P.
Barker, AT., I.L. Freeston, B. Jalinous, and J.A Jarratt (1987). Magnetic stimula-
tion of the human brain and peripheral nervous system: An introduction and the
results of an initial clinical evaluation. Neurosurgery, 20: 100-109.
Barker, AT., C.W. Garnham, and I.L. Freeston (1991). Magnetic nerve stimulation:
The effect of waveform on efficiency, determination of neural membrane time
constant, and the measurement of stimulator output. In W.J. Levy, RQ. Cracco,
AT. Barker, and J. Rothwell (eds.), Magnetic Motor Stimulation: Basic Principles
and Clinical Experience (EEG Supplement 43): 227-237.
Barlow, H.B., H.I. Kohn, and E.G. Walsh (1947a). Visual sensations aroused by
magnetic fields. Am. 1. Physiology 148: 372-375.
Barlow, H.B., H.1. Kohn, and E.G. Walsh (1947b). The effect of dark adaptation
and of light upon the electric threshold of the human eye. Am. 1. Physiol. 148:
376-38l.
Barnard, M.D., and J.A Boswick, Jf. (1976). Electrical injuries of the upper extrem-
ity. Rocky Mt. Med. 1. 73: 20-24.
Barnes, F.S. (1996). Interaction of DC and ELF electric fields with biological mate-
rials and systems. In C. Polk and E. Postow (eds.), Handbook of Biological Effects
of Biological Effects of Electromagnetic Fields, Second Edition, CRC Press, Boca
Raton, FL: 107-147.
References 505
Barta, E., D. Adam, E. Salant, and S. Siderman (1987). 3-D ventricular myocardial
electrical excitation: A minimal orthogonal pathways model. Annals Biomed.
Eng. 15: 443-456.
Basser, P.J., and B.J. Roth (1991). Stimulation of a myelinated nerve axon by
electromagnetic induction. Med. Bioi. Eng. Compo 29(3): 261-268.
Baumgardt, E. (1951). Sur Ie seuil phosphene electrique. Quantite liminaire et
pseudo-chronaxie. Comptes Rendeu Soc. Bioi. (Nov. 24): 1654-1657.
Bauwens, P. (1971). Introduction to electro diagnostic procedures. In S.H. Licht
(ed.), Electrodiagnosis and Electromyography (3rd ed.), E. Licht, New Haven,
CT, chap. 7.
Bawin, S.M., A.R. Sheppard, M.D. Mahoney, and W.R. Adey (1984). Influences
of sinusoidal electric fields on excitability in the rat hippocampal slice. Brain
Research, 323: 227-237.
Bawin, S.M., AR. Sheppard, M.D. Mahoney, M. Abu-Assai, and W.R. Adey
(1986). Comparison between the effects of extracellular direct and sinusoidal
currents on excitability in hippocampal slices. Brain Res. 362: 350-354.
Baxter, CR. (1970). Present concepts in the management of major electrical injury.
Surg. C/in. North Am. 50(6): 1401-1418.
Beck, CS., W.H. Pritchard, and S.H. Feil (1947). Ventricular fibrillation of long
duration abolished by electric shock. f. Am. Med. Assoc. 135: 985-995.
Beck, C, and B.S. Rosner (1968). Magnitude scales and evoked potentials to percu-
taneous electrical stimulation. Physiol. Behav. 3: 947-953.
Beeler, G.W., and H. Reuter (1977). Reconstruction of the action potential of
ventricular myocardial fibers. 1. Physiol. 268: 177-210.
Benz, R. and U. Zimmerman (1980). Relaxation studies on cell membranes and
lipid bilayers. Bioelectrochem. Bioenerg. 7: 723-739.
Bergeron, J.A, M.R. Hart, J.A. Mallick, and L.H. String (1995). Strength-duration
curve for human electro- and magnetophosphenes. Proc. Bioelectromagnetics
Soc. Annual Meeting, Boston.
Bernhardt, J.H. (1985). Evaluation of human exposures to low frequency fields. In
The Impact of Proposed Frequency Radiation Standards on Military Operations,
Lecture Series 138, Advisory Group for Aerospace Research and Development
(NATO), Surseine, France.
Bernhardt, J.H. (1988). The establishment of frequency dependent limits for electric
and magnetic fields and evaluation of indirect effects. Radiat. Env. Biophys. 27:
1-27.
Biegelmeier, G. (1978). Report on the electrical impedance of the human body.
Presented to the International Electrotechnical Commission, Technical Commit-
tee No. 23, Austria.
Biegelmeier, G. (1982). Report on the electrical impedance of the human body.
IEC Technical Committee No. 64, Electrical Installations of Buildings, Austin,
Texas.
Biegelmeier, G. (1985a). New knowledge of the impedance of the human body. In
J.E. Bridges, G.L. Ford, LA Sherman, and M. Vainberg (eds.), Electrical Shock
Safety Criteria, Pergamon, New York: 115-132.
Biegelmeier, G. (1985b). New experiments with regard to basic safety measures for
electrical equipment and installations. In J.E. Bridges, G.L. Ford, LA Sherman,
and M. Vainberg (eds.), Electrical Shock Safety Criteria, Pergamon, New York:
161-172.
506 References
Biege1meier, G. (1985c). The impedance of the human body. Rev. Gen. L'Electricite
11: 817-832.
Biegelmeier, G. (1986). Wirkungen des Elektrischen Stroms auf Menschen und
Nutztiere. Vde-Verlag, Berlin.
Biegelmeier, G. (1987). Effects of current passing through the human body and the
electrical impedance of the human body: A guide to IEC-Report 469. ETZ
Report 20, VDE-Verlag, Berlin.
Biegelmeier, G., and E. Hornberger (1982). Uber die wirkungen von unipo-
laren impulsatromen auf den menschlicken korper. Bulletin ASEIUSE 73, Sept.
18.
Biegelmeier, G., and W.R Lee (1980). New considerations on the threshold of
ventricular fibrillation for a.c. shocks at 50-60Hz. lEE Proc. 127(2), Pt. A: 103-
110.
Biegelmeier, G., and J. Miksch (1980). Effect of the skin on the body impedance of
humans (translated from German). Electrotechnik and Maschinenbau, mit
lndustrieller Electronik und Nachrichtentchnik 97(9): 369-378.
Biegelmeier, G., and K. Rotter (1971). Electrical resistances and currents in the
human body. Electrotechnic and Maschinenbau: 104-114.
Bigland-Ritchie, B., R Johansson, O.c. J. Lippold, S. Smith, and J.J. Woods (1983).
Changes in motoneuron firing rate during sustained maximal voluntary contrac-
tions. 1. Physiol. 340: 335-346.
Bigland-Ritchie, B., F. Bellemare, and J.J. Woods (1986). Excitation frequencies
and sites of fatigue. In N.L. Jones, N. McCartney, and AJ. McComas (eds.),
Human Muscle Power, Human Kinetics Publishers, Champaign, IL: 197-211.
Bini, G., G. Cruccu, K.E. Hagbarth, W. Schady, and E. Torebjork (1984). Analgesic
effect of vibration and cooling on pain induced by intraneural electrical stimula-
tion. Pain 18: 239-248.
Birks, R, RE. Huxley, and B. Katz (1960). The fine structure of the neuromuscular
junction of the frog. 1. Physiol. 150: 134-144.
Bishop, G.H. (1943). Responses to electrical stimulation of single sensory units of
skin. 1. Neurophysiol. 6: 361-382.
Bishop, G.H. (1946). Neural mechanisms of cutaneous sense. Psychol. Rev. 26: 77-
102.
Blackman, C.F., et al. (1988). Influence of Electromagnetic fields on the efflux of
calcium ions from brain tissue in vitro. Bioelectromagnetics 9: 215-227.
Blackman, C.F., J.P. Blanchard, S.G. Benane, and D.E. House (1994). Empirical
tests of an ion parametric resonance model for magnetic field interactions with
PC-12 cells. Bioelectromagnetics 15: 239-260.
Blair, E.A, and J. Erlanger (1933). A comparison of the characteristics ofaxons
through their individual electrical responses. Am. 1. Physiol. 106: 524-564.
Blair, H.A (1932a). On the intensity-time relations for stimulation by electric
currents. 1.1. Gen. Physiol. 15: 709-729.
Blair, H.A (1932b). On the intensity-time relations for stimulation by electric
currents. II. 1. Gen. Physiol. 15: 731-755.
Blanchard, J.P. and C.F. Blackman (1994). Clarification and application of an ion
parametric resonance model for magnetic field interactions with biological
systems. Bioelectromagnetics 14: 217-238.
Blank, M. (ed.) (1993). Electricity and Magnetism in Biology and Medicine. San
Francisco Press, San Francisco.
References 507
Bo, WJ., N.T. Wolfman, W.A. Krueger, and L Meschan (1990). Basic Atlas of
Sectional Anatomy with Correlated Imaging. 2nd edition, W.B. Saunders,
Philadelphia.
Bostock, H. (1983). The strength-duration relationship for excitation of myelinated
nerve: Computed dependence on membrane parameters. f. Physiol. 341: 59-74.
Bostock, H., T.A. Sears, and RM. Sherratt (1983). The spatial distribution of
excitability and membrane current in normal and demyelinated mammalian
nerve fibers. f. Physiol. 341: 41-58.
Bouman, M.A., and RA. van der Velden (1947). The two-quanta explanation of
the dependence of the threshold values and visual acuity on the visual angle and
the time of observation, f. Opt. Soc. Am. 37: 908-919.
Bourland, J.D., W.A. Tacker, and L.A. Geddes (1978). Strength-duration curves
for trapezoidal waveforms of various tilts for transchest defibrillation in animals.
Med. Instrum. 12: 38-41.
Bourland, J.D., J.A. Nyenhuis, G.A. Mouchawar, L.A. Geddes, and D.J. Schaefer
(1990). Human peripheral nerve stimulation from z-gradients. Society for Mag-
netic Resonance in Med., Proc. 9th Annual Meeting, Aug. 18-24, New York:
1157.
Bourland, J.D., J.A. Nyenhuis, G.A. Mouchawar, L.A. Geddes, D.J. Schaefer and
M.E. Riehl (1991a). Z-gradient coil and eddy-current stimulation of skeletal and
cardiac muscle in the dog. Society for Magnetic Resonance in Med., Proc. 10th
Annual Meeting, Aug. 10-16, San Francisco.
Bourland, J.D., J.A. Nyenhuis, G.A. Mouchawar, T.Z. Elabbady, L.A. Geddes, D.J.
Schaefer, and M.E. Riehl (1991b). Physiologic indicators of high MRI gradient-
induced fields. Society for Magnetic Resonance in Med., Proc. 10th Annual
Meeting, San Francisco: 1276.
Bourland, J.D., J.A. Nuyenhuis, G.A. Mouchawar, L.A. Geddes, D.J. Schaefer, and
M.E. Riehl (1991c). Effect of body position on z-gradient coil cardiac stimulation.
Cardiac arrhythmias induced in the dog. Society for Magnetic Resonance in Med.,
Proc. 10th Annual Meeting, Berkeley, San Francisco.
Bourland, J.D., J.A. Nyenhuis, D.J. Schaefer, K.S. Foster, W.E. Schoelein, T.Z.
Elabbady, L.A. Geddes, and M.E. Rielh (1992). Gated, gradient-induced cardiac
stimulation in the dog: Absence of ventricular fibrillation. Society for Magnetic
Resonance in Med., Proc. 11th Annual Meeting, Berlin: 4804.
Bourland, J.D., J.A. Nyenhuis, W.A. Noe, DJ. Schaefer, K.S. Foster, and L.A.
Geddes (1996). Motor and sensory strength-duration curves for MRI gradient
fields. Proc. Soc. Mag. Res. Med., 4th Annual. Meeting, NY, Apr. 27-May 3: 1724.
Bourland, J.D., J.A. Nyenhuis, K.S. Foster, L.A. Geddes (1997). Threshold and pain
strength-duration curves for MRI gradient fields. Proc. Soc. Mag. Res. Med., 5th
Ann. Meeting, Vancouver, Apr. 12-18: 1974.
Bowman, B.R., and R.C. Erickson (1985). Acute and chronic implantation of coiled
wire interaneural electrodes during cyclical electrical stimulation. Annals.
Biomed. Eng. 13: 75-93.
Bowman, B.R., and D.R McNeal (1986). Response of single alpha motoneurons
to high-frequency pulse trains. Appl. Neurophysiol. 49: 121-138.
Boxtel, A. (1977). Skin resistance during square-wave electrical pulses of 1 to
lOrnA. Med. Bioi. Eng. Compo 15: 679-687.
Boyd, LA., and M.R Davey (1968). Composition of Peripheral Nerves, E & S
Livingstone Ltd, Edinburgh.
508 References
Bracken, T.D. (1976). Field measurements and calculations of electrostatic effects

of overhead transmission lines. IEEE Trans. Pwr. Apparat. Sys. PAS-95: 494-502.
Brazier, M.A (1977). Electrical Activity of the Nervous System. Williams & Wilkins,
Baltimore.
Brengelmann, G., and AC Brown (1965). Temperature regulation. In T.C Ruch
and H.D. Patton (eds.), Physiology and Biophysics, 19th ed., W.B. Saunders,
Philadelphia: 1050-1068.
Bridges, J.E. (1985). Potential distributions in the vicinity of the hearts of primates
arising from 60Hz limb-to-limb body currents. In J.E. Bridges, G.L. Ford, LA
Sherman, and M. Vainberg (eds.), Electrical Shock Safety Criteria, Pergamon,
New York: 61-70.
Bridges, J.E., G.L. Ford, LA Sherman, and M. Vainberg (eds.) (1985). Electrical
Shock Safety Criteria, Pergamon, New York.
Bridges, J.E., M. Vainberg, and M.C Wills (1987). Impact of recent developments
in biological electrical shock safety criteria. IEEE Trans. Pwr. Del. PWRD-2(1):
238-248.
Brindley, G.S. (1955). The site of exectrical excitation of the human eye. 1. Physiol.
127: 189-200.
Brindley, G.S., and W.S. Lewin (1968). The sensations produced by electrical stimu-
lation of the visual cortex. J. Physiol. 196: 479-493.
Brindley, G.S., and D.N. Rushton (1977). Observations of the representation of the
visual field of the human occipital cortex. In F.T. Hambrecht and J.B. Reswick,
Marcel Dekker, New York: 261-276.
Brooke, M.H., and K.K. Kaiser (1970). Muscle fiber types: How many and what
kind? Arch. Neurol. 23: 369-379.
Brooks, CM., B.F. Hoffman, E.E. Suckling, and 0.0. Orias (1955). Excitability of
the Heart. Grune & Stratton, New York.
Budinger, T.F. (1992). Emerging nuclear magnetic resonance technologies: Health
and safety. In RL. Magin, RL. Liburdy, and B. Persson (eds.), Biological Effects
and Safety Aspects of Nuclear Magnetic Resonance Imaging, New York Academy
of Sciences, New York (1950).
Budinger, T.F., C Cullander, and R Bordow (1984). Switched magnetic field
thresholds for the induction of magnetophosphenes. Proc. Annual Meet. Soc.
Mag. Res. Med., New York: 118.
Budinger, T.F., R Fischer, D. Hentschel, RE. Reinfelder, and F. Schmitt (1990).
Neural Stimulation dB/dt thresholds for frequency and number of oscillations
using sinusoidal magnetic gradient fields. Society for Magnetic Resonance in Med.,
Proc. 9th Annual Meeting, Aug, 18-24, NY: 276.
Budinger, T.F., H. Fischer, D. Hentschel, H. Reinfelder, and F. Schmitt (1991).
Physiological effects of fast oscillating magnetic field gradients. J. Computer As-
sisted Tomography 15(6): 909-904.
Buechler, D.N., Durney, CH., and Christensen, D.A (1997). Calculation of electric
fields induced near metal implants by magnetic resonance imaging switched-
gradient magnetic field. Magnetic Resonance Imaging.
Burke, RE. (1968). Firing patterns of gastrocnemius motor units in the decerebrate
cat. J. Physiol. 196: 631-654.
Burke, RE. (1981). Motor units: Anatomy, physiology and functional organization.
In V.B. Brooks (ed.), Handbook of Physiology Section 1: The Nervous System.
Vol. III. Motor Systems, American Physiology Society, Bethesda, MD: 345-422.
References 509
Burke, RE. (1986). The control of muscle force: Motor unit recruitment and firing
patterns. In N.L. Jones, N. McCartney, and AJ. McComas (eds.), Human Muscle
Power, Human Kinetics Publishers, Champaign, IL: 97-106.
Burke, RE., D.N. Levine, P. Tsairis, and F.E. Zajac (1973). Physiological types of
histochemical profiles in motor units of the cat gastrocnemius. 1. Physiol. 234:
723-748.
Burton, CE., RM. David, W.M. Portnoy, and L.A Akers (1974). The application
of Bode analysis to skin impedance. Psychophysiology 11(4): 517-525.
Btitikoffer, R., and P.D. Lawrence (1978). Electrocutaneous nerve stimulation-I:
Model and experiment. IEEE Trans. Biomed Eng. BME-25(6): 526-53l.
Btitikoffer, R., and P.D. Lawrence (1979) Electrocutaneous nerve stimulation-II:
Stimulus waveform selection. IEEE Trans Biomed Eng. BME-26(2): 69-75.
Butler, E.D., and T.D. Gant (1978). Electrical injuries, with special reference to the
upper extremities-A review of 182 cases. Am. J. Surg. 134: 95-1Ol.
Cabanes, J., and C Gary (1981). La perception directe du champ electrique. Paper
233-08. CIGRE Symp. 22-81, Stockholm: 1-6.
Caldwell, CW., and J.B. Reswick (1975). A percutaneous wire electrode for chronic
research use. IEEE Trans, Biomed. Eng. 22: 429-432.
Campbell, J.N., and RA Meyer (1983). Sensitization of unmyelinated nociceptive
afferents in monkey varies with skin type. 1. Neurophysiol. 49(1): 98-110.
Campbell, J.N., R.A Meyer, and R.H. LaMotte (1979). Sensitization of myelinated
afferents that innervate monkey hand. J. Neurophysiol. 42(6): 1669-1679.
Carmeliet, E. (1977). Repolarization and frequency in cardiac cells. J. Physiol.
(Paris). 73: 903-923.
Carmeliet, E. (1992). Potassium channels in cardiac cells. Cardiovasc. Drugs Ther. 6:
305-312.
Carpentier, A and J.C Chachques (1985). Myocardial substitution with a stimu-
lated skeletal muscle: First successful clinical case. Lancet 8440: 1267.
Carpentier, A, J.C Chachques, and P.A Grandjean (eds.) (1991). Cardio-
myoplasty. Futura Publishing, Mount Kisco, NY.
Carstensen, E.L. (1985). Sensitivity of the human eye to power frequency electric
fields. IEEE Trans. Biomed. Eng. BME-32(8): 561-565.
Carstensen, E.L. (1987). Biological Effects of Transmission Line Fields. Elsevier,
New York.
Carter, AO., and R Morley (1969a). Electrical current flow through human skin at
power frequency voltages. Br. J. Ind. Med. 26: 217-223.
Carter, AO., and R Morley (1969b). Effects of power frequency voltages on
amputated human limb. Br. J. Ind. Med. 26: 224-230.
Carter, B.L., J. Morehead, S.M. Wolpert, S.B. Hammerschlag, H.J. Griffiths, and
P.e. Kahn (1977). Cross-Sectional Anatomy: Computed Tomography and Ultra-
sound Correlation. Appleton-Century-Crofts, New York.
Caruso, P.M., J.A Pearce, and D.P. DeWitt (1979). Temperature and current
density distributions at electro surgical dispersive electrode sites. Proc. 7th N.
Engl. Bioeng. Conf, Troy, NY, March 22-23,1979: 373-376.
Cattell, M.K., and R.W. Gerard (1935). The "inhibitory" effect of high frequency
stimulation and excitation state of nerve. 1. Physiol. 83: 407-415.
CENELEC (1995). Human exposure to electromagnetic fields: Low-frequency
(0 to 10kHz). European Pre standard ENV 50166-1: 1995 E, European Commit-
tee for Electrotechnical Standardization, Brussels.
510 References
Chachques, J.e., M. Radermercker, J. Tolan, E.I.e. Fischer, P.A Grandjean, and

AF. Carpentier (1996). Aortomyoplasty counterpulsation: Experimental results
and early clinical experience. Ann. Thorac. Surg. 61: 420-425.
Chakravarti, K., and G.J. Pontrelli (1976). The measurement of carpet static. Textile
Res. 1. 46(2): 129-134.
Chatterjee, I., D. Wu, and a.p. Gandhi (1986). Human body impedance and thresh-
old currents for perception and pain for contact hazard analysis in the VLF-MF
band. IEEE Trans. Biomed. Eng. BME-33(5): 486-494.
Chen, J.Y., and a.p. Gandhi (1988). Thermal implications of high SAR's in the
body extremities at the ANSI-recommended MF-VHF safety levels. IEEE Trans.
Biomed. Eng. 35(6): 435-44l.
Chen, P., G.H. Myers, V. Parsonnet, K. Chatterjee, and P. Katz (1975). Relationship
between pacemaker fibrillation thresholds and electrode area. Med. Instrum. 9(4):
165-170.
Chilbert, M.A, A Sances, J.B. Mykelbust, T. Swiontek, and T. Prieto (1983). Post-
mortem resistivity studies at 60Hz. 1. Clin. Eng. 8(3): 219-224.
Chilbert, M., D. Maiman, A Sances Jr., J. Myklebust, T.E. Prieto, T. Swiontek, M.
Heckman, and K. Pintar (1985a). Measure of tissue resistivity in experimental
electrical burns. 1. Trauma 25(3): 209-215.
Chilbert, M., T. Swiontek, T. Prieto, A Sances, J. Myklebust, J. Ackmann, e.
Brown, and J. Szablya (1985b). Resistivity changes of tissue during the applica-
tion of injurious 60 Hz currents. In J.E. Bridges, G.L. Ford, I.A Sherman, and M.
Vainberg (eds.), Electrical Shock Safety Criteria, Pergamon, New York: 193-20l.
Chilbert, M., D.J. Moretti, T. Swiontek, J.B. Myklebust, T. Prieto, A Sances, and
e. Leffingwell (1988). Instrumentation design for high-voltage electrical injury
studies. IEEE Trans. Biomed. Eng. 35(7): 565-568.
Chilbert, M., T. Swiontek, e. Leffingwell, T. Prieto, J. Myklebust, A Sances, and
T. Schneider (1989). Fibrillation induced at high current levels. IEEE Trans.
Biomed. Eng. 36(8): 864-868.
Chiu, M.e. (1982). Fuel ignition by high voltage capacitive discharges. The Johns
Hopkins University Applied Physics Laboratory, JHU PPSE T-18.
Chiu, S.Y., and J.M. Ritchie (1981). Evidence for the presence of potassium chan-
nels in the paranodal region of acutely demyelinated mammalian single nerve
fibers. 1. Physiol. 313: 415-437.
Chiu, S.Y., J.M. Ritchie, R.B. Rogart, and D. Stagg (1979). A quantitative descrip-
tion of membrane currents in rabbit myelinated nerve. 1. Physiol. 292: 149-166.
Chizeck, H.J., R. Kobetic, E.B. Marsolais, J.J. Abbas, I.R. Donner, and E. Simon
(1988). Control of functional neuromuscular stimulation systems for standing and
locomotion in paraplegics. Proc. IEEE 76(9): 1155-1165.
Chokroverty, S., and J. DiLullo (1989). Percutaneous magnetic stimulation of the
human lumbosacral spinal column: Physiological mechanism and clinical applica-
tion. Neurology 39: 376.
Christensen, J.A, R.T. Sherman, G.A Balis, and J.D. Wuamett (1980). Delayed
neurologic injury secondary to high-voltage current, with recovery. 1. Trauma
20(2): 166-168.
Clairmont, B.A, G.B. Johnson, L.E. Zafanella, and S. Zelingher (1989). The effect
of HVAC-HVDC line separation in a hybrid corridor. IEEE Trans. Pwr. Del.
4(2): 1338-1350.
Clar, E.J., e.P. Her, and e.G. Sturelle (1975). Skin impedance and moisturization.
1. Soc. Cosmet. Chem. 26: 337-353.
References 511
Clark, W.e., and S.B. Clark (1980). Pain responses in Nepalese porters. Science 209:
410-412.
Clerc, L. (1976). Direction differences of impulse spread in trabecular muscle from
mammalian heart. J. Physiol. (London). 255: 335-346.
Coburn, B. (1989). Neural modeling in electrical stimulation. Critical Reviews in
Biomed. Eng. 17(2): 133-178.
Coers, e. (1955). Les variations structurelles norm ales et pathologiques de la
jonction neuromusculaire. Acta Neurol. Psychiatr. Belg. 55: 741.
Cohen, M.S., RM. Weisshoff, RR. Rzedzian, and H.e. Kantor (1990). Sensory
stimulation by time-varying magnetic fields. Magnetic Resonance in Medicine 14:
409-414.
Cook, M.R., e. Graham, H.D. Cohen, and M. Gerkovich (1992). A replication study
of human exposure to 60Hz fields. Bioelectromagnetics 13: 261-285.
Cooley, J.W., and F.A. Dodge (1966). Digital computer solutions for excitation and
propagation of the nerve impulse. Biophys. 1. 6: 583-599.
Cowburn, J., and RH. Fox (1974). A technique for studying thermal perception.
J. Physiol. (London). 239: 77-78.
Cracco, RO. (1987). Evaluation of conduction in central motor pathways: Tech-
niques, pathophysiology, and clinical interpretation. Neurosurgery 20: 199-203.
Craggs, J.D. (1978). High-frequency breakdown of gases. In J.M. Meek and J.D.
Craggs (eds.), Electrical Breakdown of Gases. John Wiley & Sons, Chichester,
UK: 689-715.
Crago, P.E., P.R Peckham, J.T. Mortimer, and J.P. Van Der Meulen (1974). The
choice of pulse duration for chronic electrical stimulation via surface, nerve, and
intramuscular electrodes. Annals Biomed. Eng. 2: 252-264.
Crago, P.E., P.R. Peckham, and G.B. Thorpe (1980). Modulation of muscle force
by recruitment during intramuscular stimulation. IEEE Trans. Biomed. Eng. 27:
679-684.
Crago, P.E., N. Lan, P.H. Veltink, J.1. Abbas, and e. Kantor (1996). New control
strategies for neuroprosthetic systems. Rehab. Res. Develop. 33(2): 158-172.
Craig, K.D., and S.M. Weiss (1971). Vicarious influences on pain-threshold determi-
nations. J. Personality and Social Psychol. 19: 53-59.
Cranefield, P.F., B.F. Hoffman, and A.A. Siebens (1957). Anodal excitation of
cardiac muscle. Am. J. Physiol. 190(2): 383-390.
Cranefield, P.F., RO. Klein, and B.F. Hoffman (1971). Conduction of the cardiac
impulse: I. Delay, block, and one-way block in depressed Purkinje fibers. Circ.
Res. 28: 199-219.
Creasey, G., J. Elifteriades, A. DiMarco, P. Talonen, M. Bijak, W. Girsch, and
e. Kantor (1996). Electrical stimulation to restore respiration. J. Rehab. Res.
Develop. 33(2): 123-132.
Critchley, M. (1934). Neurological effects of lightning and of electricity. Lancet. 1:
68-72.
Crochetiere, W.J., L. Vodovnik, and J.B. Reswick (1967). Electrical stimulation of
skeletal muscle-A study of muscle as an actuator. Med. BioI. Eng. 5: 111-125.
Currence, H.D., B.J. Stevens, D.F. Winter, W.K. Dick, and G.F. Krause (1990).
Dairy cow and human sensitivity to short duration 60Hz currents. App. Eng. in
Agriculture, 6(3): 349-353.
Dalziel, e.F. (1938). Danger of electric shock. Electrical West 80(4): 30-31.
Dalziel, e.F. (1943). Effect of wave form on let-go currents, AlEE Trans. 62: 739-
744.
512 References
Dalziel, c.F. (1953). A study of the hazards of impulse current. Trans. AlEE, Pt. III,
72: 1032-1043.
Dalziel, C.F. (1954). The threshold of perception current. Trans. AlEE, Pt. III B:
990-996.
Dalziel, C.F. (1959). The effects of electric shock on man. IRE Trans. Med. Elect.
PGME-5: 44-62.
Dalziel, C.F. (1960). Threshold 60-cycle fibrillating currents. Trans. AlEE, Pt. III,
79: 667-673.
Dalziel, C.F. (1968). Reevaluation of lethal electric currents. IEEE Trans. Ind.
Appl. IGA-4(5): 467-476.
Dalziel, C.F. (1972). Electric shock hazard. IEEE Spectrum (9): 41-50.
Dalziel, C.F. (1978). Recent developments in ground fault circuit interrupters and
ground fault receptacles. Profession. Safety, November: 31-40.
Dalziel, C.F., and W.R. Lee (1968). Reevaluation of lethal electric currents. IEEE
Trans. Ind. Appl. IGA-4(5): 467-476.
Dalziel, C.F., and W.R. Lee (1969). Lethal electric currents. IEEE Spectrum, Feb.:
44-50.
Dalziel, C.F., and T.R. Mansfield (1950a). Perception of electric currents. Electrical
Eng. 69: 794-800.
Dalziel, C.F., and T.R. Mansfield (1950b). Effects of frequency on perception
currents. AlEE Trans. 69: 1162-1168.
Dalziel, C.F., and F.P. Massoglia (1956). Let-go currents and voltages. AlEE Trans.
P. II, Appl. Ind. 75: 49-56.
Dalziel, C.F., J.B. Lagen, and J.L. Thurston (1941). Electric shock. AlEE Trans. 60:
1073-1079.
Dalziel, C.F., E. Ogden, and C.E. Abbott (1943). Effect of frequency on let-go
currents. AlEE Trans. 62: 745-750.
Darien-Smith, I. (1982). Touch in primates. Ann. Rev. Psychol. 33: 155-194.
Darien-Smith, I., K.O. Johnson, C. LaMotte, P. Kenins, Y. Shigenaga, and V.C.
Ming (1979). Coding of incremental changes in skin temperature by single warm
fibers in the monkey. 1. Neurophysiol. 42: 1316-1331.
Davey, K., K.c. Kalaitzakis, and C. Epstein (1988). Transcranial magnetic stimula-
tion of the cerebral cortex. Proc. 10th Ann. Con! IEEE-EMBS: 922-923.
Davis, R. (1923). The relationship of the "chronaxie" of muscle to the size of the
stimulating electrode. Proc. Physiol. Soc., July 7: lxxxi-lxxxii.
de Mello, W.C. (1972). The healing-over process in cardiac and other muscle
fibres. In Electrical Phenomena of the Heart, Academic Press, New York: 323-
325.
Dean, D., and P.D. Lawrence (1983). Application of phase analysis of the
Frankenhaeuser-Ruxley equations to determine threshold stimulus amplitudes.
IEEE Trans. Biomed. Eng. BME-30(12): 810-818.
Dean, D., and P.D. Lawrence (1985). Optimization of neural stimuli based upon a
variable threshold potential. IEEE Trans. Biomed. Eng. 32(1): 8-14.
Deleze, J. (1970). The recovery of resting potential and input resistance in sheep
heart injured by knife or laser. 1. Physiol. (London) 208: 547-562.
Deno, D.W. (1975a). Calculating electrostatic effects of overhead transmission
lines. IEEE Trans. Pwr. Apparat. Sys. PAS-93(5): 1458-1471.
Deno, D.W. (1975b). Electrostatic effect induction formulae. IEEE Trans. Pwr.
Apparat. Sys. PAS-94(5): 1524-1536.
References 513
Deno, D.W. (1977). Currents induced in the human body by high voltage transmis-
sion line electric field-measurement of calculation of distribution and dose.
IEEE Trans. Pwr. Apparat. Sys. PAS-96(5): 1517-1527.
Deno, D.W. (1978). Electrostatic and electromagnetic effects of ultrahigh-voltage
transmission lines. Report EL-802. Electric Power Research Institute, Palo Alto,
CA
Deno, D.W., and L.E. Zaffanella (1982). Field effects of overhead transmission lines
and stations. In Transmission Line Reference Book, Electric Research Institute,
Palo Alto, CA: 329-419.
DiFrancesco, D., and D. Noble (1985). A model of cardiac electrical activity incor-
porating ionic pumps and concentration changes. Phil. Trans. R. Soc. London, B:
307: 353-398.
DiMarco, AF., M.D. Altose, A Cropp, and D. Durand (1987). Activation of the
inspiratory intercostal muscles by electrical stimulation of the spinal cord. Am.
Rev. Respir. Dis. 136: 1385-1390.
DiMarco, AF., G.S. Supinski, J.A Petro, and Y. Takaoka (1994). Evaluation of
intercostal pacing to provide artificial respiration in quadriplegics. Amer. 1.
Respir. Crit. Care Med. 150: 934-940.
Dixon, G.F. (1983). The evaluation and management of electrical injuries. Crit. Care
Med. 11: 384-387.
Dominguez, G., and H.A Fozzard (1970). Influence of extracellular K+ concentra-
tion on cable properties and excitability of sheep cardiac Purkinje fibers. Circ.
Res. 26: 565-574.
Donly, K.J., and AJ. Nowak (1988). Oral electrical burns: Etiology, manifestations,
and treatment. Gen. Dent. 36(2): 103-107.
Dorfman, Y.G. (1971). Physical phenomena occurring in live objects under the
effect of constant magnetic fields. In Y.A Kholodov (ed.), Influence of Magnetic
Field on Biological Objects, National Technical Information Service, Springfield,
VA, Report JPRS63038: 11-19.
Dowling, J.E., and B.B. Boycott (1966). Organization of the primate retina: Elec-
tron miscroscopy. Proc. Royal Soc. London, Series B, 166: 80-11I.
Dowse, C.M., and C.E. Iredell (1920). The effective resistance of the human body to
high frequency currents. Arch. Radiol. Electrother. 25(2): 34-46.
Drouhard, J.P., and F.A Roberge (1982a). A simulation study of the ventricular
myocardial action potential. IEEE Trans. Biomed. Eng. BME-29(7): 494-502.
Drouhard, J.P., and F.A Roberge (1982b). The simulation of repolarization events
of the cardiac Purkinje fiber action potential. IEEE Trans. Biomed. Eng. BME-
29(7): 481-493.
DuBois, E.F., and D. DuBois (1916). Formula to estimate approximate surface area
if height and weight be known. Arch. Intern. Med. 17: 863.
Dudel, J. (1989). Transmission of excitation from cell to cell, Chap. 3 (pp 41-60)
in R.F. Schmidt and G. Therw (eds.), Human Physiology, Springer-Verlag,
Berlin.
Durney, C.F., C.C. Johnson, and H. Massoudi (1975). Long wavelength analysis of
plane wave irradiation of a prolate spheroid model of a man. IEEE Trans.
Microwave Theory, MTT-23(2): 246-253.
Durney, C.F., H. Massoudi, and M.F. Isadander (1985). Radiofrequency Radiation
Handbook, Report USAFSAM-TR-85-73, USAF School of Medicine, Brooks
Air Force Base, Texas.
514 References
Eaton, H.AC. (1992). The electric field induced in a spherical volume conductor
from arbitrary coils: Application to magnetic stimulation and MEG. Med. Bio!.
Eng. Compo 30: 433-440.
Ebihara, L., and E.A. Johnson (1980). Fast sodium current in cardiac muscle.
Biophys.l. 32: 779-790.
Edelberg, R (1971). Electrical properties of the skin. In RR Elden (ed.), Biophysi-
cal Properties of the Skin, Wiley-Interscience, New York: 513-550.
Edwards, RH.T., D.K. Hill, D.A Jones, and P.A Merton (1977). Fatigue of long
duration in human skeletal muscle after exercise. 1. Physiol. 272: 769-778.
Edwards, RRT. (1981). Human muscle function and fatigue. In R Porter and
J. Whelan (eds.), Human Muscle Fatigue: Physiological Mechanisms, Pitman
Medical, London: 1-18.
Eifler, W.J., and R Plonsey (1975). A cellular model for the simulation of activation
in the ventricular myocardium. 1. Electrocardiol. 8(2): 117-128.
EI-Sherif, N., B.J. Scheriag, and R Lazzara (1975). Electrode catheter recordings
during malignant ventricular arrhythmias following experimental acute myocar-
dial ischemia. Circulation, 51: 1003-1014.
Elden, RR (ed.) (1971). Biophysical Properties of the Skin, Wiley-Interscience,
New York.
Elharrar, V., P.R Forster, T.L. Jirak, W.E. Gaum, and D.P. Zipes (1977). Alter-
ations in canine myocardial excitability during ischemia. Circ. Res. 40: 98-105.
Engelmann, T.W. (1875). Leitung und Erregung im Herzmuskel, Pflugers Arch. 11:
465-480.
Epstein, B.R, and K.R Foster (1983). Anisotropy in the dielectric properties of
skeletal muscle. Med. Bioi. Eng. Compo 21: 51-55.
Evans, B.A., W.J. Litchy, and J.R Daube (1988). The utility of magnetic stimulation
for routine peripheral nerve conduction studies. Muscle & Nerve, 11: 1074-
1078.
Fang, Z.P., and J.T. Mortimer (1987). A method for attaining natural recruitment
order in artificially activated muscles. Proc. 9th IEEE-EMBS Conf: 657-658.
FDA (1992). FDA safety parameter action levels. In RL. Magin, RL. Liburdy, and
B. Persson (eds.), Biological Effects and Safety Aspects of Nuclear Magnetic
Resonance Imaging, New York Academy of Sciences, New York: 399-400.
FDA (1995). MRI guidance update. Draft statement, RA Phillips (Chief), Com-
puted Imaging Devices Branch, ODE/CDRH, Rockville, MD (Nov. 11, 1995).
Ferris, L.P., E.G. King, P.W. Spence, and H.B. Williams (1936). Effect of electric
shock on the heart. AlEE Trans. 55: 498-515.
Fisun, 0.1. (1993). 2D plasmon excitation and nonthermal effects of microwaves on
biological membranes. Bioelectromagnetics 14: 57-66.
Fitts, RH. (1994). Cellular mechanisms of muscle fatigue. Physiol. Rev. 74(1): 49-94.
Fitzhugh, R (1962). Computation of impulse initiation and saltatory conduction in
a myelinated nerve fiber. Biophys. 1. 2: 11-21.
Fitzhugh, R (1966). Theoretical effects of temperature on threshold in the
Hodgkin-Huxley nerve model. J. Gen. Physio!. 49: 989-1005.
Flottropp, O. (1953). Effect of different types of electrodes in electrophonic hearing.
1. Acoust. Soc. Amer., 25: 236-245.
Forbes, T.W., and AL. Bernstein (1935). The standardization of sixty-cycle electric
shock for practical use in psychological experimentation. 1. Gen. Psycho!. 12/13:
436-441.
References 515
Foster, K.R, and RP. Schwan (1996). Dielectric properties of tissues. In C. Polk
and E. Postow (eds.), CRC Handbook of Biological Effects of Electromagnetic
Fields, CRC Press, Boca Raton, FL: 25-102.
Fozzard, H.A (1979). Conduction of the action potential. In RM. Berne, N.
Sperelakis, and S. Geiger (eds.), The Cardiovascular System. Handbook of
Physiology, vol. I, sec. 2, American Physiological Society, Bethesda, MD: 335-
356.
Fozzard, RA, and M.F. Arnsdorf (1986). Cardiac electrophysiology. In RA
Fozzard, E. Haber, RB. Jenings, AM. Katz, and RE. Morgan (eds.), The Heart
and Cardiovascular System, Raven, New York: 1-30.
Fozzard, R.A, and M. Schoenberg (1972). Strength-duration curves in cardiac
Purkinje fibers: Effects of liminal length and charge distribution. 1. Physiol. 226:
593-618.
Frankel, RB., and RP. Liburdy (1996). Biological effects of static magnetic fields.
In C Polk and E. Postow (eds.), Biological Effects of Electromagnetic Fields, CRC
Press, Boca Raton, FL.
Frankenhaeuser, B., and AF. Huxley (1952). A quantitative description of mem-
brane currents and its application to conduction and excitation in nerve.
1. Physiol. 117: 500-544.
Frankenhaeuser, B., and AF. Huxley (1964). The action potential in the myelinated
nerve fiber of Xenopus laevis as computed on the basis of voltage clamp data.
1. Physiol. 171: 302-315.
Freiberger, R (1934). Der elektrische Widerstand des menschlichen Korpers gegen
technischen Gleich-und Wechselstrom, Springer-Verlag, Berlin. Translation TR
79-45. The electrical resistance of the human body to commercial direct and
alternating currents. Translated from German by Allen Translation Service,
Maplewood, NJ, Item no. 9005. Published by Bell Laboratories, Maplewood,
NJ.
Frey, AH. (ed.) (1994). On the Nature of Electromagnetic Field Interactions. R.G.
Landes Co., Austin, TX.
Friedli, W.G., and M. Meyer (1984). Strength-duration curve: A measure for assess-
ing sensory deficit in peripheral neuropathy. 1. Neurol. Psychiat. 47: 184-189.
Furman, S., B. Parker, and D.J.W. Escher (1967). Endocardial electrical threshold
of human cardiac response as a function of electrode surface area. Digest 7th Int.
Con! Med. and Bioi. Eng., Stockholm, Sweden Aug. 14-19, p. 71.
Furman, S., J. Garvey, and P. Hurzeler (1975). Pulse duration variation and elec-
trode size as factors in pacemaker longevity. 1. Thorac. Cardiovas. Surg. 69(3):
382-389.
Furnary, AP., J.C Chachques, L.F. Moreira, G.L. Grunkemeier, J.S. Swanson,
N. Stolf, S. Haydar, C. Acar, A. Starr, A.D. Jatene, and AF. Carpentier (1996).
Long-term outcome, survival analysis, and risk stratification of dynamic
cardiomyoplasty.l. Thorac. Cardiovasc. Surg. 112(6): 1640-1650.
Gaffey, CT., and T.S. Tenforde (1981). Alterations in the rat electrocardiogram
induced by a stationary magnetic field. Bioelectromagnetics 2: 357-370.
Galvani, L. (1791). Commentary on the Effect of Electricity on Muscular Motion,
translated by R.M. Green. Waverly, Baltimore, MD: 1953.
Gandhi, G.P. (1988). Advances in dosimetry of radiofrequency radiation and their
past and projected impact on safety standards. Proc. Instrumentation and Tech-
nology Conf, San Diego, California, April 20-22.
516 References
Gandhi, O.P. (ed.) (1990). Biological Effects and Medical Applications of Electro-
magnetic Energy. Prentice-Hall, Englewood Cliffs, NJ.
Gandhi, O.P., and J. Chen (1992). Numerical dosimetry at power line frequencies
using anatomically based models. Bioelectromagnetics, Supplement 1: 43-60.
Gandhi, O.P., J.F. DeFord, and R Kanai (1984). Impedance method for calculation
of power deposition patterns in magnetically induced hyperthermia. IEEE Trans.
Biomed. Eng. BME-31(10): 644-651.
Garrey, W.E. (1914). The nature of fibrillary contractions. Its relation to tissue mass
and form. Am. J. Physiol. 33: 397-402.
Gauger, J.R (1985). Household appliance magnetic field survey. IEEE Trans. Pwr.
Apparat. Sys. P AS-104(9): 2436-2444.
Gaylor, D.C. (1989). Physical mechanism of cellular injury in electrical injury. In
RC. Lee (ed.), Electric Trauma: Biophysical Mechanisms of Tissue Injury and
Clinical Concepts, Chicago, IL, July 13-14.
Gaylor, D.C., K. Prakah-Asante, and RC. Lee (1988). Significance of cell size and
tissue structure in electrical trauma. J. Theor. Bioi. 133: 223-237.
Gaylor, D.C., D.L. Bhatt, and RC. Lee (1992). Skeletal muscle cell membrane
electrical breakdown in electrical trauma. In RC. Lee, E.G. Gravalho, and J.F.
Burke (eds.) (1992) Electrical Trauma. Cambridge University Press, Cambridge/
New York: 401-425.
Geddes, L.A (1972). Electrodes and the Measurement of Bioelectric Events. Wiley-
Interscience, New York.
Geddes, L.A (1985). The conditions necessary for the electrical induction of ven-
tricular fibrillation. In J.E. Bridges, G.L. Ford, I.A Sherman, and M. Vainberg
(eds.), Electrical Shock Safety Criteria, Pergamon, New York: 45-59.
Geddes, L.A (1987). Optimal stimulus duration for extracranial cortical stimula-
tion. Neurosurgery 20: 94-99.
Geddes, L.A, and H. Antoni (panel chairmen) (1985). Panel meeting on physiology
of electrical shocks. In J.E. Bridges, G.L. Ford, I.A Sherman, and M. Vainberg
Geddes, L.A, and J.D. Bourland (1985). Tissue stimulation: Theoretical consider-
ations and practical applications. Med. BioI. Eng. Compo 23: 131-137.
Geddes, L.A, and L.F. Baker (1967). The specific resistance of biological mate-
rial-A compendium for the biomedical engineer and physiologist. Med. Bioi.
Eng. 5: 271-293.
Geddes, L.A, and L.E. Baker (1971). Response to passage of electric current
through the body. J. Assoc. Adv. Med. Instrum. 5(1): 13-18.
Geddes, L.A, and L.E. Baker (1975). Principles of Applied Biomedical Instrumen-
tation, 2nd ed. John Wiley & Sons, New York.
Geddes, L.A, L.E. Baker, AG. Moore, and T.W. Coulter (1969). Hazards in the
use of low frequencies for the measurement of physiological events by imped-
ance. Med. Bioi. Eng. 7: 289-296.
Geddes, L.A, L.E. Baker, P. Cabler, and D. Brittain (1971). Response to passage of
sinusoidal current through the body. In RN. Wolfson and A Sances (eds.), The
Nervous System and Electric Currents, Vol. II, Plenum, New York.
Geddes, L.A, N. Morehouse, and A Surawiez (1972). Effect of premature depolar-
ization on the duration of action potentials in Purkinje and ventricular fibers of
moderator band of the pig heart. Circ. Res. 30: 55-66.
References 517
Geddes, L.A., P. Cabler, AG. Moore, J. Rosborough, and W.A Tacker (1973).
Threshold 60-Hz current required for ventricular fibrillation in subjects of various
body weights. IEEE Trans. Biomed. Eng. BME-20: 465-468.
Geddes, L.A, W.A Tacker, and P. Cabler (1975). A new hazard associated with the
electrocautery. Med. Instr. 9: 112-113.
Geddes, L.A, M.1. Niebauer, e.F. Babbs, and J.D. Bourland (1985a). Fundamental
criteria underlying the efficacy and safety of defibrillating current waveforms.
Med. BioI. Eng. Compo 23: 122-130.
Geddes, L.A, W.D. Voorhees, e.F. Babbs, and J.A DeFord (1985b). Electro-
ventilation. Amer. J. Emerg. Med. 3: 337-339.
Geddes, L.A, J.D. Bourland, and G. Ford (1986). The mechanism underlying
sudden death from electric shock. Med. Instrum. 20(6): 303-315.
Geddes, L.A, W.D. Voorhees, R. Lagler, e. Riscili, K. Foster, and J.D. Bourland
(1988). Electrically produced artificial ventilation. Med. Instrum. 22(5): 263-271.
Geddes, L.A, W.D. Voorhees, J.D. Bourland, and e.E. Riscili (1990). Optimum
stimulus frequency for contracting the inspiratory muscles with chest-surface
electrodes to produce artificial respiration. Annals Biomed. Eng. 18: 103-108.
Geddes, L.A, G. Mouchawar, J.D. Bourland, and 1. Nyenhuis (1991). Inspiration
produced by bilateral electromagnetic cervical phrenic nerve stimulation in man.
IEEE Trans. Biomed. Engin. 38(9): 1047-1048.
Geldard, F.A. (1972). The Human Senses. John Wiley & Sons, New York.
George, M.S., E.M. Wassermann, W.A Williams, A Callahan, T.A Ketter, P.
Basser, M. Hallett, and R.M. Post (1995). Daily repetitive transcranial magnetic
stimulation (rTMS) improves mood in depression. Neuroreport (Rapid Science
Publishers), 6(14): 1853-1856.
Gerst, P.H., W.H. Fleming, and J.R. MaIm (1996). Increased susceptibility of the
heart to ventricular fibrillation during metabolic acidosis. Circulation Res. 19: 63-
70.
Gescheider, G.A, and J.H. Wright (1968). Effects of sensor adaptation on the form
of the psychophysical magnitude function for cutaneous vibration. J. Exp.
Psycho I. 77: 308-313.
Gibson, R.R. (1968). Electrical stimulation of pain and touch. In D.R. Kenshalo
(ed.), The Skin Senses, Charles e. Thomas, Springfield, IL: 223-261.
Gilbert, T.e. (1939). What is a safe voltage? The Electrical Review 119(3062): 145-
146.
Girvin, J.P., L.E. Marks, J.L. Antunes, D.O. Quest, M.D. O'Keefe, P. Ning,
and W.H. Dobelle (1982). Electrocutaneous stimulation-I. The effects of
stimulus parameters on absolute threshold. Perception & Psychophys. 32(6): 524-
528.
Glenn, W.W.L., and M.L. Phelps (1985). Diaphragm pacing by electrical stimulation
of the phrenic nerve. Neurosurgery 17: 974-984.
Glenn, W.W.L., et a1. (1986). Twenty years of experience in phrenic nerve stimula-
tion to pace the diaphragm. PACE 9: 780-784.
Glenn, W.W.L., et a1. (1988). Fundamental considerations in pacing of the dia-
phragm for chronic respiratory insufficiency: A multi-center study. PACE 11:
2121-2127.
Goellner, D., B. Zackheim, and M. Bockleman (1993). Safety of high speed guided
ground transportation systems: Review of existing EMF guidelines, standards,
518 References
and regulations. U.S. Dept. of Transportation, Washington, DC, Report DOTI

FRAIORD-93/27.
Goff, G.D., B.S. Rosner, T. Detre, and D. Kennard (1965). Vibration perception in
normal man and medical patients. 1. Neurol. Neurosurg. Psychiatr. 18: 503-509.
Gorman, P.H., and J.T. Mortimer (1983). The effect of stimulus parameters on the
recruitment characteristics of direct nerve stimulation. IEEE Trans. Biomed.
Eng. BME-30(7): 407-414.
Gracely, RH., R Dubner, and P.A McGrath (1979). Narcotic analgesia: Fentanyl
reduces the intensity but not the unpleasantness of painful tooth pulp sensations.
Science 203(3): 1261-1263.
Graham, e., M.R Cook, H.D. Cohen, and M.M. Gerkovich (1994). Dose response
study of human exposure to 60Hz electric and magnetic fields. Bioelec-
tromagnetics 15: 447-463.
Grandjean, P.A, and J.T. Mortimer (1986). Recruitment properties of monopolar
and bipolar epimysial electrodes. Annals Biomed. Eng. 14: 53-66.
Grandjean, P.A, R Leinders, and I. Bourgeois (1991). Implantable stimulation
systems for systolic and diastolic biomechanical cardiac assistance. Sem. Thorac.
Cardiovasc. Surg. 3(2): 119-123.
Grandjean, P., M. Acker, R Madoff, N.S. Williams, J. Woloszko, and e. Kantor
(1996). Dynamic myoplasty: Surgical transfer and stimulation of skeletal muscle
for functional substitution or enhancement. 1. Rehab. Res. Develop. 33(2): 133-
144.
Grandori, F., and P. Ravazzani (1991). Magnetic stimulation of the motor cortex-
theoretical considerations. IEEE Trans. Biomed. Eng. 38(2): 180-19l.
Graupe, D. (1989). EMG pattern analysis for patient-responsive control of FES
in paraplegics for walker-supported walking. IEEE Trans. Biomed. Eng. 36(7):
711-719.
Gray, H. (1985). Gray's Anatomy, 30th ed. e.D. Clemente (ed.), Lea & Febiger,
Philadelphia.
Gray, R (1989). Gray's Anatomy, 37th ed., P.L. Williams, R Warwick, M. Dyson,
and L.R Bannister (eds.), Churchill Livingstone, New York.
Grayson, Ae. (1931). Treat low voltages with respect. National Safety News 23: 32,
34,64.
Green, E.G. (1977). The effect of skin temperature on vibrotactile sensitivity. Per-
ception & Psychophys. 21: 243-248.
Green, D.M. (1960). Psychoacoustics and detection theory. 1. Acoust. Soc. Am. 32:
1189-1203.
Green, D.M., and J.A Swets (1966). Signal Detection Theory and Psychophysics.
John Wiley & Sons, New York.
Green, RL., E.B. Rafferty, and I.e. Gregory (1972). Ventricular fibrillation thresh-
old of healthy dogs to 50Hz current in relationship to earth leakage currents of
electromedical equipment. Biomed. Eng. 7: 408-414.
Green, R.L., J. Ross, and P. Kurn (1985). Danger levels of short (1 msec. to 15
msec.) electrical shocks from 50-Hz supply. In J.F. Bridges, G.L. Ford, LA
Sherman, and M. Vain berg (eds.), Electrical Shock Safety Criteria, Pergamon,
New York: 259-272.
Greenberg, AW. (1940). Experimental radiological observations on the action of
electrical current upon the respiratory and circulatory organs. L Respiratory
organs. 1. Ind. Hyg. Toxicol. 22: 104-110.
References 519
Grimnes, S. (1983a). Dielectric breakdown of human skin in vivo. Med. Bioi. Eng.
Compo 21: 379-381.
Grimnes, S. (1983b). Electrovibration, cutaneous sensation of microampere cur-
rents. Acta Physiol. Scand. 118(1): 19-25.
Grissom, C.B. (1995). Magnetic field effects in biology: A survey of possible mecha-
nisms with emphasis on radical-pair mechanisms. Chem. Rev. 95: 3-24.
Guinard, J.P., R Chiolero, E. Buchser, A. Delaloye-Bischof, M. Payot, A. Grbic,
S. Krupp, and J. Freeman (1987). Myocardial injury after electrical burns: Short
and long term study. Scand. 1. Plast. Reconstr. Surg. Hand Surg. 21(3): 301-
302.
Gumbel, E.J. (1958). Statistics of Extremes. Columbia University Press, New
York.
Gustafson, RJ., T.M. Brennan, and RD. Appleman (1985). Behavioral studies of
dairy cow sensitivity to AC and DC electric currents. Trans. ASAE, 28(5): 1680-
1685.
Gustafson, R.J., Z. Sun, and T.D. Brennan (1988). Dairy cow sensitivity to short
duration electrical currents. ASAE Paper No. 88-3522, Amer. Soc. Ag. Eng, St.
Joseph, MI.
Gybels, J., RO. Handwerker, and J.N. Hees (1979). A comparison between the
discharges of human nociceptive nerve fibers and the subject's ratings of his
sensations. 1. Physiol. 292: 193-206.
Haberal, M. (1986). Electrical burns: A five-year experience-1985 Evans lecture.
1. Trauma 26(2): 103-109.
Hahn, J.F. (1958). Cutaneous vibratory thresholds for square-wave electrical pulses.
Science 127: 879-880.
Hallgren, R (1973). Inductive Neural Stimulator. IEEE Trans. Biomed. Eng. BME-
20(6): 470-472.
Hammond, E., and T.D. Robson (1955). Comparison of electrical properties of
various cements and concretes. The Engineer (London), 199(5156): 78-80 and
199(5166): 114-115.
Hammond, J.S., and e.G. Ward (1988). High-voltage electrical injuries: Manage-
ment and outcome of 60 cases. South. Med. 1. 81(11): 1351-1352.
Han, J. (1969). Ventricular vulnerability during acute coronary occlusion. Am. 1.
Cardiol. 24: 857-864.
Han, J., and G.K. Moe (1964). Nonuniform recovery of excitability in ventricular
muscle. Circulation Res. 14: 44-60.
Han, J., G. Dejalon, and G.K. Moe (1966). Fibrillation threshold of premature
ventricular responses. Circulation Res. 18: 18-25.
Handa, Y., N. Hoshimiya, Y. Iguchi, and T. Oda (1989). Development of percutane-
ous intramuscular electrode for multichannel FES system. IEEE Trans. Biomed.
Eng. 36(7): 705-710.
Hardy, J.D., RG. Wolff, and R Goodbell (1952). Pain Sensations and Reactions.
Rafner, New York.
Hare, P.R. (1968). Detection threshold for electric shock in psychopaths. 1. Abnor-
mal Psychol. 73(3): 268-272.
Harkins, S.W., and e.R Chapman (1976). Detection and decision factors in pain
perception in young and elderly men. Pain 2: 253-264.
Harkins, S.W., and C.R Chapman (1977). The perception of induced dental pain in
young and elderly women. 1. Gerontol. 32: 428-435.
520 References
Harkness, RD. (1971). Mechanical properties of skin in relation to its biological

function and its chemical components. In RR Elden (ed.), Biophysical Properties
of the Skin, Wiley-Interscience, New York: 393-436.
Harris, R (1971). Chronaxy. In S. Licht (ed.), Electrodiagnosis and Electromyogra-
phy, 3rd ed., E. Licht, New Haven, CT, chap. 9.
Hart, FX. (1992). Numerical and analytical methods to determine the current
density distributions produced in human and rat models by electric and magnetic
fields. Bioelectromagnetics, Supplement 1: 27-42.
Hart, W.F. (1985). A five-part resistor-capacitor network for measurement of volt-
age and current levels related to electric shock and burns. In J.E. Bridges, G.L.
Ford, I.A. Sherman, and M. Vainberg (eds.), Electrical Shock Safety Criteria,
Pergamon, New York: 183-192.
Hauf, G., K. Haap, M. Lay, and R Antoni (1977). Beziehungen zwischen der
Richtung des Stromdurchgangs und der Flimmerschwelle bei perfundierten
Tierherzen. In R Rauf (ed.), Beitriige Zur Ersten Hilfe, Forschungsstelle fur
Elektropathologie, Freiburg i. Br.: 146-163.
Hauf, R (1986). Beitriige zur Ersten Hilfe und Behandlung von Unfiillen durch
elektrischien Strom. Proceedings of Conference on Electropathology, Sept. 11-13,
1986, Forschungsstelle fur Elektropathologie, Freiburg, Germany.
Havel, W.J., J.A. Nyenhuis, J.D. Bourland, K.S. Foster, L.A. Geddes, G.P. Graber,
M.S. Waniger, and D.l Schaefer (1997). Comparison of rectangular and damped
sinusoidal dB/dt waveforms in magnetic stimulation. IEEE Trans. Magnetics,
33(5): 4269-4271.
Hawkes, G.R, and J.S. Warm (1960). The sensory range of electrical stimulation of
the skin. Am. J. Psychol. 73: 485-487.
Hawkes, G.R (1962). Effect of skin temperature on absolute threshold for electrical
current. J. Appl. Physiol. 17: 110-112.
Heckmann, J.R (1972). Excitability curve: A new technique for assessing human
peripheral nerve excitability in vivo. Neurology 22: 224-230.
Heft, M. (1982). Conjoint measurement analysis of verbal category judgments
for electrocutaneous stimulation. Ph.D. dissertation, American University,
Washington, DC.
Heinz, M., and F. Lippay (1928). Uber die beziehungen zwischen der
unterscheidsempfindlichkeit und der zahl der erregten sinneselemente: I. Pflugers
Arch. ges. Physiol. Menschien Tiere 218: 437-447.
Henneman, E., and L.M. Mendell (1981). Functional organization of motoneuron
pool and its inputs. In V.B. Brooks (ed.), Handbook of Physiology, Section 1: The
Nervous System. Vol. II. Motor Control, Part II, American Physiological Society,
Bethesda, MD: 423-507.
Henneman, E., G. Somjen, and D.O. Carpenter (1965a). Functional significance of
cell size in spinal motoneurons. J. Neurophys. 28: 581-598.
Henneman, E., G. Somjen, and D.O. Carpenter (1965b). Excitability and inhib-
itability of motoneurons of different sizes. J. Neurophys. 28: 599-620.
Henriquez, C.S. (1993). Simulating the electrical behavior of cardiac tissue using the
bidomain model. Crit. Rev. Biomed. Eng. 21(1): 1-77.
Henriquez, C.S., and R. Plonsey (1987). Effect of resistive discontinuities on wave-
shape and velocity in single cardiac fiber. Med. Bioi. Eng. Comput. 25: 428-438.
Henriques, F.C. (1947). Studies of thermal injury v. the predictability and the
significance of thermally induced rate processes leading to irreversible epidermal
injury. Arch. Pathol. 43: 489-502.
References 521
Henriques, F.e., and AR Moritz (1947). Studies of thermal injury I. The con-
duction of heat to and through skin and the temperatures attained therein. A
theoretical and experimental investigation. Am. J. Pathol. 23: 531-549.
Hensel, R (1973). Cutaneous thermoreceptors. In A Iggo (ed.), Handbook of
Sensory Physiology: Somatosensory System, vol. 2, Springer-Verlag, New York:
79-110.
Henshaw, D., AN. Ross, AP. Fews, and AW. Preece (1996). Enhanced deposition
of radon daughter nuclei in the vicinity of power frequency electromagnetic fields.
Int. J. Radiat. Bioi. 69(1): 25-38.
Heroux, P. (1992). Thermal damage: Mechanisms, patterns, and detection in electri-
cal burns. In RC. Lee, E.G. Gravalho, and J.F. Burke (eds.) (1992). Electrical
Trauma, Cambridge University Press, Cambridge/New York: 189-215.
Hess, e.W., K.R Mills, N.M.P. Murray, and T.N. Schriefer (1987). Magnetic brain
stimulation: Central motor conduction studies in multiple sclerosis. Annals
Neurol. 22: 744-752.
Higashiyama, A, and G.B. Rollman (1991). Perceived locus and intensity of
electrocutaneous stimulation. IEEE Trans. Biomed. Eng. 38(7): 679-686.
Higgins, J.D., B. Tursky, and G.E. Schwartz (1971). Shock-elicited pain and its
reduction by concurrent tactile stimulation. Science 172: 866-867.
Hill, AV., RS. Fullerton, B. Katz, and D.Y. Solandt (1937). Nerve excitation by
alternating current. Proc. R. Soc. London, Ser. B 121: 74-132.
Hille, B. (1984). Ionic Channels in Excitable Membranes, Sinauer Associates,
Sunderland, M.A Hodgkin, AL., and AF. Huxley (1952). A quantitative de-
scription of membrane current and its application to conduction and excitation in
nerve. J. Physiol. 117: 500-544.
Hodgkin, AL., and B. Katz (1949). The effect of temperature on the electrical
activity of the giant axon of the squid. J. Physiol (London). 109: 240-249.
Hoffer, J.A, RB. Stein, M. Haugland, T. Sinkjaer, W.K. Durfee, AB. Schwartz,
G.E. Loeb, and e. Kantor (1996). Neural signals for command control and feed-
back in functional neuromuscular stimulation. Rehab. Res. Develop. 33(2): 145-
157.
Hoffman, B.P., and P.F. Cranefield (1960). Electrophysiology of the Heart. McGraw-
Hill, New York.
Hoffmeister, B., W. Janig, and S.l Lisney (1991). A proposed relationship
between circumference and conduction velocity of unmyelinated axons from
normal and regenerated cat hindlimb cutaneous nerves. Neuroscience 42(2): 603-
60l.
Hohnloser, S., S. Weirich, and R Antoni (1982). Influence of direct current on the
electrical activity of the heart and on its susceptibility to ventricular fibrillation.
Basic Res. Cardiol. 77: 237-249.
Holle, J., M. Frey, R Gruber, R Kern, R Stohr, and R Thoma (1984). Functional
electrostimulation of paraplegics: Experimental investigations and first clinical
experience with an implantable stimulation device. Orthop. 7: 1145-1155.
Hooker, D.R, W.B. Kouwenhoven, and O.R Langworthy (1932). The effect of
alternating currents on the heart. Am. J. Physiol. 103: 444-454.
Hooshmand, H., F. Radfar, and E. Beckner (1989). The neurophysiological aspects
of electrical injuries. Clin. Electroencephalogr. 20(2): 111-120.
Hoque, M., and O.P. Gandhi (1988). Temperature distributions in the human leg for
VLF-VHF exposures at the ANSI-recommended safety levels. IEEE Trans.
Biomed. Eng. 35(6): 442-449.
522 References
Horowitz, L.N., J.F. Spear, M.E. Josephson, J.A. Kastor, and E.N. Moore (1979).
The effects of coronary artery disease on the ventricular fibrillation threshold in
man. Circulation Res. 60: 792-797.
Hoshimiya, N., A. Naito, M. Yajima, and Y. Handa (1989). A multichannel FES
system for the restoration of motor functions in high spinal cord injury patients:
A respiration-controlled system for multijoint upper extremity. IEEE Trans.
Biomed. Eng. 36(7): 754-760.
Hosono, A., T. Andoh, T. Goto, T. Kawakami, F. Okumura, K. Takayama, T.
Takenaka, S. Ueno, M. Yamaguchi, and I. Yamamoto (1992). Effective combina-
tion of stimulating coils for magnetic heart stimulation. Jpn. J. Appl. Phys. 3(Pt. 1,
no. 11): 3759-3762.
Howatson, A.M. (1965). An Introduction to Gas Discharges. Pergamon, Oxford.
HSRI (1977). Anthropometry of infants, children, and youths to age 18 for product
safety design. Final Report, May 31, 1977, prepared for the U.S. Consumer
Product Safety Commission by the Highway Safety Research Institute, University
of Michigan Contract CPSC-C-75-0068.
Hufeland, C.W. (1783). Usum uis electriciae in asphyxia experimentis illustratum.
In: Dissertatio Inauguralis Medica. Gottingen, Germany.
Hunt, J.L. (1979). The use of technetiun-99m stannous pyrophosphate scintigraphy
to identify muscle damage in acute electric burns. J. Trauma. 19(6): 409-413.
Hunt, J.L., RM. Sato, and C.R Baxter (1980). Acute electric burns. Current diag-
nostic and therapeutic approaches to management. Arch. Surg. 115: 434-438.
Huxley, RE., and J. Hanson (1954). Changes in the cross-atriation of muscle during
contraction and stretch and their structural interpretation. Nature 173: 973-976.
Hylten-Cavallius, N. (1975). Certain ecological effects of high voltage power lines.
Institut de recherche de I'Hydro-Quebec (IREQ), IREQ-1160.
IAEI News Bulletin (Anon.) (1940). Oregon's first death from an electric fence. Vol.
12, p. 70.
ICNIRP (1994). Guidelines on limits of exposure to static magnetic fields. Health
Physics 66(1): 100-106.
ICNIRP (1998). Guidelines limiting exposure to time-varying electric, magnetic,
and electromagnetic fields (up to 300 Ghz). Health Physics 74(3): 494-522.
Ichioka, S., M. Iwasaka, M. Shibata, K. Harii, A. Kamiya, and S. Veno (1998).
Biological effects of static magnetic fields on the micocirculatory blood flow in
vivo: a preliminary report. Med. Bioi. Eng. Comput. 36: 91-95.
IEC (1982). Report of Technical Committee 64: Electrical installation of buildings.
International Electrotechnical Commission, Geneva, Switzerland.
IEC (1984). Effects of current passing through the human body, Part 1: General
Aspects. Publication 479-1, International Electrotechnical Commission, Geneva,
Switzerland.
IEC (1985). Meeting Minutes TC74IWGS, Everett, Washington, April 2, 3, 4, 1985.
International Electrotechnical Commission, Geneva, Switzerland, July 16, 1985.
IEC (1986). Safety of information technology equipment including electrical busi-
ness equipment. Publication 950 (including Amendment No.1, Nov., 1988).
IEC (1987). Effects of current passing through the human body, Part 2: Special
Aspects. Publication 479-2, International Electrotechnical Commission, Geneva,
Switzerland.
IEC (1990a). Safety requirements for electrical equipment for measurement, con-
trol, and laboratory use. Publication 1010-1, International Electrotechnical
Commission, Geneva, Switzerland.
References 523
IEC (1990b). Methods of measurement of touch current and protective conductor

current. Report of TC74/WGS, Publication 990, International Electrotechnical
Commission, Geneva, Switzerland.
IEC (1993). Diagnostic Imaging Equipment. IECITC 62B(Secretariat)145, Interna-
tional Electrotechnical Commission, Geneva, Switzerland.
IEC (1995). Medical Electrical Equipment-Part 2: Particular Requirements for the
Safety of Magnetic Resonance Equipment for Medical Diagnosis. International
Electrotechnical Commission Publication 601-2-33, Geneva, Switzerland.
IEEE (1982). IEEE Recommended Practice for Grounding of Industrial and Com-
mercial Power Systems. Std. 142-1982. Institute of Electrical and Electronics
Engineers, New York.
IEEE (1986). Guide for safety in ac substation grounding. ANSI/IEEE Std 80-1986,
Institute of Electrical and Electronics Engineers, New York.
IEEE (1990). National Electrical Safety Code. ANSI C2-1990, Institute of Electrical
and Electronics Engineers, New York.
IEEE (1992). IEEE Standard for Safety Levels with Respect to Human Exposure to
Radio Frequency Electromagnetic Fields, 3kHz to 300GHz. Document IEEE
C95.1-1991, published by Institute of Electrical and Electronics Engineers, New
York.
Inancsi, W., and T.L. Guidotti (1987). Occupation-related burns: Five-year experi-
ence of an urban burn center. J. Occup. Med. 29(9): 730-733.
Irnich, W. (1973). Physikalische Oberlegungen zur Elektrostimulation (Physical
consideration on electrostimulation). Biomed. Technik. 18: 97-104.
Irnich, W. (1980). The chronaxie time and its practical importance. PACE, 3: 292-
30l.
IRPA (1988). Guidelines on limits of exposure to radiofrequency electromagnetic
fields in the frequency range from 100kHz to 300 Ghz. Health Physics, 54(1): 115-
123.
IRPA (1990). Interim guidelines an limits of exposure to SO/60Hz electric and
magnetic fields. Health Physics 58(1): 113-122.
IRPA (1991). Protection of the patient undergoing a magnetic resonance examina-
tion. Health Physics 61(6): 923-928.
Irwin, D., S. Rush, R Everling, E. Lepeschkin, D.B. Montgomery, and RJ. Weggel
(1970). Stimulation of cardiac muscle by a time-varying magnetic field. IEEE
Trans. Magnet. MAG-6(2): 321-322.
lIT (1979). Reference Data for Radio Engineers. Howard W. Sams, New York.
Jack, J.B., D. Noble, and RW. Tsien (1983). Electric Current Flow in Excitable Cells.
Oxford University Press (Clarendon). London/New York.
Jackson, J.D. (1962). Classical Elflctrodynamics, John Wiley & Sons, New York.
Jackson, T.A., and B.F. Riess (1934). Electric shock with different size electrodes.
J. Gen. Psychol. 45: 262-266.
Jacobsen, J., S. Buntenkutter, and RJ. Reinhard (1975). Experimentelle
untersuchungen an schweinen zur frage der mortalitat durch sinusformige,
phasengeschnittene sowie gleichgerichtete elektrische strome. Biomed. Technik
20: 99-107.
Jaeger, RJ., G.M. Yarkony, and RM. Smith (1989). Standing the spinal cord
injured patient by electrical stimulation: Refinement of a protocol for clinical use.
IEEE Trans. Biomed. Eng. 36(7): 720-728.
Jalife, J., and G.K. Moe (1976). Effect of electrotonic potentials on pacemaker
activity of canine Purkinje fibers in relation to parasystole. Circ. Res. 39: 801-808.
524 References
Jalife, J., and G.K. Moe (1981). Excitation, conduction, and reflection of impulses in
isolated bovine and canine cardiac Purkinje fibers. Circ. Res. 49: 233-247.
Janse, M.J., F.J.L. Van Capelle, R Morsink, AG. Kleber, F. Wilms-Schopman, R
Cardinal, C. Naumann D'Alnoncourt, and D. Durrer (1980). Flow of "injury"
current and patterns of excitation during early ventricular arrhythmias in acute
regional myocardial ischemia in isolated porcine and canine hearts. Circ. Res. 47:
151-165.
Janzekovic, Z. (1975). The burn wound from the surgical point of view. J. Trauma
15: 42.
Jeheson, P., D. Duboc, T. Lavergne, L. Guize, F. Guerin, M. Degeorges, and A.
Syrota (1988). Change in human cardiac rhythm induced by a 2-T static magnetic
field. Radiology 166: 227-230.
Jex-Blake, AJ. (1913). Br. Med. J. 1: 425.
Johna, R (1989). Elektrophysiologische Eigenschaften gekoppelter Herzmus-
kelzellen in Zellkultur. Dissertation, Universitat Freiburg i. Br., Germany.
Johnston, F.E., and RM. Malina (1966). Age changes in the composition of the
upper arm in Philadelphia children. Human Bioi. 38: 1-2l.
Jones, M., and L.A Geddes (1977). Strength-duration curves for cardiac pace-
making and ventricular fibrillation. Cardiovasc. Res. Bull. 15(4): 101-112.
Jones, ROo, J.C. Wright, W.T. Jones, and A Berger (1983). A case study of high
voltage electrical injury. J. Am. Podiatry Assoc. 73: 638-642.
Joyner, RW., F.R Ramon, and W. Moore (1975). Simulation of action potential
propagation in an inhomogeneous sheet of electrically coupled excitable cells.
Circ. Res. 36: 654-66l.
Kaczmarek, K.A, J.G. Webster, and RG. Radwin (1992). Maximal dynamic range
electrotactile stimulation waveforms. IEEE Trans. Biomed. Eng. 39(7): 701-715.
Kalmijn, AJ. (1990). Transduction of nanovolt signals: Limits of electric-field detec-
tion. Bioelectromagnetics Soc. Newsl. JanlFeb. issue, no. 92: 1-7.
Kandel, E.R, and J.B. Schwartz (1981). Principles of Neural Science. Elsevier/
North-Rolland, New York.
Kandel, R, J.H. Schwartz, and T.M. Jessel (1991). Principles of Neural Science, 3rd
edition. Elsevier, New York.
Kanitkar, S., and AR Roberts (1988). Paraplegia in an electrical burn: A case
report. Burns Incl. Therm. Ini. 14(1): 49-50.
Kantrowitz, A (1990). Autologous muscle to assist the failing heart: First experi-
ments. J. Heart Transp. 9: 146-150.
Kaplan, E.B. (1984). Kaplan's Functional and Surgical Anatomy of the Hand, 3rd ed.
Morton Spinner, Philadelphia.
Katims, 1.J., E.R Naviasky, M.S. Randell, K.Y. Lorenz, and M.L. Bleecker (1987).
Constant current sinewave transcutaneous nerve stimulation for the evaluation of
peripheral neuropathy. Archives Phys. Med. Rehab. 68: 210-213.
Katims, 1.J., D.N. Taylor, and S.A Wesely (1991). Sensory perception in uremic
patients. ASAIO Transactions 37(3): M370-M372.
Kato, M., S. Ohta, T. Kobayashi, and G. Matsumoto (1986). Response of sensory
receptors of the cat's hindlimb to a transient, step function DC electric field.
Kato, M., S. Ohta, K. Shimizu, Y. Ysuchida, and G. Matsumoto (1989). Detection-
threshold of 50-Hz electric fields by human subjects. Bioelectromagnetics 10:
319-327.
References 525
Katz, B. (1939). Nerve excitation by high-frequency alternating current. J. Physiol.

96: 202-224.
Katz, B. (1966). Nerve, Muscle, and Synapse. McGraw-Hill, New York.
Katz, B., and R Miledi (1967). The study of synaptic transmission in the absence of
nerve impulses. J. Physiol. (London) 192: 407-436.
Kaune, W.T. (1981). Power frequency electric fields averaged over the body sur-
faces of grounded humans and animals. Bioelectromagnetics 2: 403-406.
Kaune, W.T., and W.C. Forsythe (1985). Current densities measured in human
models exposed to 60-Hz electric fields. Bioelectromagnetics 6(1): 13-32.
Kaune, W.T., and M.F. Gills (1981). General properties of the interaction between
animals and ELF electric fields. Bioelectromagnetics 2: 1-11.
Kaune, W.T., and RD. Phillips (1980). Comparison of grounded humans, swine and
rats to vertical, 60-Hz electric fields. Bioelectromagnetics 1: 117-129.
Kaune, W.T., RG. Stevens, N.J. Callahan, RK Steverson, and D.B. Thomas
(1987). Residential magnetic and electric fields. Bioelectromagnetics 8: 315-
335.
Keith, M.W., KL. Kilgore, P.H. Peckham, KS. Wuolle, G. Creasey, and M. Lemay
(1996). Tendon transfers and functional electrical stimulation for restoration of
hand function in spinal cord injury. J. Hand Surg. Am. 21(1): 89-99.
Keltner, J.R, M.S. Roos, P.R Brakeman, and T.F. Budinger (1990). Magnetohy-
drodynamics of blood flow. Magnetic Resonance in Medicine 16: 139-149.
Kenshalo, D.R (1979). Aging effects on cutaneous and kinesthetic sensibilities. In
S.S. Han and D.R Coon (eds.), Special Senses in Aging: A Current Biological
Assessment, Institute of Gerontology, University of Michigan, Ann Arbor, MI:
189-217.
Kiang, N.Y.S. (1965). Discharge patterns of single fibers in the eat's auditory nerve.
M.I.T. Press, Cambridge, MA
Kieback, D. (1988). International comparison of electrical accident statistics.
J. Occup. Accidents 10: 95-106.
Kieffer, S.A, and E.R Heitzman (eds.) (1979). An Atlas of Cross-Sectional
Anatomy: Computed Tomography, Ultrasound, Radiography, Gross Anatomy.
Harper & Row, New York.
Kinnunen, E., M. Ojala, H. Taskinen, and E. Matikainen (1988). Peripheral nerve
injury and Raynaud's syndrome following electric shock. Scand. J. Work Environ.
Health 14(5): 332-333.
Kirk, RE. (1982). Experimental design: Procedures for the behavioral sciences, 2nd
ed. Brooks/Cole, Belmont, CA
Kirschvink, J.L. (1989). Magnetic biomineralization and geomagnetic sensitivity in
higher animals. Bioelectromagnetics 10: 239-259.
Kirschvink, J.L. (1996). Microwave absorption by magnetite: A possible mechanism
for coupling non thermal levels of radiation to biological systems. Bio-
electromagnetics 17: 187-194.
Kiselev, AP. (1963). Threshold values of safe current at commercial frequency (in
Russian). Vopr. Elektoborud, Elekt-snabzh, i Elekt. Izmerenii Sob. MITT 17: 47-
58. CEGB Information Services, Translation no. 1167.
Klauenberg, F.J., M. Gradolfo, and D.N. Erwin (eds.) (1995). Radiofrequency
Radiation Standards, Plenum, New York, 1995.
Kloss, D.A, and E.L. Carstensen (1982). Effects of ELF electric fields on isolated
frog heart. IEEE Trans. Biomed. Eng. BME-30(6): 347-348.
526 References
Klump, D., and M. Zimmerman (1980). Irreversible differential block of A- and C-

fibers following local nerve heating in the cat. J. Physioi. (London) 298: 471-482.
Knickerbocker, G.G. (1973). Fibrillating parameters of direct and alternating
(20-Hz) currents separately and in combination-an experimental study. IEEE
Trans. Comm. COM-21(9): 1015-1027.
Knighton, RW. (1975a). An electrically evoked slow potential of the frog's retina:
I. Properties of response. J. Neurophysiol. 38: 185-197.
Knighton, RW. (1975b). An electrically evoked slow potential of the frog's retina:
II Identification with PII component of Electroretinogram. J. Neurophysiol. 38:
198-209.
Knisley, S.B., W.M. Smith, and RE. Idecker (1992). Effect of intrastimulus polarity
reversal on electric field stimulation thresholds in frog and rabbit myocardium.
J. Cardiovascular Electrophysiology 3(3): 239-254.
Kolin, A (1945). An alternating field induction flow meter of high sensitivity. Rev.
Sci. lnst. 16: 109-116.
Kolin, A (1952). Improved apparatus and techniques for electromagnetic determi-
nation of blood flow. Rev. Sci. lnst. 23: 235-242.
Koniarek, J.P. (1989). Mechanical and electrical effects of high-frequency and high-
intensity stimulation of muscle. Bioelectromagnetics 10: 335-345.
Koning, G., R Schneider, and AJ. Hoelen (1975). Amplitude-duration relation for
direct ventricular defibrillation with rectangular current pulses. Med. Bioi. Eng.
May: 388-395.
Kopeliowitch, J. (1946). The physiological effects of an alternating current and the
danger of shock in ac electrical installations. Assoc. Eng. Architects Palestine J.
(Tel Aviv, Palestine) 7(6): 2-8.
Kouwenhoven, W.B. (1949). Effects of electricity on the human body. Elec. Eng. 68:
199-203.
Kouwenhoven, W.B. (1956). Effect of capacitor discharges on the heart. AlEE
Trans. Pwr Apparat. Sys. 75, part 3, no. 23.
Kouwenhoven, W.B. (1964). The effects of electricity on the human body. Bulletin
Johns Hopkins Hosp. 114: 425.
Kouwenhoven, W.B., and O.R Langworthy (1931). Effects of electric shock-II.
Trans. AlEE 50: 1165-117l.
Kouwenhoven, W.B., and W.R Milnor (1958). The effects of high-voltage, low-
capacitance electrical discharges in the dog. IRE Trans. Biomed. Eng. PGME 11:
41-45.
Kouwenhoven, W.B., D.R Hooker, and O.R Langworthy (1932). Heart injury
from electric shock. Trans. AlEE 51: 242-244.
Kouwenhoven, W.B., D.R Hooker, and E.L. Lotz (1936). Electric shock effects of
frequency. AlEE Trans. 55: 384-386.
Kouwenhoven, W.B., G.G. Knickerbocker, RW. Chestnut, W.R Milnor, and D.J.
Sass (1959). AC shocks of varying parameters affecting the heart. Trans. AlEE
73, part III: 163-169.
Kralj, A, T. Bajd, R Turk, J. Krajnik, and R. Benko (1983). Gait restoration in
paraplegic patients: A feasibility demonstration using multichannel surface elec-
trode FES. J. Rehabil. R&D 20: 3-20.
Kralj, A, T. Bajd, R Turk, and R Benko (1986). Posture switching for prolonging
functional electrical stimulation standing in paraplegic patients. Paraplegia 24:
221-230.
References 527
Kraus, K.H., W.J. Levy, L.D. Gugino, R. Ghaly, V. Amassian, and J. Cadwell
(1994). Clinical application of transcranial magnetic stimulation for
intraoperative mapping of the motor cortex. In S. Ueno (ed.), Biomagnetic Stimu-
lation, Plenum, New York: 59-73.
Krugilikov, L.L., and H. Dertinger (1994). Stochastic resonance as a possible
mechanism of amplification of weak electric signals in living cells. Bio-
electromagnetics 14: 539-547.
Kugelberg, J. (1976). Electrical induction of ventricular fibrillation in the human
heart. Scand. I. Cardiovasc. Surg. 10: 237-240.
LaCourse, J.R, M.e. Vogt, W.T. Miller, and S.M. Selikowitz (1985). Effect of
high-frequency current on nerve and muscle tissue. IEEE Trans. Biomed. Eng.
32: 82-86.
LaCourse, J.R., M.e. Vogt, W.T. Miller, and S.M. Selikowitz (1988). Spectral analy-
sis interpretation of electrosurgical generator nerve and muscle stimulation.
Lamb, J.F., C.G. Ingram, I.A. Johnston, and RM. Pitman (1984). Essentials of
Physiology. Blackwell, Oxford.
LaMotte, R.H., RE. Torebjork, e.J. Robinson, and J.G. Thalhammer (1984).
Time-intensity profiles of cutaneous pain in normal and hyperalgesic skin: A
comparison with C-fiber nociceptor activities in monkey and human.
I. Neurophysiol. 51(6): 1434-1450.
LaMotte, R.R., and J.N. Campbell (1978). Comparison of warm and nociceptive C-
fiber afferents in monkey with human judgments of thermal pain. I. Neurophysiol.
41(6): 509-528.
Lane, J.F., and T.1. Zebo (1967). Volume potential fields developed in cats' limbs
during the passage of constant current pulses. Digest 7th Int. Con! Med. and Bioi.
Eng., Stockholm, Sweden, Aug. 14-19: 207.
Lapicque, L. (1907). Recherches quantitatives sur I'excitation electrique des nerfs
traitee comme une polarization. I. Physiol. Paris 9: 620-635.
Larkin, W.D., and J.P. Reilly (1984). Strength/duration relationships for
electrocutaneous sensitivity: Stimulation by capacitive discharges. Perception
Phychophys. 36(1): 68-78.
Larkin, W.D., and J.P. Reilly (1986). Electrocutaneous sensitivity: Effect of skin
temperature. Somatosensory Res. 3(3): 261-271.
Larkin, W.D., J.P. Reilly, and L.B. Kittler (1986). Individual differences in sensitiv-
ity to transient electrocutaneous stimulation. IEEE Trans. Biomed. Eng. BME-
33(5): 494-504.
Lassek, A.M. (1942). The human pyramidal tract. II Comparative Neurology 76:
217-225.
Lawrence, A.F., and W.R Adey (1982). Non-linear wave mechanisms in interac-
tions between excitable tissue and electromagnetic fields. I. Neurol. 4: 115.
Lazzara, R, N. EI-Sherif, and B.J. Scherlag (1975). Disorders of the cellular electro-
physiology produced by ischemia of the canine His bundle. Circ. Res. 36: 444-
454.
Leake, P.A., D.K. Kessler, and M.M. Merzenich (1990). Application and safety of
cochlear protheses. In W.F. Agnew and D.B. McCreery (eds.), Neural Protheses,
Prentice-Hall, Englewood Cliffs, NJ: 253-296.
Lednev, V.V. (1991). Possible mechanisms for the influence of weak magnetic fields
on biological systems. Bioelectromagnetics 12: 71-75.
528 References
Lee, R.C. (1990). Biophysical injury mechanisms in electrical shock victims. IEEE
Int. Con! Eng. Med. BioI. 12(4): 1502-1504.
Lee, RC. (1992). The pathophysiology and clinical management of electrical injury.
In RC. Lee, E.G. Gravalho, and J.F. Burke (eds.) (1992). Electrical Trauma,
Cambridge University Press, Cambridge/New York: 33-77.
Lee, RC., and M.S. Kolodney (1987a). Electrical injury mechanisms: Electrical
breakdown of cell membranes. Plast. Reconstr. Surg. 80(5): 672-679.
Lee, RC., and M.S. Kolodney (1987b). Electrical injury mechanisms: Dynamics of
the thermal response. Plast. Reconstr. Surg. 80(5): 663-671.
Lee, R.C., D.C. Gaylor, D. Bhatt, and D.A. Israel (1988). Role of cell membrane
rupture in the pathogenesis of electrical trauma. 1. Surg. Res. 44(6): 709-719.
Lee, R.C., E.G. Gravalho, and J.F. Burke (eds.) (1992). Electrical Trauma,
Cambridge University Press, Cambridge/New York.
Lee, RC., M. Capelli-Schellpfeffer, and K.M. Kelley (eds.) (1994a). Electrical In-
jury: A Multidisciplinary Approach to Therapy, Prevention, and Rehabilitation.
Annals New York Academy of Sciences, vol. 720, New York.
Lee, R.C., A. Myerov, and c.P. Maloney (1994b). Promising therapy for cell mem-
brane damage. In RC. Lee, M. Capelli-Schellpfeffer, and K.M. Kelley (eds.)
(1994), Electrical Injury: A Multidisciplinary Approach to Therapy, Prevention,
and Rehabilitation. Annals New York Academy of Sciences, vol. 720, New York:
239-245.
Lee, W.R (1961). A clinical study of electrical accidents. B. 1. Ind. Med. 18: 260-269.
Lee, W.R (1964). Electrophysiology. In Proc. Int. Symp. on Electrical Accidents,
International Labour Office, Geneva, Switzerland, Chap. 2.
Lee, W.R (1966). Death from electric shock. Proc. IEEE 111(1): 144-148.
Lee, W.R, and S. Zoledziowski (1964). Effects of electric shock on respiration in the
rabbit. B. 1. Ind. Med. 21: 135-144.
Lefcourt, A. (1982). Behavioral responses of dairy cows subjected to controlled
voltages, 1. Dairy Sci. 65(4): 672-674.
Lefcourt, A.M. (ed.) (1991). Effects of electrical voltage/current on farm animals:
How to detect and remedy problems. U.S. Dept. of Agriculture, Agriculture
Handbook no. 696.
Lefcourt, A.M., and RM. Akers (1982). Endocrine responses of cows subjected to
controlled voltages during milking. 1. Dairy Sci. 65: 2125-2130.
Levitt, H. (1971). Transformed up-down methods in psychoacoustics. 1. Acoust.
Soc. Am. 49: 467-477.
Levy, W.J. (1987). Clinical experience with motor and cerebellar evoked potential
monitoring. Neurosurgery 20: 169-182.
Lewis, T.H., RS. Feil, and W.D. Stroud (1920). Observations upon flutter and
fibrillation. Heart 7: 191-233.
Leyden, J.G. (1990). Death in the hot seat: A century of electrocutions. The
Washington Post, Aug. 5, p. D5.
Li, c.L., and A. Bak (1976). Excitability characteristics of the A- and C-fibers in a
peripheral nerve. Exp. Neurol. 50: 67-79.
Liberson, W.T. (1971). Progressive and alternating currents. In S.R Licht (ed.),
Electrodiognosis and Electromyography. E. Licht, New Haven, CT: 272-285.
Libet, B., W.W. Alberts, E.W. JI. Wright, L. DeLattre, G. Levin, and V. Feinstein
(1964). Production of threshold levels of conscious sensation by electrical stimu-
lation of human somatosensory cortex. 1. Neurophysiol. 27: 546-578.
References 529
Liboff, AR (1985). Geomagnetic cyclotron resonance in living cells. 1. Bioi. Phys.

13: 99-102.
Liboff, AR, and W.e. Parkinson (1991). Search for ion-cyclotron resonance in an
Na+ -transport system. Bioelectromagnetics 12: 77-83.
Liboff, AR et al. (1993). Calcium oscillations and the ELF magnetic field interac-
tions. Abstracts: BEMS 15th annual meeting, Los Angeles, June 14.
Licht, S.H. (1971). History of electrodiagnosis. In S.H. Licht (ed.), Electrodiagnosis
and Electromyography. E. Licht, New Haven, CT: 1-23.
Lin, J.C. (1978). Microwave Auditory Effects and Applications. Charles e. Thomas,
Springfield, IL.
Lin, J.e. (ed.) (1989). Electromagnetic Interaction with Biological Systems. Plenum,
New York.
Lin, J.C. (1990). Auditory perception of pulsed microwave radiation. In O.P.
Gandhi (ed.), Biological Effects and Medical Application of Electromagnetic
Energy, Prentice-Hall, 1990.
Lindermans, F.W., and J.J. Danier van der Gon (1978). Current thresholds and
liminal size in excitation of heart muscle. Cardiovasc. Res. 12: 477-485.
Litovitz, T.A, e.J. Montrose, P. Doiniv, K.M. Brown, and M. Barber (1994a).
Superimposing spatially coherent electromagnetic noise inhibits field-induced
abnormalities in developing chick embryos, Bioelectromagnetics 15: 105-113.
Litovitz, T.A, D. Krause, e.J. Montrose, and J.M. Mullins (1994b). "Temporally
incoherent magnetic fields mitigate the response of biological systems to tempo-
rally coherent magnetic fields", Bioelectromagnetics 15: 399-409.
Lochner, J.P.A, and J.F. Burger (1961). Form of the loudness function in the
presence of masking noise. 1. Acoust. Soc. Am. 33: 1705-1707.
Lord, F.M., and M.R Novick (1968). Statistical Theories of Mental Test Scores.
Addison-Wesley, Reading, MA
Lovsund, P., P.A Oberg, S.A Nilson, and T. Reuter (1980a). Magnetophosphenes:
A quantitative analysis of thresholds. Med. Bioi. Eng. Comput. 18: 326-334.
Lovsund, P., P.A Oberg, and S.E. Nilson (1980b). Magneto- and
electrosphosphenes: A comparative study. Med. Bioi. Eng. Comput. 18: 758-
764.
Lovsund, P., P.A Oberg, and S.E. Nilson (1982). ELF magnetic fields in electrosteel
and welding industries. Radio Sci. 17(5S): 35S-38S.
Luce, E.A, and S.E. Gottlieb (1984). True high-tension electrical injuries. Ann.
Plast. Surg. 12: 321-326.
Luce, E.A, W.L. Dowden, C.T. Su, and J.E. Hoopes (1978). High tension electrical
injury of the upper extremity. Surg. Gynecol. Obstet. 147: 38.
Lundborg, G.L. (1988). Nerve Injury and Repair. Churchill Livingstone, New York.
Lykken, D.T. (1971). Square-wave analysis of skin impedance. Psychophysiology
7(2): 262-275.
Maccabee, P.J., V.E. Amassian, RQ. Cracco, and J.A Cadwell (1988a). Analysis of
peripheral motor stimulation in humans using the magnetic coil. EEG Clin.
Neurophysiol. 70: 524-533.
Maccabee, P.J., V.E. Amassian, RQ. Cracco, J.B. Cracco, and B.J. Anziska (1988b).
Intracranial stimulation of facial nerve in humans with the magnetic coil. EEG
Clin. Neurophysiol. 70: 350-354.
Magovern, G.J., S.B. Park, RL. Kao, I.Y. Christlieb, and G.J. Magovern, Jf. (1990).
Dynamic cardiomyoplasty in patients. 1. Heart Transp. 9: 258-263.
530 References
Malina, RM. (1975). Growth and Development: The First Twenty Years in Man.
Burgess, Minneapolis, MN.
Mansfield, P., and P.R Harvey (1993). Limits to neural stimulation in echo-planar
imaging. Magnetic Resonance in Medicine 29: 746-758.
Marg, E. (1991). Magnetostimulation of vision: Direct noninvasive stimulation of
the retina and the brain. Optometry and Vision Science 69(6): 427-440.
Marks, L.E. (1974). Sensory Processes: The New Psychophysics. Academic Press,
New York.
Marsolais, E.B., and R Kobetic (1987). Functional electrical stimulation for walking
in paraplegia. J. Bone Joint Surg. 69A: 728-733.
Martin, J.H. (1991). Coding and processing of sensory information. In Kandel, R,
J.H. Schwartz, and T.M. Jessell (eds.), Principles of Neural Science, 3rd edition.
Elsevier, New York: 329-340.
Martin, J.H., and T.M. Jessell (1991). Modality coding in the somatic sensory
systems. In R Kandel, J.H. Schwartz, and T.M. Jessell (eds.), Principles of
Neural Science, 3rd edition. Elsevier, New York: 341-352.
Martin, J.R (1985). Receptor physiology and submodality coding in the somatic
sensory system. In E.R Kandel and J.R Schwartz (eds.), Principles of Neural
Science, 2nd edition, Elsevier, New York.
Mason, J.L., and N.AM. Mackay (1976). Pain sensations associated with
e1ectrocutaneous stimulation. IEEE Trans. Biomed. Eng. BME-23(5): 405-
409.
Massello, W., 3d (1988). Lightning deaths. Med. Leg. Bulletin 37(1): 1-7.
McAllister, RE., D. Noble, and RW. Tsien (1975). Reconstruction of the electrical
activity of cardiac Purkinje fibres. J. Physiol. 251: 1-59.
McCammon, RW. (1970). Human Growth and Development. Chartes C. Thomas,
Springfield, IL.
McCarroll, G.D., and B.A Rowley (1979). An investigation of the existence of
electrically located acupuncture points. IEEE Trans. Biomed. Eng. BME-26(3):
177-18l.
McConville, J.T., T.D. Churchill, L. Kaleps, C.E. Clauser, and J. Cuzzi (1980).
Anthropometric Relationships of Body and Body Segment Moments of Inertia.
Air Force Aerospace Medical Research Lab., Wright-Patterson Air Force Base,
Ohio.
McCreery, D.B., and W.F. Agnew (1990). Mechanisms of stimulation-induced
neural damage and their relation to guidelines for safe stimulation. In W.F.
Agnew and D.B. McCreery (eds.), Neural Prostheses: Fundamental Studies,
McKinlay, AF., S.G. Allen, P.J. Dimbylow, C.R Muirhead, and RD. Saunders
(1993). Restrictions on human exposure to static and time varying electromag-
netic fields and radiation. Documents NRPB 4(5): 7-64.
McNeal, D.R (1976). Analysis of a model for excitation of myelinated nerve. IEEE
Trans. Biomed. Eng. BME-23: 329-337.
McNeal, D.R (1977). 2000 years of electrical stimulation. In T.F. Hambrecht and
J.B. Reswick (eds.), Functional Electrical Stimulation, Marcel Dekker, New York:
3-35.
McNeal, D.R, and L.L. Baker (1988). Effects of joint angle, electrodes and wave-
form in electrical stimulation of the quadriceps and hamstrings. Annals Biomed.
Eng. 16: 299-310.
References 531
McNeal, D.R, and B.R Bowman (1985a). Peripheral neuromuscular stimula-

tion. In J.B. Mykleburst, J.F. Cusick, A. Sances, and S.J. Larsons (eds.), Neural
Stimulation, vol. II. CRC Press, Boca Raton, FL: 95-118.
McNeal, D.R, and B.R Bowman (1985b). Selective activation of muscles using
peripheral nerve electrodes. Med. Bioi. Eng. Comput. 23: 249-253.
McNeal, D.R, and D.A. Teicher (1977). Effect of electrode placement on threshold
and initial site of excitation of a myelinated nerve fiber. In T.F. Hambrecht and
J.B. Reswick (eds.), Functional Electrical Stimulation. Marcel Dekker, New York:
405-412.
McNeal, D.R, B.R Bowman, and W.L. Momsen (1973). Peripheral block on motor
activity. In M. Gavrilovic and A.B. Wilson (eds.), Advances in External Control
of Human Extremities, Yugoslav Committee for Electronics and Automation,
Belgrade: 575-583.
McNeal, D.R, R Waters, and J. Reswick (1977). Experience with implanted elec-
trodes. Neurosurgery I: 228-229.
McRobbie, D., and M.A. Foster (1984). Thresholds for biological effects of time-
varying magnetic fields. Clin. Phys. Physiol. Measurements 5(2): 67-78.
McRobbie, D., and M.A. Foster (1985). Cardiac response to pulsed magnetic fields
with regard to safety in NMR imaging. Phys. Med. Bioi. 30(7): 695-702.
Melzak, R, and P.D. Wall (1965). Pain mechanisms: A new theory. Science
150(3699): 971-979.
Memberg, W.D., P.H. Peckham, and M. Keith (1994). Surgically-implanted intra-
muscular electrode for an implantable neuromuscular stimulation system. IEEE
Trans. Rehab. Engin. 2(2): 80-91.
Merton, P.A., and H.B. Morton (1986). A magnetic stimulation for the human
motor cortex. 1. Physiol. 381: lOP.
Meyer, R.A., and J.N. Campbell (1981). Myelinated nociceptive afferents accounts
for the hyperalgesIa that follows a burn to the hand. Science 213: 1527-1529.
Mild, K.H., and M. Sandstrom (1994). Health aspects of electric and magnetic fields
from VDTs. In J.e. Lin (ed.), Advances in Electromagnetic Fields in Living
Systems, vol. 1, Plenum, New York: 155-183.
Mills, K.R, and N.M.F. Murray (1985). Corticospinal tract conduction time in
multiple sclerosis. Annals Neurol. 18: 601-610.
Mines, G.R (1914). On circulating excitations in the heart muscle and their possible
relation to tachycardia and fibrillation. Trans. R. Soc. Can. 8: 43.
Mogul, D.I., N.V. Thakor, J.R McCullogh, G.A. Meyers, RE. Teneick, and D.R
Siniger (1984). Modified Beeler-Reuter model yields improved simulation of
myocardial action potentials. Proc. IEEE Con!, Computers in Cardiology, Salt
Lake City, Utah, Sept. 18-21: 159-162.
Moncrief, J.A., and B.A. Pruitt, Jr. (1971). Hidden damage from electrical injury.
Geriatrics 26(4): 84-85.
Monster, A.W., and R Chan (1977). Isometric force produced by motor units of
extensor digitorum communis muscle in man. 1. Neurophysiol. no. 40: 1432-
1443.
Montagu, M.F.A. (1960). A Handbook of Anthropometry. Charles C. Thomas,
Springfield, IL.
Moreira, L.F., E.A. Bocchi, N.S. Stolf, G. Bellotti, and A.D. Jatene (1996). Dynamic
cardiomyoplasty in the treatment of dilated cardiomyopathy: Current results and
perspectives. 1. Card. Surg. 11(3): 207-216.
532 References
Moriarty, B.J., and J.N. Char (1987). Electrical injury and cataracts-an unusual
case. West Indian Med. J. 36(2): 114-116.
Moritz, AR, and F.e. Henrigves (1947). Studies of thermal injury II. The relative
importance of time and surface temperature in the causation of cutaneous burns.
Am. J. Pathol. 23: 695-720.
Mortimer, J.T. (1981). Motor prostheses. In J.M. Brookhart, V.B. Mountcastle, V.B.
Brooks, and S.R Geiger (eds.), Handbook of Physiology, Section 1: The Nervous
System. Vol. II. Motor Control, Part I, Am. Physiol. Soc., Bethesda, MD.
Mortimer, J.T., W.F. Agnew, K. Horch, P. Citron, G. Creasy, and e. Kantor (1995).
Perspectives on new electrode technology for stimulating peripheral nerves with
implantable motor prostheses. IEEE Trans. Rehab. Engin. 3(2): 145-154.
Morton, D.J. (1944). Manual of Human Cross Section Anatomy. Williams &
Wilkins, Baltimore.
Morton, R, and K.A Provins (1960). Finger numbness after acute local exposure to
cold. J. Appl. Physiol. 15: 149-154.
Moskowitz, RR, B. Scharf, and J.e. Stevens (1974). Sensation and Measurement.
D. Reidel, Boston.
Motz, H., and F. Rattay (1986). A study of the application of the Hodgkin-Huxley
and the Frankenhaeuser-Huxley model for electrostimulation of the acoustic
nerve. Neuroscience 18: 699-712.
Mouchawar, G.A, L.A Geddes, J.D. Bourland, and J.A Pearce (1989). Ability of
the Lapicque and Blair strength-duration curves to fit experimentally obtained
data from the dog heart. IEEE Trans. Biomed. Eng. 36(9): 971-974.
Mouchawar, G.A, J.D. Bourland, L.A Geddes, and J.A Nyenhuis (1991). Mag-
netic electrophrenic nerve stimulation to produce inspiration. Annals Biomed.
Eng. 19: 219-221.
Mouchawar, G.A, J.D. Bourland, J.D. Nyenhuis, J.A Geddes, L.A Foster, K.S.
Jones, and G.P. Graber (1992). Closed chest cardiac stimulation with a pulsed
magnetic field. Med. Bioi. Eng. Comput. 30: 162-168.
Mueller, E.F., R Loeffel, and S. Mead (1953). Skin impedance in relation to pain
threshold testing by electrical means. J. Appl. Physiol. 5: 746-752.
Murphy, K.P., Y. Zhao, and M. Kawai (1996). Molecular forces involved in force
generation during skeletal muscle contraction. J. Exp. Bioi. 199(12): 2565-2571.
MWN (1995). Swedish VDT emissions standard goes international. Microwave
News, March/April: 8-9.
Myklebust, I., A Sances, M. Chilbert, T. Prieto, and T. Swiontek (1985). Capacitive
discharge studies. In J.E. Bridges, G.L. Ford, LA Sherman, and M. Vainberg
Nagarajan, S.S., D.M. Durand, and E.N. Warman (1993). Effects of induced electric
fields on finite neuronal structures: A simulation study. IEEE Trans. Biomed.
Eng. 40(11): 1175-1188.
Nahin, P.J. (1987). Oliver Heaviside, Sage in Solitude. IEEE Press, New York.
Nannini, N., and K. Horch (1991). Muscle recruitment with intrafascicular elec-
trodes. IEEE Trans. Biomed. Engin. 38: 769-776.
Naples, G.G., J.T. Mortimer, A Scheiner, and J.D. Sweeney (1988). A spiral nerve
cuff electrode for peripheral nerve stimulation. IEEE Trans. Biomed. Eng. BME-
35: 905-916.
Naples, G.G., J.T. Mortimer, and T.G.R Yuen (1990). Overview of peripheral
nerve electrode design and implantation. In W.F. Agnew and D.B. McCreery
References 533
(eds.), Neural Prostheses: Fundamental Studies, Prentice-Hall, Englewood Cliffs,

NJ: 107-145.
National Radiological Protection Board (1983). Revised guidance on acceptable
limits of exposure during nuclear magnetic resonance clinical imaging. Brit. J.
Radiol. 56: 974-977.
NAVSEA (1982). Electromagnetic radiation hazards. NAVSEA OP 3565/
NAVAIR 16-1-529/NACELEX 0967-LP-624-601O, vol. I, 5th revision. Published
by Naval Sea Systems Command, Washington, DC.
NCRP (1986). Biological Effects and Exposure Criteria for Radio-frequency Electro-
magnetic Fields, Pub. No. 86, National Council on Radiation Protection and
Measurements, Washington, D.C.
Netherlands (1997). Radio frequency electromagnetic fields (300 Hz-300 GHz).
Health Council of the Netherlands, Radiation Committee, Rijswijk, publication
No. 1997/0l.
Nethken, RT., and M.A Bulot (1967). Threshold of electrical signals on the upper
arm. Proc. IEEE Region III Convention, Jackson, MI: 81-83.
Neumann, E., AE. Sowers, and C.A Jordan (1989). Electroporation and
Electrofusion in Cell Biology. Plenum, New York.
Newman, AL. (1984). Self-injurious behavior inhibiting system. Johns Hopkins
APL Tech. Digest 5(3): 290-295.
NFPA (1990). National Electrical Code. ANSIINFPA 70-1990, National Fire
Protection Association.
Nicholson, P.W. (1965). Specific impedance of cerebral white matter. Exp. Bioi. 13:
386--401.
Niinami, H., K. Greer, H. Koyanagi, and L. Stephenson (1996). Skeletal
muscle ventricles: Another alternative for heart failure. J. Card. Surg. 11(4): 280-
287.
Noble, D. (1962). A modification of the Hodgkin-Huxley equations applicable to
Purkinje fibre action and pace-maker potentials. J. Physiol160: 317-353.
Noble, D. (1984). The surprising heart: A review of recent progress in cardiac
electrophysiology. J. Physiol. 353: 1-50.
Noble, D., and RB. Stein (1966). The threshold conditions for initiation of action
potentials by excitable cells. J. Physiol. 187: 129-162.
Noble, D., and S.J. Noble (1984). A model of sino-atrial node electrical activity
based on a modification of the DiFrancesco-Noble (1984) equations. Proc. R.
Soc. London B222: 295-304.
Nochomovitz, M. (1983). Electrical activation of respiration. EMBS Magazine 2(3):
27-31.
Norden, J., and C. Ramel (1992). Interaction Mechanisms of Low-Level Electromag-
netic Fields in Living Systems, Oxford University Press, Oxford, UK.
Norell, RJ., RJ. Gustafson, RD. Appleman, and J.B. Overmeire (1983). Behav-
ioral studies of dairy cattle; sensitivity to electrical currents. Trans. ASAE, 26:
1506--1511.
Notermans, S.L.H., and M.M.W.A Tophofif (1975). Sex differences in pain
tolerance and pain apperception. In M. Weisenberg (ed.), Pain: Clinical and
Experimental Perspectives, Mosby, St. Louis, MO: 111-116.
Notermans, S.L.R (1966). Measurement of the pain threshold determined by elec-
trical stimulation and its clinical application, Part I: Method and factors possibly
influencing the pain threshold. Neurology 16: 1071-1086.
534 References
Notermans, S.L.R (1967). Measurement of the pain threshold determined by elec-

trical stimulation and its clinical application, Part II: Clinical applications in
neurological and neurosurgical patients. Neurology 17: 58-73.
NRPB (1991). Board Statement: Principles for the protection of patients and volun-
teers during clinical magnetic resonance diagnostic procedures. Documents of the
NRPB 2(1): 1-5.
NRPB (1993). Documents of the NRPB, 4(5), National Radiological Protection
Board, Chilton, UK.
Nute, R (1985). Dynamic aspects of body impedance. In J.E. Bridges, G.L. Ford,
I.A Sherman, and M. Vainberg (eds.), Electrical Shock Safety Criteria,
Pergamon, New York: 173-181.
Nyenhuis, J.A, J.D. Bourland, G.A Mouchawar, T.Z. Elabbady, L.A Geddes, D.J.
Schaefer, and M.E. Riehl (1990). Comparison of stimulation effects of longitudi-
nal and transverse MRI gradient coils. Society for Magnetic Resonance in Med.,
Proc. 10th Annual Meeting, San Francisco: 1275.
Nyenhuis, J.A, J.D. Bourland, D.J. Schaefer, K.S. Foster, W.E. Schoelein, G.A
Mouchawar, T.Z. Elabbady, L.A Geddes, and M.E. Riehl (1992). Measurement
of cardiac stimulation thresholds for pulsed z-gradient fields in a 1.5-T magnet.
Society for Magnetic Resonance in Med., Proc. 11th Annual Meeting, Berlin: 586.
Nyenhuis, J.A., J.D. Bourland, G. Mouchawar, L. Geddes, K. Foster, J. Jones, W.
Schoenlein, G. Garber, and T. Elabbady (1994). Magnetic stimulation ofthe heart
and safety issues in magnetic resonance imaging. In S. Ueno (ed.), Biomagnetic
Stimulation, Plenum, New York: 75-98.
Nyenhuis, J.A, J.D. Bourland, and D.J. Schaefer (1997). Analysis from a stimula-
tion perspective of the field patterns of magnetic resonance imaging gradient
coils. 1. Appl. Phys. 81(8): 4314-4316.
Oester, Y.T., and S.H. Licht (1971). Routine electrodiagnosis. In S.H. Licht (ed.),
Electrodiagnosis and Electromyography. E. Licht, New Haven, CT: 201-217.
Oh, J.H., V. Badhwar, and RC. Chiu (1996). Mechanisms of dynamic car-
diomyoplasty: Current Concepts. 1. Card. Surg. 11(3): 194-199.
Olsen, RW., L.J. Hayes, E.R Wissler, R Nikaidoh, and Re. Eberhart (1985).
Influence of hypothermia and circulatory arrest on cerebral temperature distribu-
tions. ASME 1. Biomech. Eng. 107: 354-360.
Omura, Y. (1977). Critical evaluation of the methods of measurement of "tingling
threshold" and "pain tolerance" by electrical stimulation. Acupuncture & Electro-
therapy Res. Int. 1. 2: 161-236.
Osepchuk, J. (1989). Panel discussion on standards. In J.e. Lin (ed.), Electromag-
netic Interaction with Biological Systems, Plenum, New York: 281-289.
Osepchuk, J. (1994). Impact of public concerns about low-level electromagnetic
fields (EMF) on interpretation of EMF/radiofrequency (RFR) data base. In B.J.
Klauenberg, M. Grandolfo, and D.N. Erwin (eds.), Radiofrequency Radiation
Standards, Plenum, New York: 415-426.
Osypka, P. (1963). Quantitative investigation of current strength, duration, and
routing in ac electrocution accidents involving human beings and animals.
Elektromedizen 8, Sonderdruck Fachverlag Schiele and Schon, Berlin, Transla-
tion by SLA Translations Center, TT 66-1588 & TT-11470.
Paintal, AS. (1967). A comparison of the nerve impulses of mammalian non-
medullated nerve fibers with those of the smallest diameter medullated fibers.
1. Physiol. 193: 523-533.
References 535
Paintal, AS. (1973). Conduction in mammalian nerve fibers. In J.E. Desmedt (ed.),
New Developments in Electromyography and Clinical Neurophysiology, vol. 2,
Karger, Basel, Switzerland: 19-41.
Palin, W.E., Jr., AM. Sadove, J.E. Jones, W.F. Judson, and RD. Stambaugh (1987).
Oral electrical burns in a pediatric population. 1. Oral Med. 42(1): 17-21,34.
Panescu, D., K.P. Cohen, and J.G. Webster (1993). The mosaic characteristics of the
skin. IEEE Trans. Biomed. Eng. 40(5): 43~39.
Panescu, D., J.G. Webster, and RA Stratbucker (1994a). A nonlinear electrical-
thermal model of the skin. IEEE Trans. Biomed Eng. 41(7): 671--680.
Panescu, D., J.U.G. Webster, and RA Stratbucker (1994b). A nonlinear finite
element model of the electrode-electrolyte-skin system. IEEE Trans. Biomed.
Eng. 41(7): 681--688.
Papoulis, A (1965). Probability, Random Variables, and Stochastic Processes.
McGraw-Hill, New York.
Parry, C.B.W. (1971). Strength-duration curves. In S.H. Licht (ed.), Electro-
diagnosis and Electromyography, E. Licht, New Haven, CT: 241-271.
Parshley, P.F., J. Kilgore, J.P. Pulito, P.W. Smiley, and S.R Miller (1985). Aggressive
approach to the extremity damaged by electric current. Am. 1. Surg. 150(1): 78-82.
Pearce, J.A, J.D. Bourland, W. Neilsen, L.A Geddes, and M. Voelz (1982). Myo-
cardial stimulation with ultrashort duration current pulses. PACE 5: 52-58.
Peckham, P.R (1983). Restoration of upper extremity function. EMBS Magazine
2(3): 30-32.
Peckham, P.R (1987). Functional electrical stimulation: Current status and future
prospects of applications to the neuromuscular system in spinal cord injury.
Paraplegia 25: 279-288.
Peckham, P.H., and D.B. Gray (1996). Single topic issue: Functional neuromuscular
stimulation (FNS). 1. Rehab. Res. Develop. 33(2): ix-xi.
Peleska, B. (1963). Cardiac arrhythmias following condenser discharges and their
dependence upon the strength of current and phase of cardiac cycle. Circ. Res. 13:
21-32.
Peleska, B. (1965). Cardiac arrhythmias following condenser discharges led through
an inductance. Circ. Res. 16: 11-18.
Pelzer, D., and W. Trautwein (1987). Currents through ionic channels in multicellu-
lar cardiac tissues and single heart cells. Experientia 43: 1153-1162.
Pennes, RR (1948). Analysis of issue and arterial blood temperatures in the resting
human forearm. 1. Appl. Physiol. 1: 93-122.
Persson, B.R, and F. Stahlberg (1989). Health and Safety of Clinical NMR
Examinations, CRC Press, Boca Raton, FL.
Peter, J.B., RJ. Barnard, V.R Edgerton, C.A Gillespie, and K.E. Stempel (1972).
Metabolic profiles of three fiber types of skeletal muscle in guinea pigs and
rabbits. Biochemistry 11: 2627-2633.
Petersen, RC. (1991). Radiofrequency/microwave protection guides. Health
Physics 61(1): 59--67.
Petersen, R.C. (1995). Safety Standards Setting. Course book from Rutgers
University short course: Management of Electromagnetic Energy Hazards, Las
Vegas, Oct. 16-19.
Peterson, D.K., M. Nochomovitz, AF. DiMarco, and J.T. Mortimer (1986).
Intramuscular electrical activation of the phrenic nerve. IEEE Trans. Biomed.
Engin. 33(3): 342-351.
536 References
Peterson, D.K., M.L. Nochomovitz, T.A Stellato, and J.T. Mortimer (1994a).
Long-term intramuscular electrical activation of the phrenic nerve: Safety and
reliability. IEEE Trans. Biomed. Engin. 41(12): 1115-1126.
Peterson, D.K., M.L. Nochomovitz, T.A Stellato, and J.T. Mortimer (1994b).
Long-term intramuscular electrical activation of the phrenic nerve: Efficacy as a
ventilatory prosthesis. IEEE Trans. Biomed. Engin. 41(12): 1127-1135.
Pethig, R (1979). Dielectric and Electronic Properties of Biological Materials. John
Wiley & Sons Chichester, UK.
Petrofsky, J.S. (1978). Control of the recruitment and firing frequencies of motor
units in electrically stimulated muscles in the cat. Med. BioI. Eng. Comput. 16:
302-308.
Petrofsky, J.S., and C.A Phillips (1983). Computer controlled walking in the
paralyzed individual. J. Neurol. Orthoped. Surg. 4: 153-164.
Pette, D., and G. Vrobova (1985). Invited review: Neural control of phenotypic
expression in mammalian muscle fibers. Muscle and Nerve 8: 676-689.
Pfeiffer, E.A (1968). Electrical stimulation of sensory nerves with skin electrodes
for research, diagnosis, communication, and behavioral conditioning: A survey.
Med. Bioi. Eng. 6: 637-651.
Plonsey, R (1969). Bioelectric Phenomena. McGraw-Hill, New York.
Plonsey, R, and RC. Barr (1986). A critique of impedance measurements in cardiac
tissue. Annals Biomed. Eng. 14: 307-322.
Plonsey, R, and RC. Barr (1988). Bioelectricity. Plenum, New York.
Plonsey, R, and RC. Barr (1995). Electric field stimulation of excitable tissue.
Polk, C. (1986). Introduction. In C. Polk and E. Postow (eds.), CRC Handbook of
Biological Effects of Electromagnetic Fields. CRC Press, Boca Raton, FL: 1-24.
Polk, C. (1993). Physical mechanisms for biological effects of ELF low-intensity
electric and magnetic fields: Thermal noise limit and counterion polarization. In
M. Blank (ed.), Electricity and Magnetism in Biology and Medicine, San Francisco
Press, San Francisco: 543-546.
Polk, c., and E. Postow (eds.) (1996). Handbook of Biological Effects of Electro-
magnetic Fields, second edition, CRC Press, Boca Raton, FL.
Polson, M.J.R, AT. Barker, and I.L. Freeston (1982). Stimulation of nerve trunks
with time-varying magnetic fields. Med. BioI. Eng. Comput. 20: 243-244.
Polson, M.J.R, AT. Barker, and S. Gardiner (1982). The effect of rapid rise-time
magnetic fields on the ECG of the rat. Clin. Physiol. Measurements 3: 231-234.
Ponten, B., U. Erikson, S.R. Johansson, and L. Olding (1970). New observations on
tissue changes along the pathway of the current in an electrical injury. Case
report. Scand. J. Plast. Reconstr. Surg. 4(1): 75-82.
Postow, E., and M.L. Swicord (1996). Modulated fields and "window" effects. In
C. Polk and E. Postow (eds.), Handbook of Biological Effects of Electromagnetic
Fields, second edition, CRC Press, Boca Raton, FL: 535-580.
Prieto, T., A Sances, J. Myklebust, and M. Chilbert (1985). Analysis of cross-body
impedance at household voltage levels. In J.E. Bridges, G.L. Ford, I.A Sherman,
and M. Vainberg (eds.), Electric Shock Safety Criteria, Pergamon, New York:
151-160.
Procacci, B.G. (1968). A study on the cutaneous pricking pain threshold in normal
man. In A Soulairac, J. Cahn, and J. Charpentier (ed.), Pain, Academic Press,
London.
References 537
Procacci, P., M. Zoppi, M. Maresca, and S. Romano (1974). Studies on the pain
threshold in man. In J.J. Bonica (ed.), Advances in Neurology, vol. 4, International
Symposium on Pain, Raven, New York.
Project UHV (1982). Transmission Line Reference Book, 345 kV and Above, 2nd ed.
The Electric Power Research Institute, Palo Alto, CA.
Provins, K.A., and R Morton (1960). Tactile discrimination and skin temperature.
J. App. Physiol. 15: 155-160.
Pruna, S., C. lonescu-Tirogoviste, E. Popa, and I. Mincu (1989). Measurement of
perception threshold to an electrical stimulus using a phase-sensitive technique in
normal and diabetic subjects. Med. Bioi. Eng. Compt. 27: 111-116.
Quiring, D.P. (1944). Surface area determination. In O. Glasser (ed.), Medical
Physics, vol. 1, Year Book Publishers, Chicago: 1490-1494.
Rafferty, E.B., H.L. Green, and I.e. Gregory (1975a). Disturbances of heart
rhythms produced by 50 Hz leakage currents in dogs. Cardiovasc. Res. 9: 256-
262.
Rafferty, E.B., RL. Green, and M.R Yacoub (1975b). Disturbances of heart
rhythm by 50Hz leakage currents in human subjects. Cardiovasc. Res. 9: 263-265.
Rail, W. (1977). Core conductor theory and cable properties of neurons. In Hand-
book of Physiology: A Critical, Comprehensive Presentation of Physiological
Knowledge and Concepts, vol. 1, American Physiological Society, Bethesda, MD:
39-97.
Ranck, J.B. (1963). Specific impedance of rabbit cerebral cortex. Exp. Neurol. 7:
144-152.
Ranck, J.B. (1975). Which elements are excited in electrical stimulation of mamma-
lian central nervous system: A review. Brain Res. 98: 417-440.
Rasch, P.J. (1989). Kinesiology and Applied Anatomy, 7th ed., Lea & Febiger,
Philadelphia.
Rattay, F. (1986). Analysis of models for external stimulation ofaxons. IEEE Trans.
Biomed. Eng. 33: 974-977.
Rattay, F. (1988). Modeling the excitation of fibers under surface electrodes. IEEE
Trans. Biomed. Eng. 35(3): 199-202.
Rattay, F. (1989). Analysis of models for extracellular fiber stimulation. IEEE
Trans. Biomed. Eng. 36(3): 676--682.
Reilly, J.P. (1978a). Electric field induction on sailboats and vertical poles. IEEE
Trans. Pwr. Apparat. Sys. PAS-97(4): 1373-138l.
Reilly, J.P. (1978b). Electric and magnetic coupling from high voltage AC power
transmission lines--classification of short-term effects on people. IEEE Trans.
Pwr. Apparat. Sys. PAS-97(6): 2243-2252.
Reilly, J.P. (1979a). Electric field induction on long objects-A methodology for
transmission line impact studies. IEEE Trans. Pwr. Apparat. Sys. PAS-98(6):
1841-1849.
Reilly, J.P. (1979b). An approach to the realistic-case analysis of electric field
induction from AC transmission lines. Third Int. Symp. High Voltage Engineer-
ing, Milan, Italy.
Reilly, J.P. (1980). Spark discharge characteristics of vehicles energized by AC
electric field. JHU PPSE T-16, The Johns Hopkins University Applied Physics
Laboratory, Laurel, MD.
Reilly, J.P. (1982). Characteristics of spark discharges from vehicles energized by
AC electric fields. IEEE Trans. Pwr. Apparat. Sys. PAS-101(9): 3178-3186.
538 References
Reilly, J.P. (1988). Electrical models for neural excitation studies. fohns Hopkins
Reilly, J.P. (1989). Peripheral nerve stimulation by induced electric currents: Expo-
sure to time-varying magnetic fields. Med. BioI. Eng. Comput. 27: 101-110.
Reilly, J.P. (1991). Magnetic field excitation of peripheral nerves and the heart:
A comparison of thresholds. Med. Bioi. Eng. Comput. 29(6): 571-579.
Reilly, J.P. (1992). Electrical Stimulation and Electropathology. Cambridge Univer-
sity Press, Cambridge, 1992.
Reilly, J.P. (1993). Safety considerations concerning the minimum threshold for
magnetic excitation of the heart. Med. Bioi. Eng. Comput. 31: 651-654.
Reilly, J.P. (1994). Transient current effects in stray voltage exposure: Biophysical
principles and mechanisms. Paper No. 943594, Amer. Soc. Ag. Eng. International
Meeting, Atlanta.
Reilly, J.P. (1995). Nerve stimulation of cows and other farm animals by time-
varying magnetic fields. Trans. Am. Soc. Ag. Eng. 38(5): 1487-1494.
Reilly, J.P. (1998). Maximum Pulsed Electromagnetic Field Limits Based on Periph-
eral Nerve Stimulation: Application to IEEE/ANSI C95.1 Electromagnetic Field
Standards. IEEE Trans. Biomed. Eng. 45(1): 137-14l.
Reilly, J.P., and RH. Bauer (1987). Application of a neuroelectric model to
electrocutaneous sensory sensitivity: Parameter variation study. IEEE Trans.
Biomed. Eng. BME-34(9): 752-754.
Reilly, J.P., and M. Cwiklewski (1978). A realistic-case analysis of electric field
induction on vehicles near AC transmission lines. IEEE Can. Conf. Communica-
tions and Power, Montreal.
Reilly, J.P., and M. Cwiklewski (1981). Rain gutters near high-voltage power lines:
A study of electric field induction. IEEE Trans. Pwr. Apparat. Sys. PAS-100(4):
2068-2076.
Reilly, J.P., and A.M. Diamant (1997). Theoretical evaluation of peripheral nerve
stimulation during MRI with an implanted spinal fusion stimulator, Mag. Res.
Imaging 15(10): 1145-1156.
Reilly, J.P., and W.D. Larkin (1983). Electrocutaneous stimulation with high
voltage capacitive discharges. IEEE Trans. Biomed. Eng. BME-30: 631-64l.
Reilly, J.P., and W.D. Larkin (1984). Understanding electric shock. fohns Hopkins
Reilly, J.P., and W.D. Larkin (1985a). Human reactions to transient electric
currents-summary report. PPSE T-34(NTIS No. PB 86-117280/AS), The Johns
Hopkins University Applied Physics Laboratory, Laurel, MD.
Reilly, J.P., and W.D. Larkin (1985b). Mechanisms for human sensitivity to
transient electric currents. In J.E. Bridges, G.L. Ford, I.A. Sherman, and M.
Vainberg (eds.), Electrical Shock Safety Criteria, Pergamon, New York: 241-
249.
Reilly, J.P., and W.D. Larkin (1987). Human sensitivity to electric shock induced by
power frequency electric fields. IEEE Trans. Electromagnetic Compatibility
EMC-29(3): 221-232.
Reilly, J.P., W. Larkin, RJ. Taylor, and V.T. Freeman (1982). Human reactions to
transient electric currents, annual report, July 1981-July 1982. CPE-8203(NTIS
No. PB83 204628), The Johns Hopkins University Applied Physics Laboratory,
Laurel, MD.
Reilly, J.P., W. Larkin, RJ. Taylor, V.T. Freeman, and L.B. Kittler (1983). Human
reactions to transient electric currents, annual report, July 1982-June 1983.
References 539
Report CPE-8305(NTIS No. PB84-112895), The Johns Hopkins University Ap-

plied Physics Laboratory, Laurel, MD.
Reilly, J.P., W.D. Larkin, L.B. Kittler, and V.T. Freeman (1984). Human reactions
to transient electric currents-annual report, July 1983-June 1984. Report CPE-
8313(NTIS No. PB84--231463), The Johns Hopkins University Applied Physics
Laboratory, Laurel, MD.
Reilly, J.P., V.T. Freeman, and W.D. Larkin (1985). Sensory effects of transient
electrical stimulation-evaluation with a neuroelectric model. IEEE Trans.
Biomed. Eng. BME-32(12): 1OOl-101l.
Reilly, J.P., H.A Eaton, and D. Gluck (1993). A novel method for improving
focality of magnetic stimulation of the brain. Abstracts: Bioelectromagnetics
Society Annual Meeting, Los Angeles CA, June 14-17.
Reinemann, D.J., L.E. Stetsen, and N. Laughlin (1994). Effects of frequency and
duration on the sensitivity of dairy cows to transient voltages. Paper No. 943597,
Amer. Soc. Ag. Eng. International Meeting, Atlanta.
Reinemann, D.J., L.E. Stetson, J.P. Reilly, N.K. Laughlin, S. McGuirk, and S.D.
LeMire (1996). Dairy cow sensitivity and aversion to short duration transient
currents. Paper 963087, American Society Agricultural Engineers International
Meeting, Phoenix, AZ.
Reinemann, D.J., L.W. Stetson, J.P. Reilly, and N.K. Laughlin (1998). Sensitivity of
dairy cows to short duration currents. Trans. Amer. Soc. Ag. Eng.
Riscili, C.E., K.S. Foster, W.D. Voorhees, J.D. Bourland, and L.A Geddes (1988).
Electroventilation in the baboon. Amer. 1. Emerg. Med. 6: 561-565.
Robblee, L.S., and T.L. Rose (1990). Electrochemical guidelines for selection of
protocols and electrode materials for neural stimulation. In W.F. Agnew and
D.B. McCreery (eds.), Neural Prostheses: Fundamental Studies, Prentice-Hall,
Englewood Cliffs, NJ: 25-66.
Rodgers, S.J. (1981). Radiofrequency burn hazards in the MF/HF band. In J.e.
Mitchell (ed.), Proceeding: Workshop on the protection of personnel against
radiofrequency electromagnetic radiation, Review 3-81, USAF School of Medi-
cine, Brooks AFB, Texas: 76-89.
Rollman, G.B. (1969). Electrocutaneous stimulation: Psychometric functions and
temporal integration. Perception Psychophys. 5(5): 289-293.
Rollman, G.B. (1974). Electrocutaneous stimulation. In F.A Geldard (ed.), Confer-
ence on Cutaneous Communication Systems and Devices, Monterey, CA, 1973,
The Psychonomic Society, Austin, TX: 38-5l.
Rollman, G.B. (1975). Behavioral assessment of peripheral nerve function. Neurol-
ogy 26: 339-342.
Rollman, G.B., and G. Harris (1987). The detectability and perceived magnitude of
painful electrical shock. Perception Phychophys. 42(3): 257-268.
Ronner, S.F. (1990). Electrical excitation of CNS neurons. In W.F. Angew and
D.B. McCreery (eds.), Neural Protheses, Prentice-Hall: 169-196.
Rosenberg, D.B., and M. Nelson (1988). Rehabilitation concerns in electrical burn
patients: A review of the literature. 1. Trauma 28(6): 808-812.
Rosenblueth, A, and J. Garcia Ramos (1947). Studies on flutter and fibrillation. II.
The influence of artificial obstacles on experimental, auricular flutter. Am. Heart
1.33: 677.
RosIer, K.M., C.W. Hess, R. Reckmann, and H.P. Ludin (1989). Significance of
shape and size of the stimulating coil in magnetic stimulation of the human motor
cortex. Neurosci. Lett. 100: 347-352.
540 References
Rosner, B.S. (1961). Neural factors limiting cutaneous spatio-temporal discrimina-

tion. In W.A Rosenblith (ed.), Sensory Communication, MIT Press, Cambridge,
MA: 725-737.
Rosner, B.S., and W.R Goff (1967). Electrical response of the nervous system and
subjective scales of intensity. In W.D. Neff (ed.), Contributions to Sensory Physi-
ology, vol. 2, Academic Press, New York.
Roth, B. (1989). Interpretation of skeletal muscle four-electrode impedance mea-
surements using spatial and temporal frequency-dependent conductivities. Med.
Bioi. Eng. Comput. 27: 491-495.
Roth, B.J. (1995). A mathematical model of make and break electrical stimulation
of cardiac tissue by a unipolar anode or cathode. IEEE Trans. Biomed. Eng.
42(12): 1174-1184.
Roth, B.J., J.M. Saypol, M. Hallett, and L.G. Cohen (1991). A theoretical calcula-
tion of the electric field induced in the cortex during magnetic stimulation.
Electroencephalog. Clin. Neurophysiol. 81: 47-56.
Rothberger, C.J., and R Winterberg (1941). Uber Vorhofflimmern und Vorhof-
fiattern. Pflugers Arch. 160: 42-90.
Rouse, RG., and AR Dimick (1978). The treatment of electrical injury compared
to burn injury: A review of pathophysiology and comparison of patient manage-
ment protocols. f. Trauma 18: 43-47.
Roy, O.Z. (1980). Summary of cardiac fibrillation thresholds for 60-Hz currents
and voltages applied directly to the heart. Med. Bioi. Eng. Comput. 18: 657-659.
Roy, O.Z., J.R Scott, and G.c. Park (1976). 60-Hz ventricular fibrillation and pump
failure thresholds versus electrode area. IEEE Trans. Biomed. Eng. BME-23(1):
45-48.
Roy, O.Z., G.c. Park, and J.R Scott (1977). Intracardiac catheter fibrillation thresh-
olds as a function of the duration of 60-Hz current and electrode area. IEEE
Trans. Biomed. Eng. BME-24(5): 430-435.
Roy, O.z., AJ. Mortimer, B.J. Trollope, and E.J. Villeneuve (1985). Electrical
stimulation of the isolated rabbit heart by short duration transients. In J.E.
Bridges, G.L. Ford, LA Sherman, and M. Vainberg (eds.), Electrical Shock Safety
Criteria, Pergamon, New York: 77-86.
Roy, O.Z., J.R Scott, and B.J. Trollope (1986). 60Hz ventricular fibrillation
thresholds for large-surface-area electrodes. Med. BioI. Eng. Comput. 24: 471-
474.
Roy, O.Z., B.J. Trollope, and J.R Scott (1987). Measurement of regional cardiac
fibrillation thresholds. Med. BioI. Eng. Comput. 25: 165-166.
Rozman, J., B. Sovinec, M. Trlep, and B. Zorko (1993). Multielectrode spiral
cuff for ordered and reversed activation of nerve fibres. f. Biomed. Engin. 15:
113-120.
Rubenstein, J.T. (1993). Axon termination conditions for electrical stimulation.
Ruch, S., J.A Abildskov, and R McFee (1963). Resistivity of body tissues at low
frequencies. Circ. Res. 12: 40-50.
Ruch, T.C. (1979). Somatic sensation: Receptors and their axons. In T.e. Ruch and
H.D. Patton (eds.), Physiology and Biophysics, W.B. Saunders, Philadelphia:
157-200.
Ruch, T.e., and H.D. Patton (1979). Physiology and Biophysics. W.B. Saunders,
Philadelphia.
References 541
Ruch, T.C., RD. Patton, J.W. Woodbury, and AL. Towe (1968). Neurophysiology.
W.B. Saunders, Philadelphia.
Ruiz, E.V., J.A Russo, G.y' Savino, and M.E. Valentinuzzi (1985). Ventricular
fibrillation threshold in the dog determined with defibillating paddles. Med. Bioi.
Eng. Compo 23: 281-284.
Rush, S., and D.A Driscoll (1968). Current distribution in the brain from surface
electrodes. Current Researches 47(6): 717-723.
Sachs, RM., J.D. Miller, and K.W. Grant (1980). Perceived magnitude of multiple
electrocutaneous pUlses. Perception Psychophys. 28: 255-262.
Sackeim, H.A, P. Decina, S. Portnoy, P. Neeley, and S. Maliz (1987). Studies of
dosage, seizure threshold, and seizure duration in ECT. Bioi. Psychiatry 22: 249-
268.
Sagan, L.A (1996). Electric and Magnetic Fields: Invisible Risks? Gordon & Breach,
Australia.
Sagan, P.M., M.E. Stell, G.K. Bryan, and W.R Adey (1987). Detection of 6O-Hertz
vertical electric fields by rats. Bioelectromagnetics 8: 303-313.
Sances, A, S.J. Larson, J. Myklebust, and J.F. Cusick (1979). Electrical injuries.
Surg. Gynecol. Obstet. 149(1): 97-108.
Sances, A, J.B. Myklebust, S.J. Larson, J.e. Darin, T. Swiontek, T. Prieto,
M. Chilbert, and J.F. Cusick (1981a). Experimental electrical injury studies.
J. Trauma 21(8): 589-597.
Sances, A, J.B. Myklebust, J.F. Szablya, T.J. Swiontek, S.J. Larson, M. Chilbert,
T. Prieto, and J.P. Cusick (1981b). Effects of contacts in high voltage injuries.
IEEE Trans. Pwr. Apparat. Sys. PAS-100(6): 2987-2992.
Sances, A, J.F. Szyblya, J.D. Morgan, J.B. Myklebust, and S.J. Larson (1981c). High
voltage powerline injury studies. IEEE Trans. Pwr. Apparat. Sys. PAS-100(2):
552-558.
Sances, A, J.B. Myklebust, J.F. Szablya, T.J. Swiontek, S.J. Larson, and M. Chilbert,
et al. (1983). Current pathways in high-voltage injuries. IEEE Trans. Biomed.
Eng. BME-30(2): 118-124.
Sandove, AM., J.E. Jones, T.R Lynch, and P.W. Sheets (1988). Appliance therapy
for perioral electrical bums: A conservative approach. J. Burn Care Rehabil. 9(4):
391-395.
Sanguinetti, M.e., and N.K. Jurkiewicz (1991). Delayed rectifier outward K+
current is composed of two currents in guinea pig atrial cells. Amer. J. Physiol.,
Heart and Circulatory Physiol. 260: H393-H399.
Sasyniuk, G.!', and C. Mentez (1971). A mechanism for reentry in canine ventricular
tissue. Circ. Res. 28: 3-15.
Sato, M., and J. Ushiyama (1950). On the relation of strength-frequency curve in
excitation by low frequency AC to the minimal gradient of the nerve fiber. Jpn. J.
Physiol. 1: 141-146.
Saunders, F.A (1974). Electrocutaneous displays. In F.A Geldard (ed.), Confer-
ence on Cutaneous Communication Systems and Devices, Monterey, CA, 1973,
Psychonomic Society, Austin, TX: 20-26.
Saunders, RD. (1991). Limits on patient and volunteer exposure during clinical
magnetic resonance diagnostic procedures: Recommendations for the practical
implementation of the Board's statement. Documents of the NRPB 2(1) 5-29.
Schaefer, D.J. (1992). Dosimetry and effects of MR exposure to RF and switched
magnetic fields. In RL. Magin, RP. Liburdy, and B. Persson (eds.), Biological
542 References
Effects and Safety Aspects of Nuclear Resonance Imaging and Spectroscopy,

New York Academy of Sciences, New York.
Schaefer, D.J., J.D. Bourland, J.A Nyenhuis, K.S. Foster, W.F. Wirth, L.A Geddes,
and M.E. Riehl (1994). Determination of gradient-induced human peripheral
nerve stimulation thresholds for trapezoidal pulse trains. Soc. Mag. Res., Proc.
2nd Annual Meeting, San Francisco: lOI.
Schaefer, D.J., J.D. Bourland, J.A Nuyenhuis, K.S. Foster, P.E. Licato, and L.A
Geddes (1995). Effects of simultaneous gradient combinations on human periph-
eral nerve stimulation thresholds. Society for Magnetic Resonance in Med., Proc.
12th Annual Meeting, Nice, France, Poster No. 1220.
Schenk, J.F., C.L. Dumoulin, C.L. Redington, RW. Kressel, H.Y. Elliot, and I.L.
McDougall (1992). Human exposure to 4.0 Tesla magnetic fields in a whole-body
scanner. Medical Physics 19(4): 1089-1098.
Scherf, D. (1947). Studies on auricular tachycardia caused by aconitine administra-
tion. Proc. Soc. Exp. Bioi. Med. 64: 233-239.
Schludermann, E., and J.P. Zubeck (1962). Effect of age on pain sensitivity.
Perceptual and Motor Skills. 14: 295-30l.
Schmid, E. (1961). Temporal aspects of cutaneous interaction with two-point
electrical stimulation. J. Exp. Psychol. 67: 191-192.
Schmidt, R (ed.) (1978). Fundamentals of Sensory Physiology. Springer-Verlag,
New York.
Schmidt-Nielsen, K. (1984). Scaling: Why Is Animal Size So Important? Cambridge
University Press, Cambridge.
Schmitt, F.P. Wielopolski, H. Fischer, and RR Edelman (1994). Peripheral stimu-
lations and their relation to gradient pulse shapes. Proc. Soc. Magnetic Resonance
in Medicine, 2nd Annual Meeting, San Francisco: 102.
Schriefer, T.N., K.R Mills, N.M. Murray, and C.W. Hess (1988). Evaluation of
proximal facial nerve conduction by transcranial magnetic stimulation. J. Neurol.
Neurosurg. Psychiatr. 51: 60--66.
Schwan, H.P. (1954). Die elektrischen Eigenschaften von muskelgewebe bie
Niederfrequenz. Z. Naturforsche, 96: 245-25l.
Schwan, H. (1968). Electrical impedance of the human body. Technical Report TR-
2199, U.S. Naval Weapons Laboratory, Dahlgren, VA, NTIS No. AD 842306.
Schwan, H.P. (1966). Alternating current electrode polarization. Biophysik. 3: 181-
20l.
Scott, J.P., W.P. Lee, and S. Zoledziowski (1973). Ventricular fibrillation thresholds
for AC. shock of long duration in dogs with normal acid-base states. Br. J. Ind.
Med. 30: 155.
Scott, W.T. (1966). The Physics of Electricity and Magnetism. John Wiley & Sons,
New York.
Sepulveda, N.G., J.P. Wikswo, and D.S. Echt (1990). Finite element analysis of
cardiac defibrillation current distributions. IEEE Trans. Biomed. Eng. 37(4): 354-
365.
Sharma, M., and A Smith (1978). Paraplegia as a result of lightning injury. Br. Med.
J. 12: 1464-1465.
Sharp, G.H., and RW. Joyner (1980). Stimulated propagation of cardiac action
potential. Biophys. J. 31: 403-424.
Shellock, F.G., and E. Kanal (1994). Magnetic Resonance Bioeffects, Safety, and
Patient Management. Raven, New York.
References 543
Shields, C.B., H.L. Edmonds, M. Paloheimo, J.R Johnson, and RT. Holt (1988).
Intraoperative use of transcranial magnetic motor-evoked potentials. Proc. 10th
Ann. Conf IEEE-EM BS: 926-927.
Shimada, Y., K. Sato, E. Abe, H. Kagaya, K. Ebata, M. Oba, and M. Sato (1996).
Clinical experience of functional electrical stimulation in complete paraplegia.
Spinal Cord 34(10): 615-619.
Silny, J. (1986). The influence of threshold of the time-varying magnetic field in
the human organism. In J.H. Bernhardt (ed.), Biological Effects of Static and
Extremely Low Frequency Magnetic Fields, MMV Medzin Verlag, Munchen,
Germany.
Silva, M., N. Hummon, D. Ruttor, and C. Hooper (1989). Power frequency magnetic
fields in the home. IEEE Trans. Pwr. Del. 4(1): 465-478.
Silverglade, D. (1983). Splinting electrical burns utilizing a fixed splint technique:
A report of 48 cases. ASDC J. Dent. Child. 50(6): 455-458.
Silversides, J. (1964). The neurological sequelae of electrical injury. Calif Med.
Assoc. J. 91: 195-204.
Skoog, T. (1970). Electrical injuries. J. Trauma. 10: 816-830.
Skuggevig, W. (1992). Standards and protective measures. Chap. 11 in J.P. Reilly,
Electrical Stimulation and Electropathology, Cambridge University Press, Cam-
bridge/New York.
Slager, c.J., J.C. Schuurbiers, J.A Oomen, and N. Bom (1993). Electrical nerve and
muscle stimulation by radio frequency surgery: Role of direct current loops
around the active electrode. IEEE Trans. Biomed. Eng. 40(2): 182-187.
Smith, B.T., M.J. Mulcahey, and RR Betz (1996). Development of an upper
extremity FES system for individuals with C4 tetraplegia. IEEE Trans. Rehab.
Engin. 4(4): 264-270.
Smith, L. (1990). Electrocutions involving consumer products. Memorandum of
March 30, 1990, U.S. Consumer Product Safety Commission, Washington, DC.
Smoot, AW. (1985). The seventh meeting of IEC TC74IWG5. Memorandum of
April 11, 1985, Underwriters Laboratories.
Smoot, AW., and J. Stevenson (1968a). Investigation of reaction current. Technical
Report of April 1968, Underwriters Laboratories.
Smoot, AW., and J. Stevenson (1968b). Report on investigation of reaction cur-
rents. Technical Report of Dec. 3, 1968, Underwriters Laboratories.
Smyth, P.D., P.P. Tarjan, E. Chernof, and N. Baker (1976). The significance of
electrode surface area and stimulating thresholds in permanent cardiac pacing.
J. Thorac. Card. Surg. 71(4): 559-565.
Snyder, RG., M.L. Spencer, C.L. Owings, and L.W. Schneider (1975).
Anthropometry of US infants and children. SAE Automotive Engineering
Congress and Exposition, Paper No. 750423, as Referenced by AF. Roche and
RM. Malina (1983), Manual of Physical Status and Performance in Childhood,
Vol. IB: Physical Status, Plenum, New York.
Solem, L., RP. Fischer, and RG. Strate (1977). The natural history of electrical
injury. J. Trauma. 17(7): 487-492.
Solomonow, M., E. Eldred, J. Lyman, and J. Foster (1983). Control of muscle
contractile force through indirect high-frequency stimulation. Am. J. Phys. Med.
62: 71-82.
Song, W.J., S. Weinbaum, and L.M. Jiji (1988). A combined macro and microvascu-
lar model for whole limb heat transfer. J. Biomed. Eng. 110(4): 259-268.
544 References
Spach, M.S., W.T. Miller, D.B. Geselowitz, RC. Barr, J.M. Kootsey, and E.A.
Johnson (1981). The discontinuous nature of propagation in normal canine car-
diac muscle. Evidence for recurrent discontinuities of intracellular resistance that
affect the membrane currents. Circ. Res. 48: 39-54.
Spach, M.S., W.T. Miller, P.C. Dolber, J.M. Kootsey, J.R Sommer, and C.E.
Mosher (1982). The functional role of structural complexities in the propagation
of depolarization in the atrium of the dog. Circ. Res. 50: 175-19l.
Spiegel, RJ. (1976). Magnetic coupling to a prolate spheroid model of man. IEEE
Trans. Pwr. Apparat. Sys. PAS-96(1): 208-212.
Starmer, C.F., and Whalen, RE. (1973). Current density and electrically induced
ventricular fibrillation. Med. Instrum. 7(1): 3-6.
Stein, RB. (1980). Nerve and Muscle. Plenum, New York.
Steiner, U.E., and T. Ulrich (1989). Magnetic field effects in chemical kinetics and
related phenomena. Chern. Rev. 89: 51-147.
Sternbach, RA., and B. Tursky (1964). On the psychophysical power function in
electric shock. Psychol. Sci. 1: 217-218.
Sternbach, RA., and B. Tursky (1965). Ethnic differences among housewives in
psychophysical and skin potential responses to electric shock. Psychophysiology
1: 241-246.
Sten-Knudsen, O. (1960). Is muscle contraction initiated by internal current flow?
J. Physiol. 151: 363-384.
Stern, S., and V.G. Laties (1985). 60Hz electric fields: Detection by female rats.
Stern, S., V.G. Laties, e.V. Stancompiano, e. Cox, andJ.O. deLorge (1983). Behav-
ioral detection of 60-Hz electric fields by rats. Bioelectromagnetics. 4: 215-247.
Stevens, J.C. (1980). Thermo-tactile interactions: Some influences of temperature
on touch. In D.R Kenshalo (ed.), Sensory Functions of the Skin of Humans,
Plenum, New York: 207-222.
Stevens, J.C., J.D. Mack, and S.S. Stevens (1960). Growth of sensation on seven
continua as measured by force of handgrip. J. Exp. Psycho I. 59: 60-67.
Stevens, J.e., L.E. Marks, and D.C. Simonson (1974). Regional sensitivity and
spatial summation in the warmth sense. Physiol. Behav. 13: 825-836.
Stevens, J.C., B.G. Green, and Krimsley (1977). Punctuate pressure sensitivity:
Effects of skin temperature. Sensory Processes 1: 238-243.
Stevens, RG., B.W. Wilson, and L.F. Anderson (1997). The Melatonin Hypothesis:
Breast Cancer and Use of Electric Power. Battelle Press, Columbus.
Stevens, S.S. (1959). Cross-modality validations of subjective scales for loudness,
vibration, and electric shock. J. Exp. Psychol. 57: 201-209.
Stevens, S.S. (1966). Matching functions between loudness and ten other continua.
Perception Psychophys. 1: 5-8.
Stevens, S.S. (1975). Psychophysics: Introduction to Its Perceptual Neural, and Social
Prospects. John Wiley & Sons, New York.
Stevens, W.G.S. (1963). The current-voltage relationship in human skin. Med.
Electron. Bioi. Eng. 1: 389-399.
Stevenson, J. (1969). Progress report on investigation of reaction current. Technical
Report Nov. 18, 1969, Underwriters Laboratories.
Stevenson, J. (1971). Reaction to leakage currents. UL Lab. Data 2(1): 16--19.
Underwriters Laboratories.
References 545
Stoy, RD., K.R Foster, and H.P. Schwan (1982). Dielectric properties of mamma-
lian tissues from 0.1 to 100MHz: A summary. Phys. Med. Bioi. 27(4): 501-513.
Strasser, E.J., RM. David, and M.J. Mehshey (1977). Lightning injuries. l. Trauma.
17(4): 315-319.
Struijk, J.J., J. Holsheimer, and H.B. Boom (1993). Excitation of dorsal root fibers
in spinal cord stimulation: A theoretical study. IEEE Trans. Biomed. Eng. 40(7):
632-653.
Stuchly, M.A, and D.W. Lecuyer (1989). Exposure to electromagnetic fields in arc
welding. Health Phys. 56(3): 297-302.
Stuchly, M.A, and S.S. Stuckly (1980). Dielectric properties of biological
substances-tabulated. l. Mic. Pwr. 15(1): 19-26.
Stuchley, M.A, and S.S. Stuchley (1996). Experimental radio and microwave dosim-
etry. In C. Polk and E. Postow (eds.), Biological Effects of Electromagnetic Fields,
CRC Press, Boca Raton, FL.
Suchi, T. (1954). Experiments on electrical resistance of the human epidermis. lpn.
l. Physiol. 5: 75-80.
Sugimoto, T., S.F. Schaal, and AG. Wallace (1967). Factors determining vulnerabil-
ity to ventricular fibrillation induced by 60-cps alternating current. Circ. Res. 21:
601-608.
Sunde, E.D. (1968). Earth Conduction Effects in Transmission Systems, Macmillan,
New York.
Sunderland, S. (1978). Nerves and Nerves Injuries, Churchill Livingstone, New
York.
Sweeney, J.D. (1993). A theoretical analysis of the "let-go" phenomenon. IEEE
Trans. Biomed. Eng. 40(12): 1335-1338.
Sweeney, J.D., J.T. Mortimer, and D. Durand (1987). Modeling of mammalian
myelinated nerve for functional neuromuscular stimulation. Proc. 9th Ann. Int.
Conf of the IEEE-EMBS: 1577-1578.
Sweeney, J.D., K. Deng, E. Warman, and J.T. Mortimer (1989). Modeling of electric
field effects on the excitability of myelinated motor nerve. Proc. 11th Ann. Int.
Conf of the IEEE-EMBS: 1281-1282.
Sweeney, J.D., D. Ksienski, and J.T. Mortimer (1990). A nerve cuff technique for
selective activation of peripheral nerve trunk regions. IEEE Trans. Biomed. Eng.
BME-31: 706-715.
Sweeney, J.D., N.R Crawford, and T.A Brandon (1995). Neuromuscular stimula-
tion selectivity of multiple-contact nerve cuff electrode arrays. Med. BioI. Eng.
Comput. 33: 418-425.
Sweeney, J.D., S.F. Cogan, J.T. Mortimer, and K. Horch (1996). Electrodes, leads,
and connectors. l. Rehab. Res. Develop. 33(2): 194-197.
Szeto, AY., and F.A Saunders (1982). Electrocutaneous stimulation for sensory
communication in rehabilitation engineering. IEEE Trans. Biomed. Eng. BME-
25(4): 300-308.
Takagi, T., and T. Muto (1971). Influence upon human bodies and animals of
electrostatic induction caused by 500kV transmission lines (English transl.),
Tokyo Electric Power.
Takemoto-Hambleton, RM., W.J. Dunseath, and W.T. Joines (1988). Electromag-
netic fields induced in a person due to devices radiating in the 10Hz to 100kHz
range. IEEE Trans. Elect. Compat. 30(4): 529-537.
546 References
Takeuchi, A, and N. Takeuchi (1962). Electrical changes in pre- and post-synaptic

axons of the giant synapse of Loligo, J. Gen. Physiol. 45: 1181-1193.
Tanner, J.A (1962). Reversible blocking of nerve conduction by alternating current
excitation. Nature 195: 712-713.
Tasaki, I. (1953). Nervous Transmission. Charles e. Thomas, Springfield, IL.
Tasaki, I. (1982). Physiology and Electrochemistry of Nerve Fibers. Academic Press,
New York.
Tasaki, I., and M. Sato (1951). On the relation of the strength-frequency curve in
excitation by alternating current to the strength-duration and latent addition
curves of the nerve fiber. 1. Gen. Physiol. 34: 373-388.
Tashiro, T., and A Higashiyama (1981). The perceptual properties of
electrocutaneous stimulation: Sensory quality, subjective intensity, and intensity-
duration relation. Perception Psychophys. 30(6): 579-586.
Taussig, H. (1968). Death from lightning and the possibility of living again. Ann.
Intern. Med. 68: 1345-1349.
Taylor, D.N., J.G. Wallace, and J.e. Masdeu (1992). Perception of different fre-
quencies of cranial transduatneous electrical nerve stimulation in normal and
HIV-positive individuals. Perceptual and Motor Skills 74: 259-264.
Taylor, RJ. (1985). Body impedance for transient high voltage currents. In
J.E. Bridges, G.L. Ford, LA Sherman, and V. Vainberg (eds.), Electric Shock
Safety Criteria, Symposium on Electric Shock Safety Criteria, Pergamon: 251-258.
Teghtsoonian, R (1973). Range effects in psychological scaling and a revision of
Stevens' law. Am. J. Psychol. 86: 3-27.
Teicher, D.A, and D.R McNeal (1978). Comparison of a dynamic and steady-state
model for determining nerve fiber threshold. IEEE Trans. Biomed. Eng. BME-
25(1): 105-107.
Teissie, J., and M. Rols (1994). Manipulation of cell cytoskeleton affects the lifetime
of cell membrane electropermeabilization. In Re. Lee, M. Capelli-Schellpfeffer,
and K.M. Kelley (eds.). Electrical Injury: A Multipliciplinary Approach to
Therapy, Prevention, and Rehabilitation, Annals New York Academy of Sciences,
vol. 720, New York: 98-110.
Tenforde, T.S. (1989). Electroreception and magnetoreception in simple and
complex organisms. Bioelectromagnetics 10: 215-22l.
Tenforde, T.S. (1993). Cellular and molecular pathways of extremely-low-frequency
electromagnetic field interactions with living systems. In M. Blank (ed.), Electricity
and Magnetism in Biology and Medicine, San Francisco Press, San Francisco: 1-8.
Tenforde, T.S., e.T. Gaffey, B.R Moyer, and T.F. Budinger (1983). Cardiovascular
alterations in Macaca monkeys exposed to stationary magnetic fields: Experimen-
tal observations and theoretical analysis. Bioelectromagnetics 4: 1-9.
Tessier-Lavigne, M. (1991). Phototransduction and information processing in the
retina. In E.R Kandell, J.H. Schwartz, and T.M. Jessell (eds.), Principles of
Neural Science, third edition, Elsevier, New York: 400-418.
Thalen, H.J.T., J.V.P. Berg, J.N. Heide, and J. Nieveen (1975). The Artificial
Cardiac Pacemaker. Van Gorum, Assem, The Netherlands.
Thompson, G. (1933). Shock threshold fixes appliance insulation resistance. Electri-
cal World 101: 793-795.
Torebjork, H.E., and RG. Hallin (1973). Perceptual changes accompanying con-
trolled preferential blocking of A and C fibre responses in intact human skin
nerves. Exp. Brain Res. 16: 321-322.
References 547
Trautwein, W., and J. Dudel (1954). Actionspotential und Mechanogramm des

Warmbliiterherzmuskels als Funktion der Schlagfrequenz. PflUgers Arch. ges.
Physiol. 260: 24-39.
Treagear, RT. (1966). Physical Functions of Skin. Academic Press, London.
Triolo, R, R Nathan, Y. Handa, M. Keith, R Betz, S. Carroll, and C. Kantor
(1996). Challenges to clinical deployment of upper limb neuroprostheses.
J. Rehab. Res. Develop. 33(2): 111-122.
Tsuji, S., Y. Mural, and M. Yarita (1988). Somatosensory potentials evoked by
magnetic stimulation of lumbar roots, cauda equina, and leg nerves. Annals
Neurol. 24: 568-573.
Tucker, RD., and D.H. Schmitt (1978). Tests for human perception of 60Hz mod-
erate strength magnetic field. IEEE Trans. Biomed. Eng. BME-25(6): 509-518.
Tucker, RD., D.H. Schmitt, c.B. Sievert, and S.E. Silvis (1984). Demodulated low
frequency currents from electrosurgical procedures. Surgery, Gynecology &
Obstetrics 159: 39-43.
Tursky, B., and P.D. Watson (1964). Controlled physical and subjective intensities
of electric shock. Psychophysiology 1(2): 151-162.
Tyler, D., and D. Durand (1993). Design and acute test of a radially penetrating
interfascicular nerve electrode. Proc. 15th Ann. Int. Con! IEEE-EMBS 15: 1247-
1248.
Tyler, D., and D. Durand (1994). Interfascicular electrical stimulation for selectively
activating axons. EMBS Magazine 13(4): 575-583.
Ueno, S. (ed., 1994a). Biomagnetic Stimulation, Plenum, New York.
Ueno, S. (1994b). Focal and vectorial magnetic stimulation of the human brain. In
S. Ueno (ed.), Biomagnetic Stimulation, Plenum, New York: 29-47.
Ueno, S., K. Harada, C. Ji, and Y. Domura (1984). Magnetic nerve stimulation
without interlinkage between nerve and magnetic flux. IEEE Trans. Magnet.
MAG-20(5): 1660-1662.
Ueno, S., T. Tashiro, and K. Harada (1988). Localized stimulation of neural tissues
in the brain by means of a paired configuration of time-varying magnetic fields.
J. Appl. Phys. 64: 5862-5864.
UL (1945). Measurement of electric shock hazard in radio equipment. Bulletin of
Research No. 33, Underwriters Laboratories.
UL (l975). Method of development-revision-implementation-standards for
safety. Publication F 200-46 3M575, Underwriters Laboratories.
UL (1981). Development of test equipment and methods for measuring potentially
lethal and otherwise damaging current levels. Consumer Product Safety Commis-
sion, Contract Number CPSC-C-79-1034, Underwriters Laboratories.
UL (1983). Standard for Safety, double insulation systems for use in electrical
equipment. UL 1097, second edition, Underwriters Laboratories.
UL (1985). Standard for safety, ground-fault circuit interrupters. UL 943, 2nd ed.,
Underwriters Laboratories.
UL (1987). Standard for safety, telephone equipment. UL 1459, 2nd ed., December
18,1987, Underwriters Laboratories.
UL (1988). Electric shock-A safety seminar on theory and prevention. Underwriters
Laboratories.
UL (1990). Report to the instrumentation committee of ANSI ClOl, Preliminary
report by A.W. Smoot, J. Stevenson, W. Myrick, and W. Tuthill. Underwriters
Laboratories.
548 References
Vallbo, AB., K.A Olsson, K.G. Westberg, and F.J. Clark (1984). Microstimulation
of single tactile afferents from the human hand. Brain 107(3): 727-749.
van Boxtel, A (1977). Skin resistance during square-wave electrical pulses of 1 to
lOrnA Med. Bioi. Eng. Compo 15: 679-fJ87.
van den Honert, C., and J.T. Mortimer (1979a). The response of the myelinated
nerve fiber to short duration biphasic stimulating currents. Annals Biomed. Eng.
7: 117-125.
van den Honert, c., and J.T. Mortimer (1979b). Generation of unidirectionally
propagated action potentials in peripheral nerve by brief stimuli. Science 206:
1311-1312.
Veltink, P.H., B.K. van Veen, J.J. Struijk, J. Holsheimer, and H.B.K. Boom (1989a).
A modeling study of nerve fascicle stimulation. IEEE Trans. Biomed. Eng. BME-
36(7): 683-fJ92.
Veltink, P.H., J.A van Alste, and H.B.K. Boom (1988). Influences of stimulation
conditions on recruitment of myelinated nerve fibers: A model study. IEEE
Trans. Biomed. Eng. BME-35: 917-924.
Veltink, P.H., J.J. Hermens, and J.A van Alste (1989b). Multielectrode
intrafascicular and extraneural stimulation. Med. Bioi Eng. Comput. 27: 19-24.
Veraart, C., W.M. Grill, and J.T. Mortimer (1993). Selective control of muscle
activation with a multipolar nerve cuff electrode. IEEE Trans. Biomed. Eng. 40:
640-fJ53.
Verrillo, RT. (1979). Comparison of vibrotactile threshold and suprathreshold
responses in men and women. Perception Psychophys. 26: 20--24.
Verrillo, RT. (1982). Effects of aging on the suprathreshold responses to vibration.
Perception Psychophys. 32: 61-68.
Verillo, RT., and G.A Gesheider (1979). Psychophysical measurements of en-
hancement, suppression, and surface gradient effects in vibrotaction. In D.R
Kenshalo (ed.), Sensory Functions of the Skin of Humans, Plenum, New York:
153-18l.
Vodovnik, L., T. Bajd, F. Gracanin, A Kralj, and P. Strojnik (1981). Functional
electrical stimulation for control of locomotor systems. CRC Crit. Rev. Bioeng. 6:
63-132.
Vodovnik, L., W.J. Crochetiere, and J.B. Reswick (1967). Control of a skeletal joint
by electrical stimulation of antagonists. Med. Bioi. Eng. 5: 97-109.
Volta, A (1800). On the electricity excited by the mere contact of conducting
substances of different kinds. Phil. Trans. R. Soc. London 90: 403-43l.
Voorhees, C.R, W.D. Voorhees, L.A Geddes, and J.D. Bourland (1992). The
chronaxie for myocardium and moter nerve in the dog with chest-surface elec-
trodes. IEEE Trans. Biomed. Eng. 39(6): 624--628.
Voorhes, W.D., K.S. Foster, L.A Geddes, and C.F. Babbs (1983). Safety factor for
transchest pacing. Proc., 36th ACEMB Conf, Columbus, Ohio, Sept. 12-14: 19.
Voorhees, W.D., L.A Geddes, J.D. Bourland, and G. Mouchawar (1990). Magneti-
cally induced contraction of the inspiratory muscles in dog. 1. Clin. Eng. 15(5):
407-409.
Wachtel, H. (1979). Firing pattern changes and transmembrane currents produced
by extremely low frequency fields in pacemaker neurons. In RD. Phillips, M.F.
Gillis, W.T. Kaune, and D.D. Mahlum (eds.), Biological Effects of Extremely
Low-Frequency Electromagnetic Fields, Publication CONF-781016, Technical
Information Center, U.S. Dept. of Energy: 132-146.
References 549
Wallace, J.G., and J.e. Masdeu (1992). Perception of different frequencies of cranial
transcutaneous electrical nerve stimulation in normal and HIV-positive individu-
als. Perceptual and Motor Skills 74: 259-264.
Walleczek, J. (1994). Immune cell interactions with extremely low frequency mag-
netic fields: Experimental verification and free radical mechanisms. Chap. 12 in
AH. Frey (ed.), On the Nature of Electromagnetic Field Interaction with Biologi-
cal Systems, RG. Landes Co., Austin TX.
Walter, J.S., P. Griffith, J. Sweeney, V. Scarpine, M. Bidnar, J. McLane, and
e. Robinson (1997). Multielectrode nerve cuff stimulation of the median nerve
produces selective movements in a raccoon animal model. J. Spinal Cord Med. 20:
233-243.
Walthard, K.M., and M. Thicaloff (1971). Motor points. In S.H. Licht (ed.),
Electrodiagnosis and Electromyography, E. Licht, New Haven, CT: 153-170.
Wang, X.W., and W.H. Zoh (1983). Vascular injuries in electrical burns-
the pathologic basis for mechanism of injury. Burns Incl. Therm. Inj. 9(5): 335-
338.
Wang, X.W., e.s. Lu, N.Z. Wang, H.e. Lin, H. Su, J.N. Wei, and W.Z. Zoh (1984),
High tension electrical burns of upper arms treated by segmental excision of
necrosed humerus. An introduction of a new surgical method. Burns Incl. Therm.
Inj. 10(4): 271-281.
Wang, X.W., B.B. Roberts, RL. Zapata-Sirvent, W.A Robinson, J.P. Waymack,
E.J. Law, B.G. MacMillan, and J.W. Davies (1985). Early vascular grafting to
prevent upper extremity necrosis after electrical burns. Commentary on indica-
tions for surgery. Burns Incl. Therm. Inj. 11(5): 359.
Wang, X.W., E.J. Bartle, and B.B. Roberts (1987). Early vascular grafting to pre-
vent upper extremity necrosis after electric burns: Additional commentary on
indications for surgery. J. Burn Care Rehabil. 8(5): 391-394.
Warren, RM. (1981). Measurement of sensory intensity. Behav. Brain Sci. 4: 175-
223.
Waters, RL., D.R McNeal, and J. Perry (1975). Experimental correction of
footdrop by electrical stimulation of the peroneal nerve. J. Bone Joint Surg. 57-A:
1047-1054.
Waters, RL., D.R McNeal, W. Faloon, and B. Clifford (1985). Functional electrical
stimulation of the peroneal nerve for hemiplegia. J. Bone Joint Surg. 67-A: 792-
793.
Watson, AB., J.S. Wright, and J. Loughman (1973). Electrical thresholds for
ventricular fibrillation in man. Med. J. Austral. 1, June 16: 1179-1182.
Weaver, J.C. (1992). Cell membrane rupture by strong electric fields: Prompt and
delayed processes. In Lee, Re., E.G. Gravalho, and J.F. Burke (eds.) (1992).
Electrical Trauma, Cambridge University Press, CambridgeINew York: 301-
326.
Weaver, J.e., and RD. Astumian (1990). The response of living cells to very weak
electric fields: The thermal noise limit. Science 247: 459-462.
Weaver, L., R Williams, and S. Rush (1976). Current density in bilateral and
unilateral EeT. Bioi. Physiciatry 11(3): 303-311.
Wedensky, N. (1884). Wie Rasch ermudet der Nerv. Z. Med. Wissen. 22: 65-68.
Weigel, RJ., RA Jaffe, D.L. Lunstorm, W.e. Forsythe, and L.E. Anderson
(1987). Stimulation of cutaneous mechanoreceptors by 60-Hz electric field.
Bioelectromagnetics 8:337-350.
550 References
Weigria, R, G.K. Moe, and c.J. Wiggers (1941). Comparison of the vulnerable
periods and fibrillation threshold of normal and idioventricular beats. Am. 1.
Physiol. 132: 651-657.
Weinstein, S. (1963). The relationship of laterality and cutaneous area to breast-
sensitivity in sinistrals and dextrals. Am. 1. Psychol. 76: 475-479.
Weinstein, S. (1968). Intensive and extensive aspects of tactile sensitivity as a
function of body part, sex, and laterality. In D.R Kenshalo (ed.), The Skin Senses,
Charles C. Thomas, Springfield, IL: 195-218.
Weinstein, S. (1978). New methods for the in-vivo assessment of skin smoothness
and skin softness. 1. Soc. Cosmet. Chem. 29: 99-115.
Weirich, J., S. Hohnloser, and H. Antoni (1983). Factors determining the suscepti-
bility of the isolated guinea pig heart to ventricular fibrillation induced by sinusoi-
dal alternating current at frequencies from 1 to 1,OOOHz. Basic Res. Cardiol. 78:
604-615.
Weirich, J., K. Haverkampf, and A. Antoni (1985). Ventricular fibrillation of the
heart induced by electric current. Rev. Gen. de ['Elect. 11: 833-843.
Weiss, G. (1901). Sur la possibilite de rendre comparables entre eux les appareils
servant a l'excitation electrique. Arch. Ital. Bioi. 35: 413-446.
Werner, G., and V.B. Mountcastle (1968). Quantitative relations between mechani-
cal stimuli to the skin and neural response evoked by them. In D.R Kenshalo
(ed.), The Skin Senses, Charles C. Thomas, Springfield, IL: 112-138.
Wessale, J.L., J.D. Bourland, W.A. Tacker, and L.A. Geddes (1980). Bipolar
catheter defibrillation in dogs using trapezoidal waveforms of various tilts.
1. Electrocardiol. 13: 359-366.
Wessale, J.L., L.A. Geddes, G.M. Ayers, and K.S. Foster (1992). Comparison of
rectangular and exponential current pulses for evoking sensation. Annals
Biomed. Eng. 20: 237-244.
Wetherill, G.B. (1963). Sequential estimation of quantal response curves. 1. R. Stat.
Soc. B25: 1-48.
Whitaker, H.B. (1939). Electric shock hazard as it pertains to the electric fence.
Underwriters Laboratories, Bulletin Res. 14: 3-56.
Whittleson, W.G., M.M. Mullord, R Kilgour, and L.R Cate (1975). Electric shocks
during machine milking. New Zealand Vet. lournal23: 105-108.
Wiggers, c.J., and R Wegria (1939). Ventricular fibrillation due to single, localized
induction and condenser shocks applied during the vulnerable phase of ventricu-
lar systole. Am. 1. Physiol. 128: 500-505.
Williams, D.O., B.J. Scherlag, RR Hope, N. EI-Sherif, and R Lazzara (1974). The
pathophysiology of malignant ventricular arrhythmias during acute myocardial
ischemia. Circulation 50: 1163-1172.
Williams, J.H., and G.A. Klug (1995). Calcium exchange hypothesis of skeletal
muscle fatigue: A brief review. Muscle & Nerve 18(4): 421-434.
Williams, N.S., J. Patel, B.D. George, et al. (1991). Development of an electrically
stimulated neoanal sphincter. Lancet 338: 1166-1169.
Willis, RJ., and W.M. Brooks (1984). Potential hazards of NMR imaging. No
evidence of the possible effects of static and changing magnetic fields on cardiac
function of the rat and guinea pig. Magn. Resonance Imag. 2: 89-95.
Wilson, B.W., RG. Stevens, and L.A. Anderson (1990). Extremely Low Frequency
Electromagnetic Fields: The Question of Cancer. Battelle Press, Columbus,
OH.
References 551
Wilson, C.M., J.D. Allen, J.B. Bridges, and AA Adgey (1988). Death and damage
caused by multiple direct current shocks: Studies in an animal model. Eur. Heart
1. 9(11): 1257-1265.
Wit, AL., P.E. Cranefield, and B.F. Hoffman (1972). Slow conduction and reentry
in the ventricular conducting system. II. Single and sustained circus movement in
networks of canine and bovine Purkinje fibers. Circ. Res. 30: 11-22.
Wolff, B.B., and S. Langley (1975). Cultural factors and the response to pain. In M.
Weisenberg (ed.), Pain: Clinical and Experimental Perspectives, Mosby, st. Louis:
144-151.
Woodbury, J.W. (1968). Action potential: Properties of excitable membranes. In
T.e. Ruch, H.D. Patton, J.W. Woodbury, and AL. Towe (eds.), Neurophysiol-
ogy, 2nd ed., W.B. Saunders, Philadelphia: 26-53.
Woodbury, J.W., AM. Gordon, and J.T. Conard (1966). Muscle. In T.e. Ruch and
H.D. Patton (eds.), Physiology and Biophysics, W.B. Saunders, Philadelphia:
113-152.
Woodrow, K.W., G.D. Friedman, AP. Siegelaub, and M.F. Collen (1975). Pain
tolerance: Differences according to age, sex, and race. In M. Weisenberg (ed.),
Pain: Clinical and Experimental Perspectives, Mosby, St. Louis: 133-140. (Re-
printed from Psychosomatic Medicine, 1972, vol. 34: 548-556.)
Woodworth, R.S., and H. Schlosberg (1954). Experimental Psychology. Holt,
Rinehart and Winston, New York.
Worthen, E.F. (1982). Surgical treatment of electrical burns of the scalp and skull:
Past and present. Clin. Plast Surg. 9(2): 161-165.
Wu, Y.e. (1979). Electrical injuries-a literature review. Natl. Bur. Std. NBSIR:
79-1710.
Wuerker, R.B., AM. McPhedran, and E. Henneman (1965). Properties of motor
units in a heterogeneous fast muscle (M. Gastrocenmius) of the cat. 1. Neurophys.
28: 85-99.
Wynn, Parry, C.B. (1971). Strength-duration curves. In S.H. Licht (ed.),
Electrodiagnosis and Electromyography, 3rd ed., E. Licht, New Haven, CT.
Wyss, AM. (1963). Die Reizwirkung sinufOrger Wechselstrome, untersucht, bis zur
oberen Grenze der Niederfrequenz (1,OOOHz). Helv. Physiol. Pharmacol. 21:
419-443.
Yamagata, H., S. Kuhara, Y. Seo, K. Sato, O. Hiwaki, and S. Deno (1991). Evalua-
tion of dB/dt thresholds for nerve stimulation elicited by trapezoidal and sinusoi-
dal gradient fields in echo-planar imaging. Society for Magnetic Resonance in
Med., Proc. 10th Annual Meeting, Berkeley, San Francisco: 1277.
Yamaguchi, M., T. Andoh, T. Goto, A Hosono, T. Kawakami, F. Okumura, T.
Takenaka, and I. Yamamoto (1992). Heart stimulation by time-varying magnetic
fields,lpn. 1. Appl. Phys., 31(7): 2310-2313.
Yamaguchi, M., T. Andoh, T. Goto, A. Hosono, T. Kawakami, F. Okumura, T.
Takenaka, and I. Yamamoto (1994). Effects of strong pulsed fields on the cardiac
activity of an open chest dog. IEEE Trans. Biomed. Eng. 41(12): 1188-1191.
Yamamoto, T., and Y. Yamamoto (1977). Analysis for the change of skin imped-
ance. Med. Bioi. Eng. Comput. 15: 219-227.
Yamamoto, T., Y. Yamamoto, and A Yoshida (1986). Formative mechanisms of
current concentration and breakdown phenomena dependent on direct current
flow through the skin by a dry electrode. IEEE Trans. Biomed. Eng. BME-33(4):
396-404.
552 References
Yoshida, K, and K Horch (1993). Selective stimulation of peripheral nerve fibers

using dual intrafascicular electrodes. IEEE Trans. Biomed. Engin. 40: 492--494.
Young, J.S., P.E. Bums, AM. Bowen, and R McCutchen (1982). Spinal Cord Injury
Statistics. Experience of the Regional Spinal Cord Injury Systems, Good Samari-
tan Medical Center, Phoenix, AZ.
Young, J.W., RF. Chandler, C.C. Snow, KM. Robinette, G.F. Zehner, and M.S.
Lofberg (1983). Anthropometric and Mass Distribution Characteristics of the
Adult Female. FAA-AM-83-16 revised ed.
Younossi, Y.K, H.Z. Rtidiger, K Haap, and H. Antoni (1973). Untersuchungen
tiber die Flimmerschwelle dis isolierten Meerchweinchen-Herzens flir Glei-
chstrom und sunusfOrmigen Wechselstrom. Basic Res. Cardiol. 68: 551-568.
Yuen, T.G.H., W.F. Agnew, L.A Bullara, and D.B. McCreery (1990). Bio-
compatibility of electrodes and materials in the central nervous system. In W.F.
Agnew and D.B. McCreery eds., Neural Prostheses: Fundamental Studies,
Zborowski, M. (1952). Cultural components in response to pain. J. Social Issue 8:
16-30.
ZeIt, RG., RK Daniel, P.A Ballard, Y. Brissette, and P. Heroux (1988). High-
voltage electrical injury: Chronic wound evolution. Plast. Reconstr. Surg. 82(6):
1027-1O4l.
Zipes, D.P. (1975). Electrophysiological mechanisms involved in ventricular fibrilla-
tion, Supplement III to Circulation, vols. 51 & 52: 111-120 to 111-130.
Zoll, P.M., RH. Zoll, RH. Flak, J.E. Clinton, D.R Eitel, and E.M. Antman (1985).
External noninvasive temporary cardiac pacing: Clinical trial. Circulation 71(5):
937-944.
Index
A cardiac excitation by, 190

Absolute refractory period. See fibrillation by, 190-191,208-211
Refractory period frequency sensitivity to, cardiac,
Accident statistics, 2-4, 412-413, 446 203-205,210-211,496-497
Accommodation, 140 and heart muscle, influence on, 164
Actin, myofibrillar, 302 let-go phenomenon for, 330-338, 495
Action potential, cardiac perception of, 260-261, 265-269
dc response of, 160 Americal College of Governmental
conduction of, 160-161 and Industrial Hygienists
dc response of, 93 (ACGIH), 463, 466
multiple, 93, 208 American National Standards Institute.
myocardial, 90-94,155,194-195 See ANSI
Purkinje, 91-92 Anisotropic conductivity, 18t, 19,238
refractory period of, 162 Anodal stimulation
sino-atrial, 92 of heart, 197
Action potential, nerve of nerve, 124-126, 133
from biphasic stimulus, 132-145, ANSI
261-262 committees, 455
conduction velocity of, 89-90 organization of, 454
described, 80-81 Safety codes, 494
excitation criteria of, 121-123 Standards Coordinating Committee,
intensity coding by, 95-98 455
of motor neurons, 98, 301, 304-305, and startle effects, 290-291
310-311 Arc. See Spark discharge
from prolonged stimulus, 87 Arcing injury, 57, 362, 415, 421, 443,
propagation of, 74, 88-90 446, 476-478. See also Spark
refractory behavior of, 88-89 discharge
repetitive, 87, 95, 98, 261-263 Area relationship. See Electrode area,
from sinusoidal stimuli, 134-135 effect on
for skeletal muscle, 301-303, 305- Auditory effects, 298, 393-394, 486
306, 310-311 Autorhythmicity, heart, 152, 163
species dependence of, 87-88
Activation function, 119-120,314,316 B
Activation variable, 82, 195 Bent fiber excitation, 130-131, 371
Adams-Stokes syncope, 153 Beta receptor, heart, 170
A fiber, 87 Bioheat equation, 385
Selective stimulation of, 234 Biphasic stimulation
Alternating current stimulation AP reversal by, 122-123
action potential effects of, 135 of heart, 203-205
553
554 Index
Biphasic stimulation (cant.) Capacitance

of muscle, 310 membrane, 19
of nerve, 123, 132-134 polarization, 28
sinusoidal,134-141 skin, 28
strength-duration form for, 135-136 of sweat ducts, 28
See also Frequency sensitive network Capacitor discharge
Block of conduction. See Conduction for magnetic stimulation, 396-400
Body impedance. See Impedance perception threshold of, 10, 250,
Body location sensitivity, 274-277 358-360
Body size effects plateau voltage, 56-58, 350
on fibrillation, 220-226, 495 safety standards for, 499
and magnetic stimulation, 371-373, stimulation by, 248-253, 347-353,
376-379 358-360
on nerve excitation, 288-290, 297 waveform, 249, 349
UL limits for, 495 Cardiac conductivity, 19
Body surface area of burns, 412 Cardiac current density, 41
Bovine. See Cow Cardiac cycle, 194-195
Brain Cardiac excitation
anatomical structures, 400 diastolic, 195-196
conductivity of, 18t pathology influences on, 227
local magnetic stimulation of, 396, threshold of, 235
402-406 See also Fibrillation
magnetic stimulation of, 392, 394- Cardiac latent injuries, 449
396,400,402-409 Cardiac output, 149
neuron diameters, 90 Cardiomyoplasty, 312-313
scales of excitation reaction, 406- Cardiovascular. See entries beginning
408 wi th Cardiac
standards for exposure of, 489, 492, Cardiovascular system, general aspects
498 of,148
synaptic effects, 392 Carpet sparks, 249
visual cortex stimulation, 393 Categorical scaling, 254-256
visual evoked potentials, 393 CENELEC standards, 466
Burns, electrical C fiber, 87
clinical treatment of, 450-453 selective stimulation of, 234
contact, 435-437 Channels, ionic, 86
current-time relationship for, 420, Chronaxie, 109-110,247
498 Circulation, coronary, 151
deep tissue, 438-442 Clinical treatment of burns, 450-453
limits for, 497-498 Codes
oral,446 National Electrical Code, 494
skin, 445 National Electrical Safety Code, 493
spark discharge, from, 57, 362, 477- See also Standards
478 Conduction
superficial, 435-438 disturbance of, in heart, 161
thermal, 8, 412-415, 435-442 heart muscle, 160
slowing of, as condition for reentry,
C 185
Cable, electrical See also Action potential, cardiac
equations of, 112-115, 119 Conduction block
and myelinated nerve model, 118- cardiac, atrio-ventricular, 153
119 cardiac, sino-atrial, 153
and space constant, 113-114 unidirectional, as condition for
Cadaver impedance, 41-42 reentry, 186-187
Calcium (+ +) oscillations, 468 See also Action potential, cardiac
Index 555
Conduction failure in nerve, 310--311 Diabetic neuropathy. See Neuropathy

Conductivity Diastolic excitation, 197
anisotropic, 19 Diastolic period, 195
of biomaterials, 17-19 Dielectric, complex, 16
cardiac tissue, 19 Dielectric breakdown, 30
of cellular membranes, 19 Dielectric dispersion, 17
complex, 16 Dipoles, 14
corneum, 21-23 Dromotropic effects on heart, 167-168
defined, 14 Drugs
of nerve bundles, 19 antiarrhythmic, 161
thermal,414-415 effects of, on heart muscle, 170--171
See also Impedance; Resistivity Duty factor, in electromagnetic
Consumer Product Safety Commission standards, 472
(CPSC),494 Dynamic myoplasty and
Corneum cardiomyoplasty, 312, 313
conductivity of, 21-23
damage of, 57, 362, 476-478 E
microampere detection by, 240 Eddy currents, 363, 373
stripping of, 22, 24, 38, 58 Efferent neurons, 74, 97
Corona discharges, 465 Einthoven's triangle, 176
Cows Electrophonics. See Auditory
electrical thresholds of, 291-296 stimulation
impedance of, 68-72 Electric fence, hazard of, 339-340, 495
magnetic stimulation of, 296 Electric field (ambient)
Current density enhancement of, by body, 346-347
for cardiac excitation, 230--236 perception of, 353-358
for dielectric breakdown, 30 spark discharge stimulation in, 348-
distribution of, in body, 37, 414-415, 353
430--435 Electric field (in situ)
beneath electrode, 26, 37-38, 426- excitation by, 129-132
428 multiple pulse stimulation by, 263-
in heart, 41 265
for nerve excitation, 129-131,274, Electric field induction
414 equivalent area of, 342, 346
for thermal effects, 413-415, 421, in humans, 344-347
430--435,441 principles of, 341-344
Current paths of skeletal muscle Electrocardiogram
stimulation, 325, 330--334, 340 definition of, 171
Cyclotron resonance. See Ion diagnostic use of, 177-178
resonance and Einthoven's triangle, 174
form and terminology of, 171-173
D origin of, 173-174
DC stimulation QRS axis orientation of, 175-176
of cardiac tissue, 166, 204, 209 recorders of, 171
of fibrillation, 210, 215, 229-230, 238 recording lead systems of, 174-177
perception by, 265 R-on-T phenomenon in, 187
deactivation variable, 82, 194 Electrochemistry, 310, 322-323
See also Excitation threshold; Electroconvulsive therapy, 407f, 408
Fibrillation Electrocution equation, 217-218
Defibrillation, 8, 192-193, 198 Electrocutions, 2, 412
electrical mechanism, 192-193 Electrode
and first aid in fibrillation, 192-193 current distribution from, 26-27, 37-
Delayed injury. See Latent injury 38,426-428
Density. See Current density epimysial,318
556 Index
Electrode (cont.) polarity effects on, 124-126

impedance, 25, 426 for repetitive stimuli, 141-143
intramuscular, 307, 309, 318 for single cardiac cells, 159
intraneural, 318-319 for sinusoidal stimuli, 134--140
nerve cuff, 318-319 with uniform field excitation, 129-
neural stimulation, 318-319 120,334,370-371
Electrode area, effect on See also Fibrillation; Perception,
body impedance, 31-36 electrical
capacitor discharge perception, 351- Excitation wave
352 length of, in heart muscle, 183
cardiac S-D parameters, 198 and reentry relationship in heart, 183
fibrillation threshold, 230-236 See also Action potential, nerve
pain threshold, 273 Exponential stimulus, 126--127
perception threshold, 270-272 Exponential strength-duration formula,
skin impedance, 26 108, 127-128
Electromagnatic exposure limits, 458- Exponential time constant, 110
468 Extrasystole
Electropathology. See Neuropathy; and cardiac rhythm influence, 180
Pathology, skeletal muscle definition of, 180
stimulation repetitive, effects of, on fibrillation,
Electroporation, 9, 416--417, 442, 453 192, 208-214
Electrostatic force, perception by, 241-
242,357 F
Electrosurgery, 140 Faraday's law, 363
Electrotonic distance. See Space Fatality, statistics on, 2-4
constant membrane and lightning, 2
Electroventilation, 340 Fatigue, muscle, 303-307, 310-311,
Endogeneous fields, 479-481 321-322
End plates, 98 Fiber diameter, 90
End structure, stimulation of, 130-131 Fibrillation
Energy, electrical atrial, 180
for fibrillation, 217-219 first aid for, 192-193
strength-duration form of, 107-109 mechanism for, 182-186
for thermal effects, 413-414, 437 ventricular, 182
Ephaptic transmission, 100 Fibrillation, electrically induced
Epidermis thickness, 21 by alternating current, 208-217
Epimysial stimulation electrodes, 318 body size effects on, 220-223, 418
Evoked response, 439, 442 by capacitor discharge, 219-220, 499
Excitation threshold current density criteria for, 230-236
and anodal break, cardiac, 160 by direct current, 204, 209
for biphasic stimuli, 122, 132-141 duration of current effects on, 202-
cardiac, 151-152 203,206-208,212-217,496
and cathodal make, cardiac, 160 electrode area effects on, 212
and drugs, effects of, on heart, 170- energy criteria for, 217-220
171 and excitation, contrast to, 196-197,
and electric field distribution, 129- 208-211
131,319-321 extrasystole effects on, 208-211
and electric field (in tissue) and frequency sensitivity, 203-208,
excitation, 129-132,334 211,497
in heart muscle, 189-190 ground-fault interrupter protection
for let-go phenomenon, 334--338 for, 500-501
magnetic, cardiac, 370-382 by high current, 417-419
magnetic, nerve, 373-375, 379-382 inductance effect on, 219-220
myelinated nerve for, 122-123 limits on, 495
Index 557
open heart, 223-224, 227 Heart

for pathological heart, 188-189,227 atrio-ventricular (AV) node in, 153
by prolonged currents, 202-203 coronary vessels, in, 149-150
statistical distribution of, 225-227 current density in, 41, 230-232
strength-duration parameters for, energetics of work of, 149
198-202 origin and spread of excitation in,
threshold of, 187-192,212-217,230- 181-183
236 pacemaker and conducting system
Figure-of-eight coil, 40~05 in, 153
Flutter, atrial and ventricular, 181- Purkinje fibers in, 153
182 sino-atrial (SA) node in, 152
Flux density, 363 vagal and sympathetic tone in, 169
Foot contact impedance, 48-51 See also entries beginning with
Frankenhaeuser-Huxley equations, 86- Cardiac
87 Heart muscle
Free nerve endings, 94-95 action potential of, 154-156
Frequency dependence, electrical all-or-nothing response of, 152
stimulation and excitation, ionic mechanism of,
in electromagnetic standards, 456, 154-156
458,490 excitation-contraction coupling of,
of heart, 203-206, 211 158-159
high frequency, 140-141,269 internal longitudinal resistance of,
for let-go, 138, 329-330, 497 161
with magnetic stimulation, 375-376 mechanism of contraction of, 156-
multiple frequencies, 143, 387, 393 158
of muscle, 311-312, 329 resting potential of, 154
of nerve, 134-141 stimulation of, 153
for perception, 265-269 Heat. See entries beginning with
for phospheres, 387-389, 393 Thermal
for thermal effects, 475 Heat equation, 414
Functional electrical stimulation, 312- Hodgkin-Huxley equations, 81-86
323
I
G IC NIRP, 462, 465
Galvani, 1, 299 IEC
Gate. See Channels, ionic body size criteria of, 215-217, 223
Gate theory, 282 capacitor discharge criteria of, 251
Gender relationships, 267, 272f, 276, fibrillation criteria of, 215-217, 494-
287-290,354 497
Generator potential, 95 impedance model of, 42
Golgi tendon organs, 300 MRI criteria, 468-471
Ground-fault circuit interrupter, 500- ventricular fibrillation criteria, 496
501 Z-relationship of, 215-217
Grounding, 500 Impedance
body, 25, 31,65
H and body size, 44-45, 48
Hair receptor, vibration, 94 of cadavers, 36, 41-42
electrical detection by, 353-357 change in, after death, 41
Hall effect, 409-410 of children, 43-45
Hand and wrist, functional anatomy of, circuit models of, 23-25
330-334 contact, 48-51, 426
Hand contact accident, 330-338 with corneal stripping, 24, 58
Hand contact stimulation, 5-7, 323- of cows, 68-71
338, 460 direct current, 42-45
558 Index
Impedance (cant.) vessel,445

distribution of, in body, 37-41, 65- Innervation, heart
68 autonomic, mode of action of, 169-
electrode, 25, 426 170
epidermis, 23 autonomic effects of, 166-170
foot contact, 48-51 chronotropic, effects of, 167-168
free space, 457 Inotropic effects on heart, 166-168
at high frequency, 45-48 Interfacial effects, 17
high-voltage, 52-54, 57-61 International Electrotechnical
of pigs, 72 Commission. See IEC
polarity effects on, 61 Intrafusal fibers, 300
skin, 28-31 Intramuscular electrodes, 307, 309, 318
spark discharge, 54-60 Intraneural electrodes, 318-319
spreading, 24-25 Ionic concentrations, 76-77
statistical distribution of, 42-45 influence of, on heart muscle, 170-
thermal effects on, 41-42, 413, 451- 171
452 Ion resonance, 411, 488
See also Conductivity; Resistance IRPA,465
Impulse currents
cow perception of, 294-296 J
energy criteria for, 107-109,217- Johnson-Nyquist noise, 479
219
strength-duration criteria for, 107- K
111 Keratin, 21
See also Capacitor discharge
Inability to let go. See Let-go L
phenomenon Latent injury, 445-446, 448-449
Inductance effects on fibrillation, 219- Leakage current
220 in consumer products, 495
Infarction related to ECG, 177 Leakage resistance, effects of, on E-
Injury (trauma) field shocks, 350, 359-360
arcing, 415, 421, 443, 446 Let-go phenomenon
clinical observations on, 444-450 analysis of, 330-338
electric burn, 413-414, 435-437 frequency effects of, 138, 329-330
contact, 424, 426-429, 436 leakage-current-measurement for,
deep tissue, 429-434 495
oral,446 and respiration interference, 338-340
on skin, 445 safety standards for, 495-497
superficial, 426-429 thresholds of, 7,138,324-330
high-current, 415, 418 Lightning injury, 2, 443-444
high-voltage, 415, 418 Liminal length, 233
of joint capsules, 441 Limits, current, 460, 495
latent, 445-446, 448-449 Lipid bilayer, 75
lesion from, 442, 445 Long wavelength solution, 363-366
from lightning, 443-444 Lorentz force, 410
model of, 415, 426-430
muscle, 438-439, 445 M
nerve, 439-440 Magnetohydrodynamic effect, 409-410,
nonthermal, 442-443 486
renal failure, 446 Magnetic field
reports of, 412 flux density of, 363
secondary, 445 induction principles of, 362-368
thermal,413-415 power deposition by, 367
tissue necrosis, 445 standards, 454ft, 473-474
Index 559
Magnetic stimulation, cardiac Metabolism, human, 456

beat rate effects, 382 Microwave hearing, 394-395, 486
excitation thresholds, 370-382 Motor line, 314-315
rheobase, 382 Motor neurons, 74
strength-duration time constant, 382 Motor point, 314-318
Magnetic resonance imaging Motor unit, 300, 303-307
FDA guidelines, 469-470 Multiple discharges in AC fields, 360-
IEC guidelines, 469-471 361
painful reactions in, 470 Multiple waveform stimulation, 143
pulsed field limits, 469-472 Muscle, skeletal
static field limits, 469t action potential of, 301-303, 305-
waveforms, 369-370 306, 310-311
Magnetic stimulation, nerve contraction of, 97-99, 302-303
and body size, 376-378 electrical stimulation of, 299-340
excitation threshold, 373-375, 379- end plates of, 98, 301
382 extrafusal fibers of, 300
painful,257-258,385-386,396,474 fatigue of, 303-307, 310-311, 321-
perception of 379-382, 385 322
phrenic nerve, 379-381 fiber types in, 303-307,321
pulsed stimulation, 374-375 force production in, 302-303, 306-
small coil systems for, 396-406 307,321-322
by sinusoidal fields, 375-376 Golgi tendon organs in, 300
strength-duration time constant of, intrafusal fibers of, 300
247,384-385,489-490 strength-duration curve of, 244-245,
Magnetic stimulation of head 308-309
auditory effects, 394-396 stretch reflex of, 300
figure-of-eight coil, 404-405 structure and function of, 74, 97-99,
focal stimulation, 402-406 299-307
headaches from, 408 temporal summation in, 303
and phosphenes, 387-388, 392, 489- tetanus, 7, 98, 138,311-312
491 thermal injury to, 438-439, 445
by small coils, 396-406 twitch, 7, 98
scales of reaction to, 406-409 See also Let-go phenomenon
taste sensations from, 410, 486 Myelin, 75
vertigo from, 410, 486 Myelinated nerve
visual evoked potential effects, 408 and a-motor neurons, 299-301, 307-
Magnitude scaling, 253-254 308
Mechanoreceptors, 94 characteristics of, 87t, 90
stimulation of, 241, 355 classes of, 90
Meisner corpuscle, 94, 97 electrical models for, 117-121
Membrane electrical stimulation of, 118-121,
capacitance of, 19 307-323,334-340
conductivity of, 416 Myosin, myofibrillar, 302, 306
electric field, 79
Frankenhaeuser-Huxley, 86-87 N
Hodgkin-Huxley, 81-86 National Council on Radiation
noise bandwidth, 479 Protection, 458, 461, 464
permeability of, 76, 416 National Electrical Code, 494
permittivity of, 19 Nernst equation, 78
potential, 113-115,416 Nernst potential, 77t
thermal noise voltage, 479-480 Nerve, myelinated. See Myelinated
space constant of, 113-115 nerve
thickness of, 75 Nerve cuff electrode, 318-319
time constant of, 106-107, 308 Neural stimulation electrode, 318-319
560 Index
Neurologic injury, 438-440 of high frequencies, 460

Neuromuscular junction, 301, 311 intensity coding of, 95-97
Neuropathy, 269, 275, 283. See also of magnetic fields, 373-382, 386
Pathology magnitude scaling of, 253-254
Neurotransmitters, 75, 97, 100 of microampere currents, 240
Nociceptors painful, 7, 254---256, 386
intensity coding by, 97 of phosphenes, 298, 387-393
role of, 95 quality of, 243
stimulation of, 240-241, 256 of repetitive stimuli, 261-265, 360-
Noise bandwidth, membrane, 479 361
Nonlinear mechanisms, 143,387,393, and safety standards, 495
487,488 of sinusoidal stimuli, 265-269
Nuclear magnetic resonance, 488 by thermal means, 8, 241, 266-267
see also Magnetic resonance imaging tolerance of, 7, 260, 386
touch contact, 460
o See also Pain
C>pen-heart surgery, 219, 224, 227 Perfusion (bloodflow), 414, 437
Permeability
p of biomaterials, 20
Pacemaker, cardiac ionic, 78-79
activity modification of, 165 membrane, 76
actual, 165 Permittivity
artificial, 153, 159-160 of air, 14
ectopic, 163-164 of biomaterials, 16-19
natural, 164---166 of cell membrane, 19
potential of, 165 change of, after death, 41
Pacinian corpuscles, 75, 94---95 complex, 16
Pain, electrical Phosphenes
AC field effects on, 360-361 charge threshold, 390
and dielectric breakdown, 30 electrically induced, 10, 298, 388-389
dynamic range of, 258-259, 360-361 exposure limits for, 489-491
electrode area effects on, 273 frequency effects, 387-389, 393
gender relationship to, 287 magnetically induced, 10, 392, 406-
by magnetic stimulation, 257-258, 408
385-386,470-471 mechanisms for, 390
and nociceptors, 97 photoreceptor stimulation of, 392
perception of, 243 strength-duration relation for, 391-
repetitive stimulus effects on, 261- 392
265,360-361 and synaptic polarization, 103, 390
strength-duration curve for, 245/ Phrenic nerve stimulation, 340, 379
See also Perception, electrical Plasmon resonance, 487
Parabolic strength-duration form, 109 Plateau voltage, 56-58, 350
Pathology Polarity effects
cardiac excitation 188-189, 227 on cardiac stimulation, 197,227-230
nerve excitation, 275, 283 on excitation thresholds, 123-127,
skeletal muscle excitation, 312, 322- 146-147
323 on perception, 269-270
Perception, electrical on spark discharges, 61-62
of biphasic stimuli, 261-262 Polarization capacitance, 28
by cows, 291-297 Power functions, 253-255, 259-260,
dynamic range of, 258-260, 361, 386 265
of electric fields, 353-357 Propagation. See Action potential
and electrode impedance, effects on, Protective systems and devices, 500-
273 501
Index 561
Psychometric function, 243 Resistivity

Pump, sodium, 79 corneum, 22-23
Purkinje fibers, 152-153 defined, 13
action potential of, 91-93 longitudinal, of heart, 161
DC stimulation of, 210 skin, 26
liminal length of, 233 soil, 51
membrane model for, 91-92 thermal effects on, 413-415, 430,
Pyramidal tract neurons, 90 438-441
See also Conductivity
Q Respiration, artificial, 340
QIO coefficient, 280 Respiratory stimulation, 7, 323-324,
338-340
R interference by, 7
Radical pain mechanism, 488 magnetic stimulation, 379-381
Radon decay effects, 487 Resting potential
Rate modulation in electrical of cardiac cells, 154
stimulation, 303, 314, 321-322 of nerve, 77t, 78-79, 115, 387
Reaction. See Startle of skeletal muscle cell, 77t
Receptor Rheobase
classes of, 94 in cardiac stimulation, 198
rapidly adapting, 355 defined, 109
recruitment, 256 in myelinated nerve model, 127
sensory, 73-74, 94-97 in nerve stimulation, 244, 307
Recruitment Rhythm, cardiac
in nerve stimulation, 243, 256 disturbances of, 178-182
physiological, 303-305, 314 origin of, in A V mode, 178
in skeletal muscle stimulation, 314, origin of, in ventricle, 180
319-322,335 Rufini ending, 94
Reentry, 196, 224
conditions for, 183 S
Reflex action, 75, 103-104 Seizures, 407/, 408
and startle, 290-291 Self-injurious behavior inhibition, 5
Refractory period SENN model, 121, 247, 385, 405
cardiac, 161-162, 195 Sensation. See Perception, electrical
and conditions for reentry, 183-185 Shoe resistance, 51-52
nerve, 88-89 Short-circuit current, 343-347
Relaxation time constant, 17 Shot noise, 479
Repetitive stimulation Sino-atrial node, 152
of muscle, 310-311 Sinusoidal excitation, 134-141, 265-
of nerve, 137, 141-143,256-265 269,311, 329, 375-377. See also
painful, 262-263 Frequency dependence
perception of, 256-257, 360-361 Skeletal muscle. See Muscle, skeletal
sensory integration of, 261-262 Skeletal muscle ventricles, 313
supra threshold growth of, 264-265 Skin
Resonance, electromagnetic in capacitance of, 28
humans, 459 conductivity of, 21-23
Retina damage, 57, 362,476-478
neural structure, 102-103 dielectric breakdown of, 30
stimulation of, 391 impedance of, 20-21, 23-25, 30, 273
strength-duration effects, 391-392 potential, 24
See also Phosphenes structure of, 21-22
Residual current device, 500. See also temperature of, 277-280
Ground-fault circuit interrupter thickness of, 21
Resistance. See Impedance See also Corneum
562 Index
Skin depth, 20 for biphasic stimuli, 132-134

Sodium pump, 78-79 for charge, 107
Soliton mechanisms, 486 for energy, 107
Somatosensory system, 94 for exponential stimulus, 107-108
Space constant, membrane, 113-114 Lapique, 109
Spark discharge linear model of, 107
breakdown process of, 54 with magnetic excitation, 373-375
in electromagnetic field standards, myelinated nerve model of, 111
459,465,476-478 with neuropathy, 269, 283
impedance of, 58-59 for perception, 244
plateau voltage of, 56 for sinusoidal stimulus, 133, 138-
polarity effects of, 61-62 140
and skin burns, 57-58, 362, 476-478 Strength-duration time constant
time constant of, 63-64 with bent fiber, 131
See also Capacitor discharge and chronaxie, 110
Specific absorption rate (SAR) for cows, 295
brain heating, 408 and current distribution, 110
defined, 478 and electrode distance, 128, 145-146
microwave hearing, 395 experimental, 127,247
resonance, human body, 455 with magnetic excitation, 371, 374,
safely factor for, 474 382,384-385
tissue heating, 485 for monophasic stimulus, 127-128
Spinal cord injury, skeletal muscle of muscle, 247, 308-309
stimulation in, 312-313, 321 parameter sensitivity of, 145-147
Spinal reflex. See Reflex action and regulated current/voltage, 111
Standards, 454 retina stimulation, 391-392
development of, 455 with uniform field excitation, 130-
international, 461 132
magnetic resonance imaging, 469- Strength-frequency forms
472 formula for, 139-140
safety, 494-499 for muscle stimuli, 128, 311-312,
See also Codes 329-330
Startle, 290-291 See also Frequeney dependence,
in UL standards, 495 electrical stimulation
Static fields Stretch receptor, 94
electromagnetic limits for, 461, 468 Stretch reflex, 300
MRI limits, 469 Suffocation. See Respiratory
taste sensations in, 461 interference
vertigo in, 461 Supernormal period, cardiac, 162-163
Stimulation. See Excitation threshold Supra threshold multiples, 257-258,
Stochastic resonance, 487 360,386
Stray voltage, 68, 291 Surface electrodes for skeletal muscle
Strength-duration form, cardiac stimulation, 310, 314-318
characteristics of, 198-199 Sweat duct
and electrode area, effects on, 199- capacitance of, 28
202 conductivity of, 23
with magnetic field excitation, 371, density of, 23
374 impedance effects by, 30, 60
with oscillatory stimuli, 206 Sympathetic tone of heart, 169
polarity effects on, 202 Synapses
and prolonged current, effects on, electrical stimulation of, 10, 101-103
202-203 electromagnetic interactions in, 484
and tissue size, effects on, 202 neural, 473
Strength-duration form, nerve and phosphenes, 103, 390
Index 563
orientation effects on stimulation, Tone, vagal and sympathetic, 169

115 Transients. See Impulse currents
polarization effects on, 101, 390 Trauma. See Injury (trauma)
of retina neurons, 103 Twitch, 7, 98
T U
Tactile discs, 94 Underwriters Laboratories, Inc. (UL)
Tactile masking, 280-282 body current limit of, 495
Taste sensations, 410, 486 and ground-fault circuit interrupter,
Temperature effects on perception, 501
277-280 leakage-current limits of, 495
Temporal summation, 261-265, 303 let-go criteria, 495, 497
Terminated fibers, 124-126, 129-132 ventricular fibrillation limit, 495-
Tetanus, 98, 311-312, 433, 497 497
See also Let-go phenomenon
Thermal conductivity, 413-415 V
Thermal noise limits, 479-480 Vagal tone of heart, 169
Thermal, impedance effects of, 41-42, Ventricular fibrillation. See Fibrillation
430,438 Vertigo effects, 410, 486
Thermal perception, 8, 241, 267, 365, Vibratory stimulus, 95-97
394,460,475 Video display terminal emissions, 467
Thermal tissue heating, 428, 435 Visual effects. See Phosphenes
Threshold of excitation. See Excitation Visual evoked potential, 393
threshold Vulnerable period of heart, 187-188
Time constant
and current distribution, 128 z
experimental, 128 Z relationship
membrane, 106 electrode area effects on, 212
spark discharge, 63-64 experimental, 212
strength-duration, 109-110 formula for, 215-216
Tolerance, 259-260 lEe criteria for, 216-217, 496

Applied Bioelectricity - Rizadian

Uploaded by

Copyright:

Available Formats

Applied Bioelectricity - Rizadian

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Applied Bioelectricity - Rizadian

Uploaded by

Copyright:

Available Formats

Springer Science+Business Media, LLC

With 241 Figures

Cover art by Anthony E. Randolph, The Johns Hopkins

Library of Congress Cataloging-in-Publication Data

Printed on acid-free paper.

© 1998 Springer Science+Business Media New York

ISBN 978-1-4612-7235-9 SPIN 10657964

The use of electrical devices is pervasive in modern society. The same

Electrical forces may be introduced into the biological medium through

(formally the Fleet Systems Department) of the Applied Physics Labora-

Laurel, Maryland J. PATRICK REILLY

2 Impedance and Current Distribution ............... 12

Distribution of Current and Impedance Within

3 Electrical Principles of Nerve and

4 Excitation Models .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 105

Nonlinear Models 116

5 Electrical Properties of the Heart .................. 148

5.7 Electrocardiogram. . . . . . . . . .. . . .. . . . . . . . . . . . . . . . .. 171

6 Cardiac Sensitivity to Electrical Stimulation . . . . . .. 194

6.6 Energy Criteria and Impulse Currents ............... 217

7 Sensory Responses to Electrical Stimulation ....... 240

8 Skeletal Muscle Response to

Anatomy and Physiology of Respiration ........... 338

9 Stimulation via Electric and Magnetic Fields ....... 341

Duration and Frequency Relationships for

10 High-Voltage and High-Current Injuries ........... 412

11 Standards and Rationale . . . . . . .. . . . .. . . . . . . . . . .. ... 454

Rationale for IEEE/ANSI C95.1 Standards ........ 455

References ................................................. 502

1.1 General Perspective

J. P. Reilly, Applied Bioelectricity 1

current stimulation, and "Faradaic" to indicate pulsed or interrupted

1.2 Electrical Exposure

TABLE 1.1. Mortality because of electrical accidents.

It is likely that the mortality data referred to above underreport the

Typical Electrical Exposures

TABLE 1.2. Examples of electrical stimulation in biomedical applications.

undesirable in one medical application may be considered beneficial in

1.3 Scales of Short-Term Reactions to Contact Current

Muscle Cardiac Thermal Electroporation

100 Excitation T = 45° C

each involving a distinct continuum of reactions. Various thresholds of

be found. Additional chapters (2-5) provide a foundation of bioelectric

Sensory Reactions (Chapter 7)

Muscle Reactions (Chapter 8)

Cardiac Reactions (Chapters 5 and 6)

heart. Such excitation is not necessarily life threatening, but is nevertheless

Thermal Reactions (Chapter 10)

Electroporation (Chapter 10)

1.4 Reactions to Electric and Magnetic

1.5 Variables Affecting Thresholds

2.1 Dielectric Properties of Biological Materials

Conductivity and Permittivity

Inversion of Eq. (2.1) yields a simple definition of the material's resistivity

(b) Area proportional resistance of membrane

between the two units of measurements is e(Qcm) = lOOe(Qm). The in-

For a sinusoidal voltage, dVldt = jwV, where w = 2nf,fis the frequency of

The macroscopic electric field is E = Vld, and the current density is J =