Journal of Applied Physiology
Vol. 83, No. 1,
pp. 317-322,
July 1997
CONTROL OF BREATHING, CIRCULATION, AND TEMPERATURE
SPECIAL COMMUNICATION
Spiral nerve cuff electrode for recordings of respiratory output
Mesut
Sahin1,
Musa A.
Haxhiu2,
Dominique M.
Durand1, and
Ismail A.
Dreshaj2
Departments of 1 Biomedical
Engineering and
2 Medicine, Case Western
Reserve University, Cleveland, Ohio 44106
ABSTRACT
- INTRODUCTION
- MATERIALS AND METHODS
- RESULTS
- DISCUSSION
- ACKNOWLEDGEMENTS
- FOOTNOTES
- REFERENCES
ABSTRACT 
Sahin, Mesut, Musa A. Haxhiu, Dominique M. Durand, and
Ismail A. Dreshaj. Spiral nerve cuff electrode for recordings of
respiratory output. J. Appl. Physiol.
83(1): 317-322, 1997.
The feasibility of using the spiral nerve
cuff electrode design for recordings of respiratory output from the
hypoglossal (HG) and phrenic nerves is demonstrated in anesthetized,
paralyzed, and artificially ventilated cats. Raw neural discharges of
the HG nerve were analyzed in terms of signal-to-noise ratios and
frequency spectra. The rectified and integrated moving average activity of the HG nerve had a peak value of 1.74 ± 0.21 µV and a baseline value of 0.72 ± 0.11 µV at elevated respiratory drive induced by
increases in CO2 or oxygen
deprivation when recorded with 10-mm-long cuffs. The frequency content
of the HG electroneurogram extended from several hundred hertz to 6 kHz. Spiral nerve cuff recordings without desheathing of the nerve
provided large enough signal-to-noise ratios that allowed them to be
used as a measure of respiratory output and had much wider frequency
bandwidths than the hook electrode preparations. A major advantage of
the cuff electrode over the hook electrode was its mechanical
stability, which significantly improved the reproducibility of the
recordings both in terms of signal amplitudes and frequency contents.
nerve recording; phrenic nerve; hypoglossal nerve; hook electrode; power spectrum analysis
INTRODUCTION
GENERALLY, THE RESPIRATORY DRIVE in anesthetized
animals is measured by using electroneurograms (ENGs) that are obtained
with standard hook electrodes. However, cutting and desheathing the nerve, which are required to obtain useful signal amplitudes, may
compromise nerve integrity and nerve output. An alternative recording
method is a nerve cuff electrode, which can provide large signals
without cutting or desheathing of the nerve (4, 5, 12-18). Nerve
cuff electrodes also increase the mechanical stability of the
nerve-electrode interface, thus improving the reproducibility of the
recordings.
The spiral nerve cuff electrode design, which was developed for
electrical nerve stimulation, wraps around the nerve and, because of
its self-coiling property, adjusts its diameter to the size of the
nerve (12). The purpose of this study was to examine the feasibility of
using the spiral nerve cuff electrode for recordings of respiratory
output in anesthetized, paralyzed, and artificially ventilated cats.
Our experience with hook electrode recordings indicates that
hypoglossal (HG) nerve discharges are relatively smaller than those of
the phrenic (Phr) nerve in amplitude. Thus HG nerve discharges are more
difficult to record with good signal-to-noise ratios. Therefore, the
focus of this study was to record and analyze the HG nerve activity by
using the spiral nerve cuff electrode. We have analyzed the cuff HG
nerve recordings in terms of signal-to-noise ratios and frequency
spectrum characteristics, and the frequency spectra were compared with
data obtained from the hook electrodes.
MATERIALS AND METHODS
Spiral nerve cuff electrode fabrication.
The details of spiral nerve cuff fabrication have been described
earlier for electrical stimulation applications (12). The spiral nerve
cuff electrode consists of two layers of Silastic sheet (Dow Corning)
bound together and platinum-foil bands placed between the layers to
provide electrical contact with the nerve (Fig.
1A). Platinum-foil
bands (in this design, thickness = 5 µm and width = 1.5 mm) were
welded to Teflon-insulated multistrand stainless steel (316 LVM) lead
wires for electrical connection. The bands were aligned on top of an
unstretched sheet of Silastic and lightly secured by using Silastic
adhesive (Medical Adhesive Silicon Type A, Dow Corning). Curing agent
(MDX4-4210, Dow Corning) was spread over the sheet and the
platinum bands. Another layer of Silastic sheet was stretched and
placed on the top. The sandwich was cured between heated pressure
plates. The cuff coils in the direction parallel with the platinum
bands and takes a cylindrical shape after it is cut out from its
surrounding (Fig. 1B). Windows (1 mm
wide) were cut inside the cylinder over the platinum bands to make
electrical contacts with the nerve for recordings. The final cuff
diameters were ~1 and 0.5 mm, which are the sizes of the HG and Phr
nerves in the cat, respectively. In this design, three platinum-foil
bands located 5 mm apart from each other (total cuff length = 10 mm)
were used. These bands were connected to the preamplifier with the lead
wires configured in either tripolar or bipolar modes (Fig.
2).
Fig. 1.
A: spiral nerve cuff electrode design.
Top layer is stretched before sandwich is put between
pressure pads for binding. B: after
binding, cuff curls in direction shown by arrow due to stress stored in
top layer.
[View Larger Version of this Image (24K GIF file)]
Fig. 2.
Tripolar (A) and bipolar
(B) recording configurations that
are used for electrical connection of metal contacts to amplifier.
[View Larger Version of this Image (21K GIF file)]
Experimental preparation.
In three anesthetized, paralyzed, and artificially ventilated animals
with intact carotid sinus nerves, HG and Phr nerve ENGs were recorded
with spiral nerve cuff and conventional hook electrodes. Cats were
anesthetized with 
-chloralose (50 mg/kg ip). Supplemental doses of
anesthesia were given every 60 min (10 mg/kg of -chloralose). A low
tracheostomy was performed, and a tracheal tube was inserted. Cats were
paralyzed with gallamine triethiodide (Flaxedil, 4 mg/kg iv) and
mechanically ventilated. Animals were allowed to periodically recover
from paralysis. If a withdrawal response to nociceptive stimuli was
present, an additional dose of -chloralose was given before
paralysis. For the recordings of the HG nerve with the cuff electrode, ~ 2 cm of the nerve length were carefully dissected. The cuff
electrode was placed around the main trunk of the HG nerve proximal to
the bifurcation point of the branches to the muscles of the tongue. For
the hook electrode recordings, the standard preparation was used as
described earlier (6, 7, 9). Briefly, the nerve (HG or Phr) was cut,
desheathed, and covered with Vaseline. The hook electrodes had
arbitrary lead separations (4-8 mm). The recordings were made
starting from implantation of the nerve cuff electrode for a period of
8-10 h.
Experimental protocol.
The recordings were obtained at different levels of respiratory drive.
The respiratory drive was increased by incremental increases in
arterial CO2 by rebreathing method
(3% CO2 in
O2) and normocapnic hypoxia (8%
O2 in
N2). A decrease in the
respiratory drive was induced by reducing the end-tidal
CO2 or switching the animal from
hypoxic to hyperoxic gas mixture.
Data acquisition.
Recording hardware consists of a custom-design preamplifier (head
stage), a commercial laboratory amplifier with a band-pass filter
(model 113, EG&G PARC), data-acquisition board/software (NB-MIO-16P-5/Labview, National Instruments), and a personal computer. A low-noise integrated instrumentation amplifier (AMP-01, Monolithic Precision) is used in the design of the preamplifier. The experimental noise is primarily due to 1) the
resistivity of the tissue (Johnson noise) and
2) the preamplifier. The
contamination from the power lines is reduced by using an optical
isolation stage (4N27, Texas Instruments) between the preamplifier and
the amplifier, and the remaining contamination is removed by the
low-frequency stop band of the band-pass filter (300 Hz) available on
the commercial amplifier. The raw data are recorded breath by breath
for study of the frequency characteristics of the neural activity. Data
are sampled at 30,000 samples/s after band-pass filtering (300 Hz-10
kHz) to prevent aliasing. For integrated moving average ENGs, the
nerve's electrical activity is first rectified and integrated by using
a 200-ms electronic averager, in addition to filtering (300 Hz-3 kHz),
and then digitized at 40 samples/s. The electrode configurations
utilized are shown in Fig. 2.
Data analysis.
The signal-to-noise ratio is defined as the peak value of the
integrated moving average signal divided by the baseline value measured
immediately before the start of the inspiratory phase (see Fig.
4B). The peak and baseline values
are measured at the maximal respiratory level in each animal.
Signal-to-noise ratios are calculated in each animal separately. Then,
the mean and SD of each parameter are calculated across all the
animals. The electronic noise due to the preamplifier is estimated by
making a recording with the input of the amplifier short circuited.
Power spectra are calculated by using the Welch method of power
spectrum estimation. The sampled version of the time sequences
consisting of samples (n) are
divided into overlapping (50%) sections (K) of M points, where M is a
power of two. Successive sections are filtered with a Hanning window,
transformed by using fast Fourier transform, and averaged.
Fig. 4.
A: typical raw ENG recorded from HG
nerve by using spiral nerve cuff electrode. Signal is filtered by using
a 300-Hz to 10-kHz band-pass filter and sampled at 30 kHz.
B: rectified and integrated moving
average (200-ms) ENG obtained from data shown in
A. Locations of peak and baseline
activity level readings are also shown.
[View Larger Versions of these Images (23 + 14K GIF file)]
RESULTS
Integrated ENGs.
An example of integrated nerve activity recorded during change in
respiratory drive in a cat with intact carotid sinus nerves is shown in
Fig. 3. The animal was first ventilated
with O2 at end-tidal
CO2 above apneic threshold
(PO2 = 359 Torr, PCO2 = 40.7 Torr, pH = 7.393).
Switching to hypoxic gas mixture (8%
O2-balance
N2) was associated with an
increase in Phr nerve activities recorded by both hook and nerve cuff
electrodes as well as an increase in HG nerve discharge. When the
animal was switched from a hypoxic to a hyperoxic gas mixture
(indicated by the arrow in Fig. 3), the activities of both nerves were
decreased and then gradually returned to their prehypoxic exposure
levels.
Fig. 3.
Rectified and integrated moving average (200-ms) electroneurograms
(ENGs) recorded from hypoglossal (HG) and phrenic (Phr) nerves with
both spiral nerve cuff and hook electrodes when animal is switched from
hypoxic to hyperoxic gas mixture.
[View Larger Version of this Image (47K GIF file)]
Raw ENGs.
A typical HG nerve ENG recorded by using tripolar configuration is
shown in Fig.
4A. The
phasic component of the HG nerve activity is modulated by the
respiratory activity. Under hypoxic conditions, the recorded amplitudes
increased and were >10 µV (peak to peak) during inspiration in all
three animals. The peak and baseline values in integrated traces (Fig.
4B) were 1.74 ± 0.21 and 0.72 ± 0.11 µV (n = 3),
respectively. The maximum signal-to-noise ratio averaged for all the
animals was 2.44 ± 0.18 (n = 3).
The electronic noise gave a baseline value of 0.43 ± 0.02 µV in
the integrated traces.
Power spectra.
The power spectra calculated from 1-s-long epochs taken during the
inspiratory phase and a preinspiratory phase are shown in Fig.
5. During inspiration, most of the power is
found between 500 Hz and 6 kHz. There is a considerable difference in
power between inspiratory and expiratory (baseline) phases. The power spectrum of the electronic noise is also shown.
Fig. 5.
Power spectra (n = 30,000, M = 128, 50% overlap; see MATERIALS AND METHODS for
description) of HG nerve activity during inspiration (1-s-long
epoch at peak activity) and baseline (second before next breath starts)
recorded by using nerve cuff electrode in tripolar mode. Third trace
(dashed line), power spectrum of electronic noise
(n = 30,000, M = 128, 50% overlap).
[View Larger Version of this Image (15K GIF file)]
In Fig. 6, the power spectra of the HG
nerve activities recorded with hook and cuff electrodes from four
separate electrode-nerve preparations during increased respiratory
drive are shown. Both cuff electrode recordings have a much wider
frequency band compared with that of the hook electrodes (100 Hz-2 kHz
vs. 500 Hz-6 kHz).
Fig. 6.
Power spectra (n = 30,000, M = 128, 50% overlap; see MATERIALS AND
METHODS) of HG nerve activity recorded with tripolar
cuff and bipolar hook electrodes. Each spectrum is obtained from a separate electrode-nerve preparation. Each plot is normalized with
respect to its own peak for comparison.
[View Larger Version of this Image (21K GIF file)]
In Fig. 7, the power spectra of the HG
nerve activity recorded by configuring the spiral nerve cuff electrode
in tripolar and bipolar modes are compared. The overall frequency
bandwidth for both spectra is approximately between 500 Hz and 4 kHz.
The spectrum shown for the bipolar cuff is bimodal with the higher mode
(2.5-4 kHz), which is similar to the spectrum from the tripolar electrode. However, the lower frequency components (500 Hz-2.5 kHz)
have much higher amplitudes.
Fig. 7.
Power spectra (n = 1,020, M = 512, 50% overlap; see MATERIALS AND METHODS) of short HG nerve
recordings (10 ms) acquired by configuring spiral nerve cuff electrode
in tripolar and bipolar modes. Raw data epochs are acquired one
immediately after another to avoid changes in respiratory drive level.
Data are sampled at 100 kHz after filtering with 300-Hz to 10-kHz band
filter. In this trial, tripolar and bipolar modes gave peak values of 1.95 and 2.42 µV, respectively, in rectified and integrated versions of recordings.
[View Larger Version of this Image (16K GIF file)]
DISCUSSION
Signal-to-noise ratios.
The signal-to-noise ratios of the HG nerve recordings obtained with the
cuff electrode are in an acceptable range and can detect modulations of
nerve activity by changes in respiratory drive (Figs. 3 and 4). The
baseline value of the integrated signals is only slightly above the
electronic noise level (0.72 ± 0.11 vs. 0.43 ± 0.02 µV). The
fact that the tonic activity level is so small can be attributed to
anesthesia (11). In unsedated cats, significantly higher levels of
tonic discharge of the genioglossus, which is innervated by the HG
nerve, were found (8).
The cuff length (i.e., electrode separation) is one of the important
parameters that determine the signal amplitude. The amplitudes of the
recorded signals first increase as the cuff length increases and then
reach saturation (16). Faster fibers require longer cuffs for
saturation. Thus we consider only the largest fibers for evaluation of
the cuff length. The largest fibers in the cat HG nerve have a diameter
of 11.5 µm (1). The estimated conduction velocity for a myelinated
fiber of this caliber is ~64 m/s (10). Although not saturated, the
signal amplitudes are large when the activity of a fiber of this
diameter is recorded with a 10-mm-long cuff from the cat sural nerve,
which has a size similar to the HG nerve (16). Thus the cuff length,
which was limited by the dissectable length of the nerve, was chosen to
be 10 mm in these experiments. In hook electrode preparations, Vaseline
is applied to the nerve/electrode, and it is very difficult to
determine the length of the restricted extracellular space surrounding
the nerve. For this reason, one can expect large variations in the signal amplitude within an experiment as well as between experiments. Temporal variations in the degree of nerve dehydration and the damage
introduced to the nerve complicate the issue further.
The signal amplitude is also dependent on the electrical configuration
of the recording electrode (Fig. 7). When configuration is switched
from tripolar to bipolar mode on the same cuff electrode, the
integrated peak signal amplitudes were increased (1.95 vs. 2.42 µV).
This increase was possibly due to the large components added in the
lower frequency range of the bipolar spectrum.
Another important factor that determines the signal amplitude is the
thickness of the extraneural medium inside the cuff that shunts the
electrical potentials to be recorded. The fluid from the surrounding
tissue fills this space in acute experiments. A snugly fitting cuff
gives considerably larger amplitudes compared with a cuff that has a
relatively larger diameter (data not shown). An advantage of the
self-coiling cuff is that it squeezes the fluid out and increases the
resistivity of the extraneural space within the cuff, thus increasing
the signal amplitudes.
Frequency content of recordings.
The power spectra shown in Fig. 6 for hook and cuff electrodes differ
considerably. The frequency bandwidth of the plots for the hook
electrodes is much narrower than that of the cuff electrodes. A
possible explanation is the difference in the transfer function of the
electrodes. When whole nerve recordings are obtained with the cuff (or
hook) electrodes, the frequency spectrum of the nerve activity is
modified by the following transfer functions:
1) the nonlinear, spatial
frequency-dependent transfer function of the nerve trunk and the
surrounding medium as a volume conductor, including the restrictions on
the extraneural space, i.e., the cuff (or Vaseline with the hook
electrodes) (3, 19); and 2) the
filter function determined by the location (separation) and the
electrical configuration of the metal contacts (discussed below).
The transfer function of the volume conductor can be altered by
restricting the extraneural space, e.g., surrounding the nerve with a
nonconductive material such as Silastic, Vaseline, paraffin, or mineral
oil. Because the restricted space provided by a nonsolid material like
Vaseline is prone to mechanical disturbances, the frequency response of
the recording system will vary throughout the experiment. Moreover, the
extent of the extraneural restriction space can vary from experiment to
experiment because the length of insulation is not constant. This may
adversely affect the reproducibility of the recordings, especially in
long trials. Mechanical stability and rigidity provided by the nerve
cuff electrodes eliminate these problems.
The design parameters of the cuff electrode, i.e., separation and
electrical configuration of the metal contacts, also affect the overall
transfer function of the recording system. The tripolar configuration
is used for its superior immunity to electrical disturbances from
surrounding muscles and power lines (17), which may not be the primary
concern in acute recordings conducted under anesthesia. Bipolar
configuration is more common with the hook electrodes in acute
experiments for practical reasons. The tripolar cuff provides the
second spatial derivative of the extraneural potentials, whereas the
bipolar configuration gives the first spatial derivative (18). Thus the
nerve cuff electrode is, in fact, a linear filter in the spatial
frequency domain, with a transfer function determined by the separation
and the configuration of the metal contacts (simulation data not
shown). These spatial transfer functions will be transformed into the
temporal frequency domain after being scaled with the propagation
velocity of the action potentials. The higher the conduction velocity,
the further the transfer function spectrum is spread over higher
frequencies. Thus the temporal frequency spectrum of the recordings
should also be a function of the cuff length and the contact
configuration. This is supported by the plots shown in Fig. 7. Although
the overall frequency range occupied by both recordings is
approximately the same, the bipolar recording has much larger
components at the low end of the spectrum than does the tripolar
recording. This suggests that the bipolar configuration has higher
gains in the lower frequency range.
Finally, the bimodal spectra of the recordings in Fig. 7 could be
explained by desynchronization of the fiber activity. It has been shown
that as a result of desynchronization in the fibers' activation, dips
can exist in the power spectrum of compound action potentials and that
the frequencies of dips would depend on the degree of desynchronization
(2).
We conclude that the signal-to-noise ratios obtained with the spiral
nerve cuff electrode recordings from intact HG and Phr nerves are large
enough to allow one to use them as a measure of respiratory output
under different respiratory drive levels. The mechanical stability
provided by the cuff electrode can be crucial for the stability of the
transfer function of the nerve-electrode system, and hence, the
reproducibility of the recordings. The frequency content and the signal
amplitude of the recordings depend on the cuff length (contact
separation) and the electrical configuration of the metal contacts. The
bandwidth of the signals obtained with our cuff electrodes was much
broader than that of the hook electrodes.
ACKNOWLEDGEMENTS
This study was supported by The Ministry of National Education of
Turkey and National Heart, Lung, and Blood Institute Grant HL-25830.
FOOTNOTES
The results of this study were partially published in abstract form in
the Proceedings of the 16th Annual International Conference of
the Institute of Electrical and Electronics Engineers/Engineering in
Medicine and Biology Society (see Ref. 15).
Address for reprint requests: D. M. Durand, Applied Neural Control
Laboratory, Bolton Bldg., Rm. 3510, Dept. of Biomedical Engineering,
Case Western Reserve Univ., 10900 Euclid Ave., Cleveland, OH 44106 (E-mail: dxd6{at}po.cwru.edu).
Received 26 April 1996; accepted in final form 7 March 1997.
REFERENCES
| 1.
| Bloom, S. Afferent influences on tongue muscle activity.
Acta Physiol. Scand. 9, Suppl. 170: 23-37,
1960.
|
| 2.
|
Dimitrov, G. V.,
Z. C. Lateva,
and
N. A. Dimitrova.
Use of potential spectral characteristics for solving the inverse problem in electroneurography.
Med. Biol. Eng. Comput.
30:
399-405,
1992[Medline].
|
| 3.
|
Ganapathy, N.,
and
J. W. Clark, Jr.
Extracellular currents and potentials of the active myelinated nerve fiber.
Biophys. J.
52:
749-761,
1987[Abstract/Free Full Text].
|
| 4.
|
Haugland, K. H.,
and
J. A. Hoffer.
Slip information provided by nerve cuff signals: Application in closed-loop control of functional electrical stimulation.
IEEE Trans. Rehabil. Eng.
2:
29-36,
1994.
|
| 5.
|
Haugland, K. H.,
and
T. Sinkjaer.
Cutaneous whole nerve recordings used for correction of footdrop in hemiplegic man.
IEEE Trans. Rehabil. Eng.
3:
307-317,
1995.
|
| 6.
|
Haxhiu, M. A.,
J. Mitra,
E. van Lunteren,
E. N. Bruce,
and
N. S. Cherniack.
Hypoglossal and phrenic responses to cholinergic agents applied to ventral medullary surface.
Am. J. Physiol.
247 (Regulatory Integrative Comp. Physiol. 16):
R939-R944,
1984.
|
| 7.
|
Haxhiu, M. A.,
J. Mitra,
E. van Lunteren,
N. Prabhakar,
E. N. Bruce,
and
N. S. Cherniack.
Responses of hypoglossal and phrenic nerves to decreased respiratory drive in cats.
Respiration
50:
130-138,
1986[Medline].
|
| 8.
|
Haxhiu, M. A.,
E. van Lunteren,
J. Mitra,
and
N. S. Cherniack.
Responses to chemical stimulation of upper airway muscles and diaphragm in awake cats.
J. Appl. Physiol.
56:
397-403,
1984[Abstract/Free Full Text].
|
| 9.
|
Haxhiu, M. A.,
E. van Lunteren,
W. B. van de Graaff,
K. P. Strohl,
E. N. Bruce,
J. Mitra,
and
N. S. Cherniack.
Action of nicotine on the respiratory activity of the diaphragm and genioglossus muscles and the nerves that innervate them.
Res. Physiol.
57:
153-169,
1984[Medline].
|
| 10.
|
Hursh, J. B.
Conduction velocity and diameter of nerve fibers.
Am. J. Physiol.
127:
131-139,
1939.
|
| 11.
|
Hwang, J. C.,
W. M. St. John,
and
D. Bartlett, Jr.
Respiratory-related hypoglossal nerve activity: influence of anesthetics.
J. Appl. Physiol.
55:
785-792,
1983[Abstract/Free Full Text].
|
| 12.
|
Naples, G. G.,
J. T. Mortimer,
A. Scheiner,
and
J. D. Sweeney.
A spiral nerve cuff electrode for peripheral nerve stimulation.
IEEE Trans. Biomed. Eng.
35:
905-916,
1988[Medline].
|
| 13.
|
Nicolic, Z. M.,
D. B. Popovic,
R. B. Stein,
and
Z. Kenwell.
Instrumentation for ENG and EMG recordings in FES systems.
IEEE Trans. Biomed. Eng.
41:
703-706,
1994[Medline].
|
| 14.
|
Popovic, D. B.,
R. B. Stein,
K. L. Jovanovic,
R. Dai,
A. Kostov,
and
W. W. Armstrong.
Sensory nerve recording for closed-loop control to restore motor functions.
IEEE Trans. Biomed. Eng.
40:
1024-1030,
1993[Medline].
|
| 15.
|
Sahin, M.,
D. M. Durand,
and
M. A. Haxhiu.
Whole nerve recordings with the spiral nerve cuff electrode.
In: Proceedings of the Sixteenth Annual International Conference of the IEEE Engineering in Medicine and Biology Society Baltimore MD 1994. Piscataway, NJ: IEEE, 1994, vol. 1, p. 372-373.
|
| 16.
|
Stein, R. B.,
D. Charles,
L. Davis,
J. Jhamandas,
A. Mannard,
and
T. R. Nichols.
Principles underlying new meth ods for chronic neural recordings.
J. Can. Sci. Neurol.
2:
235-244,
1975.
|
| 17.
|
Stein, R. B.,
T. R. Nichols,
J. Jhamandas,
L. Davis,
and
D. Charles.
Stable long term recordings from cat peripheral nerves.
Brain Res.
128:
21-38,
1977[Medline].
|
| 18.
|
Stein, R. B.,
and
K. G. Pearson.
Predicted amplitude and form of action potentials recorded from unmyelinated nerve fibers.
J. Theor. Biol.
32:
539-558,
1971[Medline].
|
| 19.
|
Trayanova, N.,
C. S. Henriquez,
and
R. Plonsey.
Extracellular potentials and currents of a single active fiber in a restricted volume conductor.
Ann. Biomed. Eng.
18:
219-238,
1990[Medline].
|
0161-7567/97 $5.00
Copyright © 1997 the American Physiological Society
