Download Response from the exposed intracranial human auditory nerve to

Hearing Research, 38 (1989) 163-176
Elsevier
HRR
163
01181
Response from the exposed intracranial human auditory nerve to
low-frequency tones: Basic characteristics
Aage R. Mprllerand Hae Dong Jho
Department
of Neurological Surgery, University of Pittsburgh School of Medicine, Pittsburgh, Pennsylvania, U.S.A.
(Received
5 June 1988; accepted
3 November
1988)
The responses recorded from the exposed intracranial
portion of the eighth nerve in man with normal hearing to short bursts of
low-frequency
tones (500, 1000, and 1500 Hz) consist of two components;
these two components
can be separated by adding and
subtracting,
respectively,
the responses to tonebursts
of opposite polarity. Subtracting
the responses to tones of opposite polarity
reveals a waveform that resembles the sinusoidal waveform of the stimulus (frequency-following
response = FFR), while adding the
responses to tones of opposite polarity reveals a slow component,
the waveform of which is more variable than the frequency-following component.
The initial deflection of the slow component
of the response to 1000 Hz and to 1500 Hz is a positive peak followed
by a slow, negative deflection, and the response to 1500-Hz tonebursts often shows a clear off-response.
The slow component
of the
response to 500-Hz tones often has an initial negative peak followed by a slow, positive or negative wave. The temporal relationship
between the stimulus tone and the frequency-following
component
changes only slightly when the intensity of the sound is changed,
whereas the latency of the slow potential decreases with increasing stimulus intensity.
The FFR can be masked by noise, and the results of masking with highpass-filtered
noise indicate that the frequency-following
response may be generated at a location on the basilar membrane that is tuned to a frequency that is higher than that of the stimulus
tone.
Human
auditory
nerve;
Frequency-following
response;
Intraoperative
Introduction
Goldstein and Kiang (1958) showed that the
gross response that can be recorded from the
auditory nerve is related to the discharges in individual nerve fibers. The former is the convolution
between the distribution of neural discharges over
the population of nerve fibers that contribute to
the response and the waveform of a single nerve
impulse. Then, in 1962, Teas et al. presented results that were fundamental to the interpretation
of auditory nerve action potentials and the distribution of excitation along the basilar membrane
in the guinea pig. They used selective masking
with narrowband noise of the response to broad-
Correspondence to: Aage R. Moller, Dept. of Neurological
Surgery, Room 9402, Presbyterian-University
Hospital,
DeSoto at O’Hara Streets, Pittsburgh,
PA 15213, U.S.A.
0378-5955/89/$03.50
0 1989 Elsevier Science Publishers
recordings
band clicks, to define areas of excitation along the
basilar membrane.
It was earlier assumed that it was only the
rarefaction phase of a sound, corresponding to a
deflection of the basilar membrane towards the
Scala vestibuli, that was excitatory. More recently,
it has been shown that the neural excitation that
occurs in the cochlea is far more complex. Thus,
Konishi and Nielsen (1973, 1978), Zwislocki and
Sokolich (1973), Zwislocki (1974), and Sokolich et
al. (1974) found that, when recording from single
auditory nerve fibers using sounds with sinusoidal
and trapezoidal waveforms, the motion of the
basilar membrane towards the Scala tympani was
excitatory in the majority of nerve fibers. Some
fibers responded to basilar membrane velocity,
depending on the frequency to which the nerve
fiber is tuned (CF) and on the frequency and
intensity of the sound. Since basilar membrane
displacement is related to stapes velocity, the
B.V. (Biomedical
Division)
164
majority of nerve fibers will respond to the derivative of the condensation phase of low-frequency
sounds.
The timing (phase angle) of excitation of single
auditory nerve fibers in response to pure tones
shows a transition in the mid CF range, and there
is thus a low and high range of CF where the
phase angle of excitation shows little change as a
function of CF for low stimulus frequencies. In
between these two ranges the phase of excitation
changes rapidly as a function of its CF (about 540
degrees/octave). Fibers with a CF that is higher
than twice that of the stimulus tone change slowly
as a function of their CF (about 90 degrees/octave) (Rim et al., 1979; Ruggero and Rich, 1983,
1987; Ruggero et al., 1986).
Since the phase-locked component of the gross
responses from the auditory nerve to lowfrequency tones of moderate to high intensity can
be assumed to be the vector sum of the phaselocked activity in many single auditory nerve
fibers, the responses from fibers with CF below
twice the frequency of the stimulus tones are
probably unlikely to contribute significantly to the
gross response, which therefore can be assumed to
be a result of the phase-locked activity in fibers
with CF above twice the frequency of the tone.
Only the phase-locked response to low-intensity
tones (near threshold) may be assumed to originate
near the location of maximal deflection on the
basilar membrane. This was in fact confirmed by
Snyder and Schreiner (1985) in experiments in
cats. These investigators used forward masking to
show that it is only tones near threshold that
originate from a region of the basilar membrane
that is tuned to the frequency of the tone, and that
the location on the basilar membrane which contributes maximally to the phase-locked gross response shifts towards the base of the cochlea as
the intensity of the tone is increased. The response
to loud tones originates from regions on the basilar membrane that are tuned to frequencies that
are about twice the frequency of the stimulus
tones (Snyder and Schreiner, 1985). In interpreting these results it must also be taken into account
that the width of the tuning increases as the sound
intensity is increased (Rhode, 1971, 1978; Sellick
et al., 1982b). Recordings from single auditory
nerve fibers show a similar widening of tuning
when broadband noise is used as the test signal,
and there is a shift in the CF that indicates that
the location of maximal excitation on the basilar
membrane shifts towards the base of the cochlea
with increasing stimulus intensity (Moller, 1977,
1983).
These results obtained in animals made it seem
likely that information about the properties of
human neural discharges could be obtained from
recordings of gross responses from the human
auditory nerve. This would be advantageous, as it
is usually not possible to obtain recordings from
single units of the human auditory nerve, but it is
relatively easy to obtain recordings of gross potentials.
The purpose of this study was to provide information about the human ear and auditory nerve
that may not be directly inferred from the results
of studies in which experimental animals such as
the cat are used. Because of possible differences in
the cochleas of man and such small animals, e.g.,
the auditory nerve is much longer in man (Lang,
1981) than it is in such animals, the results of
animal studies are not always applicable to man.
Studying the responses recorded from the human
auditory nerve to tones may permit correlation of
physiological measurements with psychoacoustic
measurements made in patients with different
types of sensorineural hearing loss. Thus, this report presents results of recording from the
intracranial portion of the human eighth nerve in
response to low-frequency tones.
Patients and Methods
Recordings were made from the exposed eighth
nerves of patients undergoing neurosurgical operations to relieve trigeminal neuralgia (TN), hemifacial spasm (HFS), or disabling positional vertigo
(DPV). The technique used was microvascular
decompression (MVD) of the respective cranial
nerve(s) (see Jannetta, 1981a,b). The patients who
were operated upon to relieve TN and DPV were
anesthetized using a balanced anesthesia regimen
with a strong analgesic (Fentanyl) and a muscle
relaxant (Pavulon) together with nitrous oxide.
Patients operated upon to relieve HFS were
anesthetized using inhalation agents (isoflurane
165
and nitrous oxide) and small amounts of Fentanyl,
but no long-lasting muscle relaxants.
The stimulus sounds consisted of bursts of pure
tones at frequencies of 500, 1000, and 1500 Hz,
with a plateau of 10 ms and a rise and fall time of
two periods. Clicks were generated by applying
rectangular pulses of 100~ps duration at a repetition rate of 19 or 9 pulses per s (Grass Instrument Co., Type SlO CTCMA) using miniature
stereo earphones (Radio Shack, Realistic) that were
sealed in the patient’s ear with adhesive tape. The
masking noise was generated by the same
audiostimulator and was added to the clicks and
tonebursts after being highpass filtered (roll-off of
54 dB/octave). The characteristic of the sound
system was determined by attaching the earphone
to a 2-0~ cavity partly filled with cotton, and
measuring the sound pressure with a calibrated
l/Cinch condenser earphone (Brtiel and Kjaer,
Type 4134).
The potentials were recorded from the exposed
eighth nerve using a monopolar electrode made
from a fine, malleable, Teflon-insulated,
multistrand, silver wire with a small, cotton wick sutured to its uninsulated tip (Moller and Jannetta,
1983). A noncephalic reference electrode (Grass
Instrument Co., needle electrode Type E2) was
placed on the shoulder or on the contralateral
earlobe. A similar needle electrode placed over the
sternum served as a ground electrode. The potentials were amplified using differential amplifiers
(Grass Instrument Co., Type 12) with a gain of
20,000 times and filter settings of 3 to 3000 Hz
(roll-off of 6 dB/octave). After additional lowpass
filtering (3.4kHz
cut-off and a roll-off of 18
dB/octave), the potentials were averaged using an
LSI 11/73 microprocessor with an analog-to-digital converter with 1Zbit resolution. Fifty to 200
responses of the tones were averaged and stored
on computer disks, after which the tones were
inverted’ and a similar number of responses were
averaged and stored. The responses to opposite
polarities of the stimuli were added or subtracted
to separate the FFR from the slow response. In
some cases digital filtering was used to remove
noise from the FFR. These filters were zero-phase,
bandpass filters that were implemented by convolving the response with a weighting function
that consisted of 3/2 sine waves with a frequency
equal to that of the stimulus tones. The bandwidths (at 3 dB points) of these filters were 360
Hz at 500 Hz and 700 Hz at 1000 Hz.
Brainstem auditory evoked potentials (BAEP)
were recorded during the operation using platinum
needle electrodes (Grass Instrument Co., Type E2)
placed on the vertex and the ipsilateral earlobe.
These potentials were amplified and processed in
the same way as were the potentials recorded
directly from the eighth nerve. The recording that
was made before the operation but after the patient was anesthetized served as the baseline recording. Comparison of the potentials that were
recorded during the operation to that of the baseline recording was made to ascertain that the
auditory nerve had not been injured before the
responses to low-frequency tones had been recorded from the nerve.
The recordings presented in this report were
made as part of the routine intraoperative monitoring of auditory evoked potentials that is done
in patients in our institution. Such monitoring
minimizes the risk of hearing loss due to manipulation of the eighth nerve. The research is done
with patients’ informed consent according to a
procedure approved by the Institutional Review
Board of the University of Pittsburgh School of
Medicine.
ReSUltS
The results of the present study are based on
recordings made in 8 patients with normal hearing. The responses that can be recorded from the
exposed eighth nerve in man to bursts of 500-Hz
pure tones typically show an initial positive peak
followed by a pattern of negative peaks (Fig. 1).
Inverting the phase of the tone (Fig. 1, bottom
tracings) results in the peaks in the response being
shifted one-half wave of the sound; this indicates
that it is predominantly one phase of the sound
that gives rise to the peaks in the response.
In order to extract the components of the responses to low-frequency tones that are related to
the envelope of the sound (slow potential) from.
those that are related to the waveform of the
sound we added and subtracted, respectively, the
responses to opposite phases of the tones. When
the response to tone is subtracted from the re-
166
sponse to the same tone presented with reversed
phase, a nearly sinusoidal pattern with the same
frequency as the stimulus tone is seen; this phaselocked part of the response lasts as long as the
toneburst (Fig. 2). We have chosen to call this part
of the response the ‘frequency-follo~ng
response’
(FFR). However, the same term was used by
Worden and Marsh (1968) and Moushegian et al.
(1978), as well as others, to describe the entire
neural response to low-frequency tones (in contrast to cochlear microphonic (CM) potentials).
When the responses to tones of opposite phase
are added, the result is a slow potential that may
be regarded to reflect the response to the envelope
SOUND
I
’
:
,
’
.,
1:
‘/
’
‘,
‘,
,’
:
I
I
I
I
I
0
5
10
15
20
i
,
25 ms
Fig. 2. Responses to 500-Hz tonebursts of both polarities
superimposed (top tracing), together with the sums and differences (middle and lower tracings), to show the separation of
the two components of the response. The sound intensity was
110 dB SPL, and the recordings were obtained in a patient who
was undergoing an operation for DPV.
,‘,
,
DIFFERENCE
I
RESPONSE
i
:
‘,
;
i
,,
,\,
‘,
:
,,$,:
it:!/
’
I
,’
(
‘, 1
,,
:’
e’
!
,8
/
,’
’
,
’
:
:
/
‘s,
:
‘,
‘,
I
:
;
,’
,,_
I
I
I
I
0
5
10
15
I
20 Ins
Fig. 1. Responses recorded from the exposed etghth nerve to
MO-Hz tonebursts of both polarities in a patient who was
operated upon for trigeminal neuralgia. The sound pressure,
delivered by the earphone and measured in a slightly damped
~-CCcavity, was 113 dB SPL. The waveform of the stimulus
sound is shown below each recordii
with rarefaction as
upward deflections. The same convention is used in all subsequent recordings. Negative potentials are shown as upward
deflections in this and all subsequent figures.
of the sound. The FFR is largely cancelled as the
result of this procedure.
The amplitude of the FFR decreases with increasing frequency of the tones (Fig. 3,A), and the
waveform of the slow component of the response
is often different for different frequencies of the
stimulus tones. Thus, the slow component of the
response to 1000 Hz and 1500 Hz tones typically
has an initial positive deflection followed by a
slow negativity that lasts as long as the sound, as
may be seen in Fig. 3,B. The response to 1500-I&
tones, in addition, has a clear off-response. The
response to 500 Hz is often different in that it
frequently has an initial negative peak followed by
a slow wave, and its amplitude is usually smaller
than that of the response to 1000 and 1500 Hz
(Fig. 3,B).
It may also be seen from Fig. 3 that the latency
of the onset of the slow potential, as well as that
167
B
DIFFERENCE
500
500
Hz
Hz
4
1500
1000
Hz
1500
Hz
H>
-d
CLICK
I
3
15pv
I
I
I
I
I
0
5
10
15
20
Ins
0
I
I
I
5
10
15
I
20 ms
Fig. 3. Differences (A) and sums (B) of the responses to tonebursts of opposite polarities. Responses to tonebursts at 500, 1000, and
1500 Hz are shown for another patient who underwent an operation to relieve DPV. The sound levels were 104, 95, and 87 dB SPL
for the three frequencies. The lower curves show the responses to click stimulation (dashed lines represent rarefaction
clicks and solid
lines are the responses to condensation
clicks).
of the FFR, is longest for the response to 500 Hz
and shortest for the response to 1500 Hz, but the
latency of the response to tones is longer than the
latencies of the responses to click sounds. These
latter responses are shown both for rarefaction
(dashed lines) and condensation (solid lines) clicks
presented at a peak SPL of 115 dB. Results of a
more detailed analysis of the latencies of the FFR
component will be given later in this paper.
The amplitudes of both the FFR and the slow
components of the response decrease as the
stimulus intensity is decreased. However, the slow
potential in response to lOOO-Hz tones can usually
be discerned at lower sound intensities than the
FFR (Fig. 4), while the slow component of the
response to 500 Hz is often poorly developed (Fig.
5). The latency of the FFR in the range of relatively high stimulus intensities used is nearly independent of the stimulus intensity, but the latency
of the initial positive peak of the slow component
increases when the stimulus intensity is decreased.
It may also be seen from Fig. 4 that the amplitude
of the individual peaks of the response at the
highest sound intensity is largest in the beginning
of the toneburst, after which it decreases, indicating adaptation of the phaselocked response. The
FFR component of the response to 500-Hz tonebursts at different sound intensities also shows
that the latency is independent of the stimulus
intensity in the intensity range that was studied
Fig. 4. The FFR component (A) and the slow component (B) of the response to lC@O-Hz tonebursts at different intensities (given in
dB SPL). The results were obtained in a patient who was undergoing a MM, operation for HFS.
(Fig. 5), but the slow component is much less
prominent. When the sound level is decreased, the
slow component becomes indiscernible when the
FFR still is clearly visible.
The responses to low-frequency sounds of high
intensities have distorted waveforms: peaks appear between the main peaks of the response (Fig.
6). Such distortion products are cancelled when
the responses to tones of opposite polarities are
subtracted to obtain the FFR, but the distortion
can be observed when the responses to stimuli of
opposite polarity are added: at this time it is seen
as a ripple superimposed on the slow potential
(Fig. 5,B). The response to 500-Hz tones usually
shows a higher degree of distortion than the response to lOO@Hz and 1500-Hz tones. However,
there is considerable variation from person to
person.
The traditional way of determining the latency
of a response is to measure the time between the
beginning of the sound and the beginning of the
response. Since the sound must have a smooth
onset to avoid transients, this method may be
inaccurate when low-frequency tones are used, so
we determined the latency of the FFR in a different way as well, namely, by measuring how much
the waveform of the sound needed to be shifted in
time to match the waveform of the response (Fig.
7). To do this, we assumed that it is the rarefaction phase of the sound that results in neural
excitation, and that neural excitation produces a
negative nerve potential. In turn, the assumption
169
B
A
DIFFERENCE
104
dB
104
il
dB
94 dB
94 dB
84 dB
84 dB
74 dB
74 dB
SOUND
I
0
I
I
I
5
10
15
I
20 Ins
I
0
I
I
I
I
5
10
15
20
ms
Fig. 5. Differences (A) and sum (B) of the responses to 500 Hz shown in a way similar to that in Fig. 4 to 500-I-k tones.
that neural excitation produces a negative potential on the surface of the nerve presumes that the
recording is made from a position on the nerve
where a propagated wave of nerve activity passes
under the electrode. This can be achieved by recording from the eighth nerve at a distance from
its entry into the brainstem (as long as the nerve is
not injured so that neural transmission is blocked).
The latency of the FFR to 500-Hz and lOOO-Hz
tones measured in that way was 3.2 and 2.2 ms,
respectively, in the patient whose responses are
shown in Fig. 7. This should be compared with the
latency of the main negative peak in the CAP,
recorded in response to rarefaction clicks immediately before the response to tonebursts was
obtained; the latency of the main negative peak in
the CAP to clicks was 3.3 ms in the patient
depicted in Fig. 7.
The time difference between the second rarefaction wave of the sound (as measured in a
calibration coupler) and the second negative peak
in the response to 500-Hz and lOOO-Hz tonebursts
in the 8 patients in this study was slightly longer
for the 500-Hz tonebursts than for the 1000~Hz
tonebursts (Table I). The latency when measured
in this way is a measure of how much the sound
waves had to be shifted to obtain the best match
between the waveform of the responses and that
of the tonebursts (c.f., Fig. 7). The latency of the
negative peak in the CAP in response to high-intensity rarefaction clicks was slightly longer than
the latency for the response to lOOO-Hz tones.
There was, however, considerable individual variation.
The assumption that the responses to the lowfrequency tones that we studied are generated by
94 de
‘,
!
’
,
(‘,j’,
,
1
‘,
,’
I
5vv
reduced to about half its amplitude
when the
cut-off frequency of the masking noise is 1800 Hz
(spectrum
level of 60 dB SPL/Hz).
When the
cut-off frequency
is increased
to 3300 Hz the
effect on the FFR is minimal.
Recordings
from
another patient show that masking with noise with
a cut-off frequency of 1900 Hz (60 dB spectrum
level) almost eliminates
the response to lOOO-Hz
tones at 95 dB SPL, and noise with a cut-off
frequency
of 2800 Hz, and the same spectrum
level causes a marked reduction in the response to
the same lOOO-Hz tones. In the same patient, noise
with a cut-off frequency of 1200 Hz reduces the
response to 500-Hz tones (104 dB SPL) by about
50% (Fig. 9,B). There is no noticeable effect on the
timing of the FFR as a result of masking.
The effect from masking on the slow component of the response is more difficult to assess,
because this component
of the response is more
variable than the FFR.
SOUND
Discussion
Fig. 6. Unprocessed
responses to SOO-Hz tones of both polarities at two different stimulus intensities to show how distortion
increases with increasing stimulus intensity. The data are the
same as shown in Fig. 5.
The results of the present study, obtained
by
monopolar
recording
from the eighth nerve in
man, are essentially
similar to those obtained by
Snyder and Schreiner (1984, 1985) by recording
from cats’ eighth nerves using bipolar electrodes.
TABLE
the intracranial
portion of the audito& nerve is
supported by the finding that the latencies of the
responses changed when the recording
electrode
was moved along the intracranial
portion of the
eighth nerve: as can be seen from Fig. 8, the
latency was shorter when the recording electrode
was placed more distally compared to when it was
placed more proximally.
Further, the latency of
the response to click sounds changed slightly less
than did the latency of the FFR.
Simultaneous
masking with continuous
highpass-filtered
noise reduces the amplitude
of the
FFR in response to low-frequency
tones when the
cut-off frequency of the noise is lower than about
twice the frequency of the stimulus sound (Fig.
9,A). The FFR to lOOO-Hz tones at 95 dB SPL is
I
TIME DIFFERENCE
BETWEEN
THE SECOND
RAREFACTION
WAVE IN THE SOUND AND THE SECOND
NEGATIVE
DEFLECTION
IN THE
RESPONSE
TO
TONEBURSTS,
TOGETHER
WITH THE LATENCY
OF
THE NEGATIVE
PEAK IN THE RESPONSE
TO RAREFACTION CLICKS
Patient
500 Hz
1000 Hz
N,
1.
2.
3.
4.
5.
6.
7.
8.
2.92
3.32
3.20
2.82
3.72
3.72
3.62
3.30
2.72
3.00
2.85
2.80
2.45
2.20
2.44
3.08
3.12
2.90
3.20
2.92
3.00
3.20
Mean
S.D.
3.33
0.32
2.67
0.27
2.98
0.23
171
I
I
I
I
I
I
0
5
10
15
20
25
In*
I
I
I
I
I
0
5
10
15
20
I
25 m*
Fig. 7. Responses to 500-Hz (left) and 1000-Hz (right) tonebursts of both polarities (A,B), together with the dtiierencc between these
recordings (FFR) (C) (solid lines). The waveform of the stimulus sound, after the sound was shifted 3.2 and 2.2 ms, respectively
(dashed lines), is also shown to demonstrate
how latency of the FFR is measured. Below (D) is the response to broadband
clicks. The
results were obtained in a patient who was operated upon to relieve trigeminal neuralgia (same patient as illustrated in Fig. 1).
The effects on the response of masking with
continuous
highpass-filtered
noise that we see in
man are similar to those described by Snyder and
Schreiner (1985) using forward mashing in the cat.
Thus, our results support these investigators’
findings that the FFR are nerve potentials
and that
also, in man, the FFR is generated by nerve fibers
that are tuned to frequencies higher than those of
the stimulus tones.
That the timing of the FFR is relatively independent of stimulus intensity, while the latency of
the slow potential
decreases
with increasing
stimulus intensity,
indicates that the two components of the response,
the FFR and the slow
potential, may not have the same origin. The latter
seems to be dependent
on a synaptic build-up of
excitatory
postsynaptic
potentials
(or generator
potentials), which occurs at a slower rate when the
stimulus intensity is low than when the stimulus
intensity
is high. At the relatively high stimulus
levels used in this study, the timing of the FFR is
largely independent
of the stimulus intensity. Similar results were obtained
when utilizing
crosscorrelograms
of the responses to continuous
amplitude-modulated
noise (Moller,
1981a,b). The
difference (0.66 ms) in the delays of the responses
to 500 Hz and to 1000 Hz may be taken as a
measure of the difference in traveling time on the
basilar membrane
from which the responses
to
these two sounds originate
(probably
locations
that are tuned to 1000 and 2000 Hz).
The method used to measure the latency of the
FFR is based on the assumption
that it is the
rarefaction
phase of a sound that is excitatory. If
instead it was the condensation
phase that was
excitatory,
then the latency measurements
would
172
2PV
I
0
I
I
5
I
,o
I
15
I
20 ms
Fig. 8. FFR recorded from two different locations on the
eighth nerve: near the porus acousticus (distal), and at about
the middle of the intracranial portion of the nerve. The CAP
recorded from these two locations on the nerve in response to
click sounds are also shown. The results were obtained in,a
patient who was undergoing an MVD operation for disabling
positional vertigo (same patient as illustrated in Fig. 3).
be in error by the duration of one-half period of
the sound. In the same way, if it is the velocity of
a sound that is excitatory, then the error would be
one quarter of a wave or three quarters of the
duration of a wave. The fact that the FFR response is well defined makes it unlikely that the
match between the sound wave and the FFR
component of the response should be in error by a
full period of the sound wave due to uncertainty
in the determination of the beginning and the end
of the FFR response.
The phase shift of the middle ear is not taken
into account when the reference for comparing the
FFR with the stimulus is based on the sound
pressure that is measured in a calibration cavity. If
the timing of the FFR is to be compared with the
timing of the motion of the cochlear fluid, then
that phase shift must be taken into account. The
phase shift in the middle ear is a function of the
frequency of the sound, and it is less for a NO-Hz
tone than for a lOOO-Hz tone. On the basis of
measurements made in cats (Msller, 1963) the
displacement of the stapes (and the cochlear fluid)
will lag the sound pressure near the tympanic
membrane by about 30 to 40 degrees at 500 Hz,
while at 1000 Hz the displacement of the stapes
may be shifted by as much as 90 to 120 degrees,
relative to the sound pressure waveform at the
tympanic membrane, because that frequency is
slightly higher than the principal
resonance
frequency of the human middle ear (about 800
Hz) (Msller, 1963).
The distortion that is seen in the recorded
potentials to 500-Hz tones of high intensity is
dominated by second harmonics, thus resembling
a (partial) full-wave rectification. The extra peaks
are indications that the condensation phase of the
stimulus may become excitatory at high sound
levels. This may indicate that neural excitation in
the cochlea for low-frequency tones occurs for
deflections of the basilar membrane in both directions. However, there may be other reasons for
this type of distortion. Thus, several investigators
have described ‘peak splitting’ that can occur in
the responses from single nerve fibers to lowfrequency tones (Sellick et al., 1982a; Ruggero
and Rich, 1983; Ruggero et al., 1986; Kiang,
1984).
It seems less likely that the observed waveform
distortion is caused by nonlinearities in sound
transmission to the cochlea. Studies of sound
transmission in the middle ear in the cat show
linearity at sound levels well above those used in
the present study. The sound produced by the
earphones has no noticeable distortion at the sound
levels that were used in the present study. It is
interesting that the responses recorded in humans
seem to have less distortion than those seen in the
cat (Snyder and Schremer, 1984). This difference
is particularly noticeable at low sound levels, at
which the waveform of the responses in the cat
has a noticeable distortion, while the responses in
man have little distortion below 95 dB SPL at 500
Hz and at 1000 Hz.
173
1
1
I
I
I
0
5
10
15
20
I
25 ms
In interpreting the results of recording of auditory evoked potentials intraoperatively it must be
taken into account that such recordings are
hampered by technical problems that do not exist
when similar recordings are made in experimental
animals in the laboratory. For instance, it is not
practical to use shielded earphones intraoperatively, nor can bipolar recordings be made, and
therefore monopolar recordings have to be used.
In addition, the electrical and acoustic noise levels
in the operating room are much higher than they
are in the laboratory, and there are limitations on
the time that can be spent for such recordings.
We, however, do not find any evidence that these
factors have introduced noticeable errors in the
recordings reported on in this paper, but the
acoustic noise level in the operating room as well
as electrical noise and the limitations of time for
recording have not allowed us to record over a
large range of sound intensities, as can be done in
the laboratory.
Contrary to what has been a problem in similar
1200
I
I
I
I
I
I
0
5
10
15
20
25
1
Ill.3
0
HZ HP
I
I
I
I
5
10
15
20
I
25 Ins
Fig. 9. The effect on the FFR of masking with highpass-filtered noise. (A) Response to lOOft-Hz tonebursts, 95 dB SPL, without and
with masking at cut-off frequencies of 1800 and 3300 Hz. The level of the masking noise was 60 dB SPL/Hz (spectrum level). The
results were obtained in a patient operated upon for HFS. (B) Recordings similar to those obtained in A, but from another patient
using SOO-Hzand MOO-Hz tonebursts at 104 dB SPL and 95 dB SPL, respectively. The spectrum level of the masking noise was 60
dB. The responses were filtered digitally. The results were obtained in a patient operated upon for TN.
174
recordings in the cat (Snyder and Schreiner, 1984)
the recordings from the intracranial portion of the
eighth nerve in man were not noticeably contaminated by the CM potential. The reason for
this is most likely due to the longer distance
between the recording site and the cochlea. In
man this distance is about 1.5 cm (Lang, 1981)
while in the cat it is probably less than 0.5 cm.
The fact that the eighth nerve in man is not
usually immersed in fluid when the recordings are
made may reduce the conduction of the CM
potential to the recording site.
When recordings from the human eighth nerve
are used to study the normal physiology of the
auditory system, it is naturally important to make
sure that surgical manipulations to expose the
eighth nerve do not injure the auditory nerve,
since if this happens the results would naturally
not reflect the normal functioning of the ear and
the auditory nerve. To help identify cases in which
such injuries may have occurred, we always record
the farfield response (BAEP) before the beginning
of the operation and compare this recording with
recordings obtained simultaneously with the recordings from the exposed eighth nerve. For this
study, we excluded the results obtained in patients
in whom there had been a noticeable change in the
latencies of the peaks in the BAEP (about 0.5 ms).
This does not, however, assure that the function of
the auditory nerve was totally unchanged by the
surgical manipulations needed to expose the nerve,
but it is the best possible way available to monitor
auditory function in anesthetized patients. These
problems are not restricted to intraoperative recording in man, but are just as applicable to
results obtained in experimental animals in which
the eighth nerve must be exposed to obtain the
recordings.
Acknowledgements
This study was supported by a grant from the
National Institutes of Health (Grant No. 1 ROlNS21378-04). The authors are grateful to Peter
Jannetta, M.D., for making the patients under his
care available for this study, and to Margareta B.
Moller, M.D., Dr. Med. Sci., for audiological and
otological evaluation of the patients in this study.
References
Goldstein, J.L. and Kiang, N.Y.S. (1958) Synchrony of neural
activity in electric response evoked by transient
acoustic
stimuli. J. Acoust. Sot. Am. 30, 107-114.
Jannetta, P.J. (1981a) Hemifacial spasm. In: M. Samii and P.J.
Jannetta @is.), The Cranial Nerves. Springer-Verlag,
New
York, pp. l-8.
Jannetta,
P.J. (1981b) Vascular decompression
in trigeminal
neuralgia.
In: M. Samii and P.J. Jannetta
(I%.),
The
Cranial Nerves. Springer-Verlag,
New York, pp. 331-340.
Kiang, N.Y.S. (1984) Peripheral neural processing of auditory
information.
In: S.R. Geiger (Exe. Ed.), J.M. Brookhart,
V.B. Montcastle
(Eds.),
I. Darian-Smith
(Vol. Ed.),
Handbook
of Physiology, The Nervous System III, Chap.
15. American
Physiological
Society, Bethesda,
Maryland,
pp. 6399673.
Kim, D.O., Siegel, J.H. and Molnar, C.E. (1979) Co&ear
nonlinear phenomena in two-tone responses. Stand. Audiol.
(Stockh.) Suppl. 9, 63-81.
Konishi, T. and Nielsen, D.W. (1973) The temporal relationship between motion of the basilar membrane
and initiation of nerve impulses in the auditory
nerve fibers. J.
Acoust. Sot. Am. 53, 325.
Konishi, T. and Nielsen, D.W. (1978) The temporal relationship between basilar membrane motion and nerve impulse
initiation in auditory nerve fibers of guinea pigs. Jpn. J.
Physiol. (Tokyo) 28, 291-307.
Lang, J. (1981) Facial and vestibulocochlear
nerve, topographic
anatomy
and variations.
In: M. Samii and P.J. Jannetta
(Eds.), The Cranial Nerves. Springer-Verlag,
New York, pp.
363-377.
Msller, A.R. (1963) Transfer
function of the middle ear. J.
Acoust. Sot. Am. 35, 1526-1534.
Msller, A.R. (1977) Frequency
selectivity of single auditory
nerve fibers in response
to broadband
noise stimuli. J.
Acoust. Sot. Am. 62. 135-142.
Moller, A.R. (1981a) Neural delay in the ascending auditory
pathway. Exp. Brain Res. 43, 93-100.
Moller, A.R. (1981b) Latency in the ascending auditory pathway determined
using continuous
sounds: Comparison
between transient
and envelope
latency.
Brain Res. 207,
184-188.
Moller, A.R. (1983) Frequency
selectivity of phase-locking
of
complex sounds in the auditory nerve of the rat. Hear. Res.
11, 267-284.
Msller, A.R. and Jamretta.
P.J. (1983) Monitoring
auditory
functions
during cranial nerve microvascular
decompression operations
by direct recording from the eighth nerve.
J. Neurosurg.
59, 493-499.
Moushegian.
G., Rupert,
A.L. and Stillman,
R.D. (1978)
Evaluation
of frequency-following
potentials
in man:
Masking and clinical studies. Electroenceph.
Clin. Neurophysiol. 45, 711-718.
Rhode, W.S. (1971) Observations
of the vibration of the basilar
membrane
in squirrel monkeys using the Mossbauer
technique. J. Acoust. Sot. Am. 49, 1218-1231.
175
Rhode, W.S. (1978) Some observations
on cochlear mechanics.
J. Acoust. Sot. Am. 64, 158-176.
Ruggero, MA. and Rich, N.C. (1983) Chinchilla auditory-nerve
responses to low-frequency
tones. J. Acoust. Sot. Am. 73,
2096-2108.
Ruggero,
M.A., Robles, L. and Rich, N.C. (1986) Basilar
membrane mechanics at the base of the chinchilla cochlea.
II. Responses
to low-frequency
tones and relationship
to
microphonics
and spike initiation
in the VIII nerve. J.
Acoust. Sot. Am. 80, 137551383.
Ruggero,
M.A. and Rich, N.C. (1987) Timing of spike initiation
in cochlear
afferents:
Dependence
on site of
innervation.
J. Neurophysiol.
58, 3799403.
Sellick, P.M., Patuzzi, R. and Johnstone,
B.M. (1982a) Modulation of responses of spiral ganglion cells in the guinea
pig cochlea by low frequency sound. Hear. Res. 7,199-221.
Sellick, P.M., Patuzzi, R. and Johnstone,
B.M. (1982b) Measurement of basilar membrane
motion in the guinea pig
using the Mossbauer
technique.
J. Acoust. Sot. Am. 72,
131-141.
Snyder, R.L. and Schreiner, C.E. (1984) The auditory neurophonic: Basic properties.
Hear. Res. 15, 261-280.
Snyder, R.L. and Schreiner, C.E. (1985) Forward masking of
the auditory nerve neurophonic
(ANN) and the frequency
following response (FFR). Hear. Res. 20, 45-62.
Sokolich, W.G., Hamemik,
R.P., Zwislocki, J.J. and Schmiedt,
R.A. (1974) Inferred polarities
of co&dear hair cells. J.
Acoust. Sot. Am. 59, 963-974.
Teas, D.C., Eldredge,
D.H. and Davis, H. (1962) Cochlear
responses
to acoustic
transients:
An interpretation
of
whole-nerve
action potentials.
J. Acoust.
Sot. Am. 34,
1438-1459.
Worden,
F.G. and Marsh, J.T. (1968) Frequency-following
(microphonic-like)
neural responses evoked by sound. Electroenceph. Clin. Neurophysiol.
25, 42-52.
Zwislocki, J.J. and Sokolich, W.G. (1973) Velocity and displacement responses in auditory-nerve
fibers. Science 182,
64-66.
Zwislocki, J.J. (1974) A possible neuro-mechanical
sound analysis in the cochlea. Acustica 31, 354-359.