Hearing Research, 21 (1986) 91-95
Elsevier
91
HRR 00681
Representation
of a low-frequency tone in the discharge rate of populations
of auditory nerve fibers
W.P. Shofner and M.B. Sachs
~epari~e~~
of 3~o~edical Engineering, Johns Hopkins University School of Medicine, 720 Rowland Ave., Baltimore, MD 2I 205, U.S.A.
(Received
19 August
1985; accepted
23 October
1985)
Discharge rates for populations
of single auditory nerve fibers in response to 1.5 kHz tone bursts were measured in anesthetized
cats. Separate plots of average rate vs. best frequency (rate-place
profiles) were made for high, medium and low spontaneous
rate
(SR) auditory nerve fibers. At the lowest sound levels studied (34 dB SPL), all three SR groups show a peak in the rate-place
profile
centered around 1.5 kHz. At the highest sound levels studied (87 dB SPL), the average rates of the high and medium SR fibers are
saturated across a wide range of best frequencies, but a peak around 1.5 kHz is maintained in the rate-place
representation
of the low
SR fibers. These results show that in addition to the temporal information
present in the discharge patterns of auditory nerve fibers, a
rate-place
representation
of a single low-frequency
tone exists in the auditory nerve over a wide range of sound levels.
auditory
nerve, spontaneous
rate, rate-place
profile,
low frequency
At low sound pressure levels, a plot of average
discharge rate versus best frequency (BF) for the
population of auditory nerve fibers responding to
a single pure tone shows a peak centered around
the frequency of the tone (Evans, 1981; Kim and
Molnar, 1979; Pfeiffer and Kim, 1975). We refer
to such a plot as a rate-place profile (Sachs and
Young, 1979). The peak in the rate-place profile is
maintained over a wide range of sound pressure
levels for high-frequency tones (Evans, 1981), but
not for low-frequency tones (Kim and Molnar,
1979; Pfeiffer and Kim, 1975). The difference in
the population behavior for high- and lowfrequency tones is presumably related to the
sharper tuning curves of high-frequency auditory
nerve fibers [Evans, 1981). One interpretation of
this result is that tone frequency can be encoded
by average rate at high frequencies, while at low
frequencies temporal information must be used.
On the other hand, studies with more complex
stimuli such as tones in noise (Barta, 1985; Costalupes, 1985) and steady-state vowels (Sachs and
Young, 1979) have shown that a rate-place repre~378-5955/84/~3.50
5 1986 Elsevier Science Publishers
sentation of the stimulus spectrum is preserved
over a wide range of sound pressure levels by the
low spontaneous rate (SR) population of auditory
nerve fibers. Low SR auditory nerve fibers have
higher thresholds (Liberman, 1978) and wider dynamic ranges (Evans and Palmer, 1980; Sachs and
Abbas, 1974; Schalk and Sachs, 1980) than do
high SR auditory nerve fibers. In addition, low
and high SR fibers have different morphological
characteristics (Fekete et al., 1984; Liberman, 1982;
Liberman and Oliver, 1984; Rouiller et al., 1983).
The earlier studies of Pfeiffer and Rim (1975) and
Kim and Molnar (1979) did not separate low SR
fibers in a way which would enable us to determine if a rate-place representation of a lowfrequency tone is maintained at high stimulus levels
in .these fibers. Therefore, we have re-examined the
rate representation of a single low-frequency tone
and analysed the contributions of low, medium
and high SR fibers separately.
Methods
Two adult cats (2.0 and 3.4 kg) whose middle
ears were free of infection were used. They were
B.V. (Biomedical
Division)
92
injected intramuscularly
with 0.3 mg of atropine to
suppress mucous secretions and initially anesthetized with an intramuscular
injection of ketamine
(100-120 mg). The saphenous vein was cannulated
and cats were maintained
under anesthesia with
intravenous
injections of sodium pentobarbital.
A
tracheotomy was performed. Cats were placed in a
sound-proof
room (IAC-1204A)
and their body
temperature was maintained
between 36 and 38°C.
The external meatus was exposed and transected.
The bulla was exposed and a small hole was
drilled through the wall. A long piece of 0.86 mm
I.D. polyethylene
tubing (approximately
90 cm in
length) was cemented to the hole in order to vent
the bulla and prevent the development
of negative
pressure in the bulla (Guinan
and Peake, 1967).
The cranium between the tentorium
and nuccal
ridge was removed and the cerebellum was gently
retracted to expose the auditory nerve. Glass micropipettes
filled with 3 M NaCl (10-40 MS2)
were used to record single auditory nerve fibers.
Acoustic stimuli were delivered from an electrostatic driver (Sokolich, 1977) through a hollow ear
bar. Sound pressure at the eardrum was measured
with a probe tube for each animal. This system
typically has a flat frequency response (k 5 dB)
from 20 Hz to 20 kHz. Search stimuli consisted of
50 ms broadband
noise bursts with 10 ms rise/fall
times. The test stimulus was a digitally synthesized
1.5 kHz tone burst (400 ms duration;
10 ms
rise/fall times) presented once per second.
When a fiber was isolated, its best frequency
(BF) and threshold were determined
with audiovisual cues. Single fiber thresholds at BF of the
sampled population
were used to monitor
the
physiological
state of the cochlea. Spontaneous
rate was estimated
as the average rate over an
interval of at least 3 s in the absence of any
controlled acoustic stimuli. The 1.5 kHz tone was
presented at four standard sound pressure levels,
and average driven discharge rate was estimated
on-line for the interval between 20 and 400 ms
after the onset of the tone. Approximately
1000
spikes were collected in response to the 1.5 kHz
tone at each level studied.
Results
Rate
auditory
profiles of the sampled populations
nerve fibers to a 1.5 kHz tone burst
of
are
8
B
5
B
300
c
1
37
dB WI_
Best
Frequency
CkHz)
Fig. 1. Plots of average driven discharge rate as a function of
best frequency for 148 fibers studied in experiment 2/21/85
and 214 fibers studied in experiment 3/12/85.
The driven rate
of each fiber is in response to a 1.5 kHz tone at the sound
pressure levels indicated in the figure. x ,driven rates for fibers
with spontaneous
rates greater than 19 spikes/s
(high SR
fibers); A, driven rates for fibers with spontaneous
rates between 1 and 19 spikes/s (medium SR fibers); 0. driven rates of
fibers with spontaneous
rates less than 1 spike/s
(low SR
fibers). Triangularly
weighted moving window averages of the
and low (.
.) auditory nerve fibers are
high SR ( -)
shown. The moving window averages have bin widths of 0.250
octaves and are moved in 0.125 octave steps. The arrows point
to the 1.5 kHz place.
shown in Fig. 1 for two cats. These profiles are
plots of average discharge rate versus fiber BF.
Average rates were computed for the last 380 ms
of the 400 ms tone bursts. Data from high SR
fibers (SR greater than 19 spikes/s)
are plotted
with crosses; data from medium SR fibers (SR
between 1 and 19 spikes/s) are plotted with triangles; and data from low SR fibers (SR less than 1
spike/s) are shown by squares. The solid lines in
Fig. 1 show a moving window ,average of the rate
of only the high SR auditory nerve fibers. For
both cats, there is a peak in the rate profile of the
high SR population
centered around 1.5 kHz at
the lowest sound pressure levels we presented (Fig.
lA, C). However, at higher sound pressure levels
(Fig. 1 B, D) there is no clear peak around 1.5 kHz,
93
and the discharge rate is roughly constant across a
range of BFs from 1.0 to 6.0 kHz. In, order to
study more fibers with BFs near 1.5 kHz, we did
not study fibers with BFs greater than 10.0 and
generally not greater than 7.0 kHz. Note also that
at the low sound pressure levels (Fig. lA, C) the
average rate around 1.5 kHz is the same as the
average rate at high levels (Fig. lB, D), suggesting
that the high SR auditory nerve fibers with BFs
near 1.5 kHz are already saturated at these low
levels. This result is similar to that previously
reported by Rim and Molnar (1979) for a 1.0 kHz
tone.
The behavior of the low SR auditory nerve
fibers is clearly different from that of the high SR
population. The dotted line in Fig. 1 shows a
moving window average of the rate of only the low
SR auditory nerve fibers. At the higher sound
pressure levels where the discharge rate of the high
SR fibers is saturated across a wide range of BFs
(Fig. lB, D), the rate profile of the low SR fibers
shows a clear peak near 1.5 kHz.
The effect of increasing sound pressure level on
the rate profiles for the three populations of audi-
3112185
-IO,_,,,
I .o
0.1
Best
Frequency
100
(kHz)
Fig. 3. Scatter diagram showing the thresholds for low (0) and
high (x) SR auditory nerve fibers. Thresholds were determined
from audiovisual cues. Thresholds
for medium SR fibers have
been omitted for clarity. Mean thresholds of high and low SR
fibers with BFs between 3 and 6 kHz are 6.2 dB SPL and 17.1
dB SPL, respectively.
tory nerve fibers is further illustrated in Fig. 2.
Here we show only the moving window averages
for the low, medium and high SR populations of
I~i~~~~
~~~~~~
0.1
1.0
I 0.0
0.1
*
1.0
IO.0
Best
Frequency
1.0
I.0
IO.0
(kHz)
Fig. 2. Plots showing the moving window averages for the average discharge rate as a function of best frequency
levels indicated for experiments
2/21/85
and 3/12/85.
Moving window averages are shown for high (and low (. . . . . .) SR auditory nerve fibers
at the sound pressure
). medium
(- - - - - -)
94
auditory nerve fibers with dotted, dashed and solid
lines, respectively.
At 34 and 37 dB SPL, the
lowest level studied in each cat (Fig. 2A, E), all
three populations
of auditory nerve fibers show a
peak in average rate near 1.5 kHz. At intermediate
sound pressure levels (Fig. 2B, F), the peak near
1.5 kHz is lost in the high SR population
as
excitation begins to spread, and fibers with BFs
above and below 1.5 kHz begin to saturate. The
1.5 kHz tone is still represented in terms of average rate by a clear peak in both the low and
medium SR populations
of auditory nerve fibers
at 54 and 57 dB SPL (Fig. 2B, F). As sound
pressure level is further increased to 84 and 87 dB
SPL (Fig. 2C, D, G, H), the peak at 1.5 kHz is lost
in the medium SR population
as fibers with BFs
removed from 1.5 kHz begin to respond and are
driven into saturation at the highest levels. Nevertheless, at these high sound pressure levels a clear
peak around 1.5 kHz remains in the rate profile
for the low SR population.
Discussion
The results of the present study show that a 1.5
kHz tone can be represented in terms of average
discharge rate over a wide range of sound levels.
At high sound pressure levels where the driven
rates of the high SR auditory
nerve fibers are
saturated across a wide range of BFs, the low SR
population
maintains
a peak in the average rate
profile around 1.5 kHz. This peak at high sound
levels in the rate profile of the low SR auditory
nerve fibers reflects the higher thresholds (Liberman, 1978) and wider dynamic ranges (Evans and
Palmer, 1980; Sachs and Abbas, 1974; Schalk and
Sachs, 1980) of these fibers relative to those of the
high SR group.
In one case, at the highest level studied (Fig.
2H), the average rate of the low SR auditory nerve
fibers in the 3-6 kHz region approaches
that of
the high and medium SR fibers. However, in the
BF range of 3-6 kHz where the spread of excitation occurs, the mean thresholds for high SR fibers
is 6.2 dB SPL and is 17.1 dB SPL for low SR
fibers, and there are no fibers included in this
sample with thresholds greater than 20 dB above
the mean high SR thresholds (Fig. 3). On the other
hand, Liberman (1978) reports low SR fibers with
thresholds close to 40 dB above those of the high
SR group for normal cats. Liberman found such
high threshold
fibers with electrical stimulation
through the recording electrode; we did not search
for fibers with electrical stimulation.
Spread of
excitation in these very high threshold fibers would
be expected to be less than that observed in our
sample. Nevertheless, even in the population shown
in Fig. 2H, the discharge rate for low SR fibers
with BFs around 1.5 kHz is still higher than those
fibers with BFs between 3 and 6 kHz where the
spread of excitation
occurs. Even such a small
peak could be sharpened and amplified by inhibitory sideband mechanisms in the cochlear nucleus.
The suggestion
that frequency
is encoded in
average rate across the population of low SR fibers
at high levels requires that the responses of these
fibers be heavily weighted at high sound levels.
Such a weighting has been proposed and modeled
by Delgutte (1982) for responses to speech sounds.
The low SR fibers represent approximately
15% of
the auditory nerve population
(Liberman,
1978).
In this regard it is interesting that low and medium
SR fibers show more complex branching patterns
(Fekete et al., 1983) within the cochlear nucleus
and have a significantly
greater number of bouton
terminals and en passant swellings (Rouiller et al.,
1983) within the cochlear nucleus than do high SR
fibers. Furthermore,
Shofner and Young (1985)
have found that some non-spontaneous
units in
the cochlear nucleus characterized by chopper discharge patterns to tone bursts have BF rate-level
functions
similar to those of low SR auditory
nerve fibers, i.e. they have sloping saturation (Sachs
and Abbas, 1974; Schalk and Sachs, 1980). These
observations
suggest that the rate representation
present at high levels in low SR auditory nerve
fibers might be preserved in at least some subsvsterns in the cochlear nucleus.
In summary, we have shown that a rate-place
representation
for the frequency of a single low
frequency tone is maintained at high stimulus levels
by the low SR population.
This result and similar
ones for more complex
stimuli
(Barta,
1985;
Costalupes,
1985; Sachs and Young, 1979) emphasize the importance of considering average rate
codes as seriously as codes based on temporal
properties of discharge patterns (Young and Sachs,
1979).
95
Acknowledgements
We wish to thank Phyllis Taylor for the illustrations and Drs. E.D. Young and R.L. Winslow for
their critical reading of the manuscript.
This research was supported by Grant No. NS12112 from
NINCDS.
W.P.S. was supported by NRSA postdoctoral fellowship No. NS07270.
References
Barta, P. (1985): Testing stimulus encoding in the auditory
nerve. Doctoral
Dissertation,
Johns Hopkins
University,
Baltimore.
Costalupes, J.A. (1985): Representation
of tones in noise in the
responses of auditory nerve fibers in cats: .I. Comparison
with detection thresholds. J. Neurosci. (in press).
Delgutte, B. (1982): Some correlates of phonetic distinctions at
the level of the auditory nerve. In: The Representation
of
Speech in the Peripheral
Auditory
System, pp. 131-149.
Editors: R. Carlson and B. Granstrom.
Elsevier Biomedical
Press, Amsterdam.
Evans, E.F. (1981): The dynamic range problem:
place and
time coding at the level of cochlear nerve and nucleus. In:
Neuronal
Mechanisms
of Hearing, pp. 69-85. Editors: J.
Syka and L. Aitkin. Plenum Press, New York.
Evans, E.F. and Palmer, A.R. (1980): Relationship
between
dynamic range of cochlear nerve fibres and their spontaneous activity. Exp. Brain Res. 40, 115-118.
Fekete, D.M., Rouiller,
E.M., Liberman,
M.C. and Ryugo,
D.K. (1984): The central
projections
of intracellularly
labeled auditory nerve fibers in cats. J. Comp. Neurol. 229,
432-450.
Guinan, J.J. and Peake, W.T. (1967): Middle ear characteristics
of anesthetized cats. J. Acoust. Sot. Am. 41, 1237-1261.
Kim, D.O. and Molnar, C.E. (1979): A population
study of
cochlear nerve fibers: comparison of spatial distributions
of
average-rate
and phase-locking
measures of responses to
single tones. J. Neurophysiol.
42, 16-30.
Liberman,
M.C. (1978): Auditory-nerve
response from cats
raised in a low-noise chamber. J. Acoust. Sot. Am. 63,
442-455.
Liberman, M.C. (1982): Single-neuron
labeling in the cat auditory nerve. Science 216, 1239-1241.
Liberman,
M.C. and Oliver, M.E. (1984): Morphometry
of
intracellularly
labeled neurons of the auditory nerve: correlations with functional
properties.
J. Comp. Neurol. 223,
163-176.
Pfeiffer, R.R. and Kim, D.O. (1975): Cochlear
nerve fiber
responses:
Distribution
along the cochlear
partition.
J.
Acoust. Sot. Am. 58, 867-869.
Rouiller, E.M., Cronin Schreiber, R., Fekete, D.M. and Ryugo,
D.K. (1983): Morphology
of auditory nerve fiber innervation of the cat cochlear nucleus in relation to spontaneous
rate activity. Sot. Neurosci. Abs. 9, 495.
Sachs, M.B. and Abbas, P.J. (1974): Rate versus level functions
for auditory-nerve
fibers in cats: tone-burst
stimuli. J.
Acoust. Sot. Am. 56, 1835-1847.
Sachs, M.B. and Young, E.D. (1979): Encoding of steady-state
vowels in the auditory nerve: Representation
in terms of
discharge rate. J. Acoust. Sot. Am. 66, 470-479.
Schalk, T.B. and Sachs, M.B. (1980): Nonlinearities
in auditory-nerve fiber responses to bandlimited
noise. J. Acoust.
Sot. Am. 67. 903-913.
Shofner, W.P. and Young, E.D. (1985): Excitatory/inhibitory
response types in the cochlear nucleus: Relationships
to
discharge patterns and responses to electrical stimulation of
the auditory nerve. J. Neurophysiol.
54, 917-939.
Sokolich, W.G. (1977): Improved acoustic system for auditory
research. J. Acoust. Sot. Am. 62, S12.
Young,
E.D. and Sachs, M.B. (1979): Representation
of
steady-state
vowels in the temporal aspects of the discharge
patterns of populations
of auditory-nerve
fibers. J. Acoust.
Sot. Am. 66, 1381-1403.