Responses Of Auditory Nerve Fibers Research Articles

Effects of crossed-olivocochlear-bundle stimulation on cat auditory nerve fiber responses to tones.

Phase, synchronization index, and average firing rate were calculated from period histograms of tone burst responses obtained from sound level series with and without electrical stimulation of the cross-olivocochlear-bundle (COCB). For most fibers, at low sound levels, COCB stimulation shifted the rate and synchronization index level functions up in sound level but did not shift the phase-level function in the same way. These effects can be accounted for if the stage at which the COCB acts precedes the stage at which analog signals are changed into neural firing patterns with a given rate and synchronization index, but does not precede the stage at which the level dependence of phase is introduced. Some level series show an abrupt phase change and "dips" in rate and synchronization-index level functions at high sound levels. COCB stimulation shifted these abrupt phase changes and dips down in sound level and usually had little effect at sound levels above these abrupt phase changes and dips. The following explanatory hypothesis is developed: excitation of an auditory-nerve fiber is the result of two factors which are out of phase and have different growth functions. The two factors cancel when they are equal in amplitude producing the dips and phase changes. COCB stimulation reduces the more sensitive factor but does not change the other factor so the two factors cancel at a lower sound level with COCB stimulation.

Responses of auditory-nerve fibers to consonant-vowel syllables.

The discharge patterns elicited by a set of synthesized consonant-vowel (CV) syllables were studied in the auditory nerve of the cat. The syllables, heard as /ba/, /da/, or /ga/, included a 25-, 50-, or 75-ms formant transition followed by a segment of steady-state vowel. The data were analyzed in terms of average discharge rate and in terms of the synchrony of discharges with respect to various spectral components of the stimuli. The results differ slightly from those of previous reports of the responses to vowels [Sachs and Young, J. Acoust. Soc. Am. 66, 470-479 (1979); Young and Sachs, J. Acoust. Soc. Am. 66, 1381-1403 (1979)], in that average discharge rates appear to provide more information about the spectra of formant transitions than they do about the spectra of steady-state vowels. This difference reflects changes in the spectrum of the syllable produced by movements of the formants. The synchrony of discharges, however, may provide more detailed information about the spectra of CVs than does average discharge rate. Each fiber's response at a particular peristimulus time may be characterized by the "dominant response component," the largest peak in the Fourier transform of the period histogram. The trajectories of the first three formants can be inferred from changes in the "dominant components" in a sample of fibers.

Interaction phenomena in a model of mechanical to neural transduction in the ear

We describe a model for the transduction of the basilar membrane displacements into activity of the auditory nerve fibers. The model consits of two parts: a one-dimensional model of basilar membrane motions and a model of sensory receptors. The first part is a linear functional model which gives the electrical analog of the displacement patterns of the basilar membrane. The second part is a non-linear physical model which is mainly based on the assumption that information issued from the two kinds of hair cell receptors interacts electrically. The model is shown to reproduce the excitation patterns of single auditory nerve fibers in response to simple input signals (one and two-tone stimuli) or to inferred trapezoidal displacements of the basilar membrane. A second set of experiments is described to show how useful such a model can be for the understanding of complex sound coding. We present results for a speech-like single formant input sound paying special attention to the fundamental frequency representation in the discharge patterns of our simulated fibers.

Nonmonotonic behavior of cat ABR amplitude at high intensities

At low intensities (< 60 dB SPL), the amplitude of the auditory brain stem evoked response (ABR) is a linear function of signal intensity. Above 60–70 dB SPL response amplitudes decline, sometimes dramatically, and reach a minimum at 80–90 dB SPL. Above 80–90 Db SPL, ABR wave amplitudes increase rapidly. These features are represented in all ABR components with the effect being most pronounced in waves I and V. The nonmonotonic ABR amplitude behavior exists at high intensities (> 100 dB SPL) in kittens less than four postnatal weeks old, and is more pronounced in kittens less than three to four months of age than in adult animals. Also, nonmonotonic behavior disappears for some time following exposure to a high-intensity sound. Nonmonotonic growth of ABR amplitude is in many respects similar to that observed in responses of auditory nerve fibers. [Work supported by NINCDS grant NS-14880 and by a Biomedical Research Supplement Grant from NIH.]

Simulated nerve fiber activity in response to one‐ or two‐formant speech‐like sound

Representations of speech‐like sounds in discharge patterns of auditory nerve fibers were studied by mean of a model. At a previous meeting we presented the model of the peripheral auditory system developed in our laboratory. The model was shown to reproduce the excitation patterns of single auditory nerve fibers in response to one‐tone, two‐tone, trapezoidal, or impulse stimuli. We present new results obtained in response to one‐and two‐formant stimuli. Time patterns of simulated fibers are studied with Fourier analysis to elicit the frequency components involved in each response. Special care is taken to fundamental frequency representation. The main conclusion is that fundamental frequency can be present in all units regardless of the tuning properties cf the fibers. All the results obtained are compared with available experimental data. [Work supported by A.D.I. convention No 80/379.]

Population study of non‐monotonic response‐intensity curves in auditory nerve fibers

For pure‐tone signals, responses of many auditory nerve fibers in cats are nonmonotonic functions of intensity. In the experiments described here, we examined responses across the fiber population to tones at 200 Hz, 500 Hz, and 1000 Hz. Below 70–90 dB SPL, responses to 200 and 500 Hz generated unimodal period histograms. Above 70–90 dB SPL, period histograms were bimodal. This was not the result of an externally generated harmonic distortion product. At yet higher intensities, period histograms became unimodal, but the average discharge phase angle changed. Typically, the second peak which emerged in the period histograms at high levels dominated the temporal response and the original peak disappeared. Only unimodal period histograms were seen at 1000 Hz; however, discharge phase shifts still occurred at high intensities. The intensity corresponding to the decline of response amplitude was independent of fiber CF and fiber sensitivity at the stimulus frequency. These findings indicate that the mechanism producing the nonmonotonic input‐output functions is not confined to a restricted cochlear region. Further, cochlear mechanics is likely to be the source of the nonmonotonicities, perhaps producing standing waves in the cochlea.

Chronic effects of loop diuretics on the guinea-pig cochlea.

Groups of pigmented guinea pigs (weighing 480 to 650 g at the start of the experiment) were given daily intraperitoneal injections of 16 mg/kg bumetanide or 10, 20 or 40 mg/kg furosemide over a period of 4 weeks. Two weeks following the last diuretic dose the animals were anaesthetized with urethane and the responses of single auditory nerve fibres to pure-tone and click stimuli were recorded. The tuning of individual units, as measured by Q10 dB values of frequency threshold curves, tended to be more affected than threshold at characteristic frequency by both loop diuretics. Light microscope examination of treated cochleas suggested greater effect of the drugs on the organ of Corti than on stria vascularis.

Efferent desensitization of the responses of turtle single auditory nerve fibers

Extracellular recordings were made from single auditory afferents and frequency-threshold curves determined in the isolated half-head of Pseudemys scripta. Constant current shocks, 20–500 μA, in trains at 250 s1, delivered at the junction between the anterior and posterior roots of the statoacoustic nerve, resulted in a prolonged elevation of the thresholds to pure tones, and a broadening of the frequency-threshold curves of the auditory afferents. As an example, in one sharply tuned afferent, a threshold elevation of 41 dB at the characteristic frequency (CF, 700 Hz) was accompanied by a reduction in Q10 dB from 4.7 to 0.9. The extent of desensitization increased steeply with number of shocks, and a train of seven shocks could elevate the threshold at the CF by 80 dB. The desensitization could be demonstrated in a majority of recordings, and observed in the absence of antidromic firing of the afferent. We suggest that a component of this efferent effect may result from a detuning of the hair cell's electrical filter [A. C. Crawford and R. Fettiplace, J. Physiol. 312, 377–412 (1981)].

Frequency analysis in the peripheral auditory system of anurans

The two auditory organs—amphibian and basilar papillae—in an anuran's inner ear are anatomically distinct and both lack basilar membranes. Of fundamental interest is whether these organs nevertheless are capable of a systematic frequency analysis (i.e., tonotopic organization) and how the response properties of their auditory nerve fibers compare with those of other vertebrates. Anatomical and electrophysiological studies in our laboratories during the past several years have focused on these issues. Each auditory fiber in the anuran's eighth nerve is characterized by a “V”-shaped tuning curve which is fairly selective about its own best excitatory frequency (BEF). In primitive frogs, the fibers fall into two disjoint populations based on their BEF's (low-frequency sensitive and high-frequency sensitive) and susceptibility to two-tone suppression, with each population derived from a separate papillae. In more advanced species a third population arises, associated with the amphibian papilla and sensitive to the mid-frequency range. Electrophysiological studies with paired clicks and pure tones predicted that a place mechanism is operative in the amphibian papilla but not in the basilar papilla. The implication of a place principle in the amphibian papilla was further reinforced by the presence of an extension of this organ's sensor macula in advanced species. Recent Lucifer yellow dye injections in single auditory fibers confirm the tonotopy in the amphibian papilla which is not present in the basilar papilla. These combined anatomical and electrophysiological studies are currently aimed at elucidating further correlations between hair cell locus, morphology, and auditory nerve fiber response properties.

Auditory-nerve fiber responses to tones in a noise masker

The phase-locked responses of single auditory-nerve fibers were measured for a continuous tonal stimulus presented in a noise background. The response amplitude, the primary Fourier component of the period histogram, was found to be dependent on the relative levels of the noise and tone. Different noise-to-tone level ration resulted in quite different response amplitude; changing overall level keeping noise-to-tone constant (constant dB difference) provided little change in response. With transient stimuli, phase-locked response to the tone at noise onset was consistently greater than to a tone presented during the steady-state noise exposure. When responses were normalized to the average rate, the difference between onset and steady-state responses were not clearcut.

Identification of local and propagating distortion products from cochlear microphonic responses.

Careful measurements show that sound pressures of 40 dB re 20 mu N/m2 are sufficient to reveal two varieties of well-behaved, nonlinear distortion products in the cochlear microphonic (CM). The first variety appears in the primary-tone area and is designated as a local CM distortion product. The second type exists apical to the primary-tone area and is identical to the mechanically propagated distortion seen in the phase-locked responses of primary auditory nerve fibers. The existence of the propagating distortion product forces the conclusion that there must be a mechanical contribution to the local "CM" distortion product as well. The intrusion of nonlinear mechanical responses at such low levels (less than 40 dB SPL) indicates that the effective mechanical input to the hair cells may be nonlinear over most of the audible range. An important but unanswered question is the range over which the transducer characteristic of CM could be effectively linear, for that would determine whether CM could be used to probe mechanical nonlinear effects in the primary tone area.

Responses of auditory-nerve fibers to characteristic-frequency tones and low-frequency suppressors

Responses of single auditory-nerve fibers to combinations of low-frequency (suppressor) and characteristic frequency (CF) tones were recorded. The shapes of period histograms were observed as the level of the suppressor was raised. For fibers with CFs above 5.0 kHz, at the lowest suppressor levels used, the phase of the suppressor tone which caused instantaneous rate to increase when presented alone caused a rate decrease when added to a CF tone. For suppressor frequencies of 500 and 100 Hz, as suppressor level is increased, two peaks appear in the period histograms. Further increase in level causes first one, then the other of these peaks to disappear and the histogram takes on the shape of the histogram to the suppressor alone. A similar progression of histogram shape follows a decrease in CF tone level for fixed suppressor level.

Experiments on the origin of two‐tone suppression

Two‐tone suppression has been shown to exist in virtually every response of cochlear origin. For comparable stimulus conditions, however, suppression magnitude measured in responses of hair cells and auditory nerve fibers is considerably greater than suppression magnitude measured in basilar membrane displacement patterns, and tho mechanical suppression threshold appears to be higher than the auditory nerve‐fiber suppression threshold. We recorded responses of single auditory nerve fibers in anesthetized cats and used discharge synchronization as an index of suppression magnitude. It was found that (a) when “notches” in input‐output rate functions exist, suppressed responses cannot be made to occur to the right of the upper leg of the unsuppressed function; (b) unlike results obtained for cochlear microphonics, the discharge phase angle of a suppressed response in auditory nerve fibers is the same as the angle of the pure‐tone response at the same intensity; and (c) for low‐frequency suppressors, the rate of growth of suppression with increasing suppressor‐tone intensity can easily exceed the presumed rate of growth of basilar membrane displacement. A possible explanation is that there are two generators for suppression. One generator operates at high intensities and originates in basilar membrane mechanics. The other generator operates at all suprathreshold intensities and originates either in the micromechanics of the subtectorial space or in the mechanical‐to‐neural transduction process occurring inside hair cells. [Work supported by NIH.]

High‐intensity “notches” in responses of auditory nerve fibers

In cat auditory nerve fibers, pure‐tone input‐output functions for both discharge rate and synchronization can be nonmonotonic. That is, there can exist an intensity range over which the response saturates, then declines, then increases again at higher intensities. In our experiments, this intensity “notch” usually began at 70–100 dB SPL and persisted for the next 10–50 dB, with 30 dB being a typical range. The upper leg of the rate‐intensity function had a steeper slope and saturated at a higher rate than the lower leg. The depth of the “notch” could vary from only a slight inflection in response to a depression down to spontaneous level. We have observed “notches” in all regions of fibers' response areas. The intensity range and magnitude of the “notch” varied both within and across fibers. “Notches” at low stimulus frequencies were usually accompanied by rapid shifts in mean discharge phase angle. The cumulative phase shift was not consistent, ranging from less than 90° to more than 200°. Reversals in the direction of phase change were common. “Notches” for tones below 600 Hz often displayed bimodal period histograms, but the two peaks usually existed only over a narrow range of intensities. [Work supported by NIH.]

Simulated auditory nerve fiber responses to speech signal

Alterations imposed to speech signal by the peripheral auditory system, have been studied by mean of the physical model described by J. M. Dolmazon and L. Bastet [J. Acoust. Soc. Am. 65, S14 (1979)]. This model can reproduce for any acoustical stimulus the mean discharge frequency of an auditory nerve fiber after a linear simulation of basilar membrane motion and a nonlinear simulation of mechanical to neural transduction for taking into account two‐tone suppression effects or combination tone generation. We present a comparison between features extracted, from the input speech signal, by the model and by conventional analysis techniques to estimate the influence of the well‐known nonlinear effects in speech perception. Special care has been taken for vowel analysis to clarify formant interaction.

Auditory-nerve fiber responses to frequency-modulated tones

Discharge patterns of cat auditory-nerve fibers were obtained in response to frequency-modulated (FM) tones. The rate and direction of frequency change and the sound-pressure level of the sweep tones were systematically varied, and aspects of the discharge patterns were compared to aspects of the discharge patterns elicited by pure tones. Increases in SPL broaden the frequency range over which the fiber responds, as is the case with pure-tone stimuli. Increases in the rate of frequency change have little effect on frequency selectivity for the rates tested. In general, the pure-tone response area is a good predictor of the response area to FM. Although approximately equal numbers of spikes are elicited by ascending and descending sweeps, the discharge patterns differ slightly; for each direction of frequency change, the FM response area is shifted in the direction of the earliest-occurring frequencies. Most of this shift can be accounted for by neural adaptation. This asymmetry is small, relative to those observed in the central nervous system.

Differences in the CF dependence of response phase of gerbil and guinea pig auditory nerve fibers

Time‐locked responses of auditory nerve fibers to acoustic sinusoids having frequencies between 100 and 400 Hz were studied in gerbils and guinea pigs. With the exception of the responses of gerbil fibers with CFs above 10 kHz, which produce double‐peaked histograms, the responses of most fibers in both species produce single‐peaked histograms when the modulation of spontaneous activity is less than 100%. Within this restricted modulation range, the preferred phase of response is relatively independent of stimulus amplitude. Nonetheless, the variation of the response phase with CF is fundamentally different in the two species. Whereas gerbil fibers show an abrupt 180° phase transition in the CF range between 2 and 4 kHz, the response phase of guinea pig fibers is CF‐independent for stimulus frequencies that are at least 12 octave below fiber CF. The finding of different phase functions in the two species indicates that caution should be exercised in basing models of the neural excitation process in the cochlea solely on data obtained from one mammalian species.

Relationships between trapezoidal EP variations and auditory nerve fiber responses in gerbils and guinea pigs

It has been shown [Sokolich et al., J. Acoust. Soc. Am., 59, 963–974 (1976)] that most gerbil auditory nerve fibers show one of three time patterns of response to acoustically‐evoked trapezoidal time patterns of EP change. Pattern “A” is characteristic of fibers with CFs below 2 kHz, pattern “B”, of fibers with CFs between 3 and 6 kHz, and pattern “C”, of fibers with CFs above 8 kHz. Although all three patterns contain four distinct response components reflecting both static and dynamic aspects of trapezoidal EP change, the patterns differ in that the sign of one or more components depends on fiber CF. The present comparative study reveals that pattern “A” is exhibited by all guinea pig fibers with CFs below and by many fibers with CFs above 7 kHz. Nonetheless, some guinea pig fibers with CFs above 7 kHz show pattern “B”. In spite of the difference in the dependence of response pattern on CF, all patterns observed in both species support the conclusion that neither excitation nor inhibition of auditory nerve fibers is uniquely related to specific aspects of either static or dynamic changes in EP or to associated aspects of basilar membrane motion.

The relationship between spike rate and synchrony in responses of auditory-nerve fibers to single tones.

Dischargae patterns of single auditory-nerve fibers in the cat responding to single tones were recorded and analyzed. Synchronization of the discharges to the stimulus tone was measured for frequencies below 6-kHz. The synchronized response and the average discharge rate each varied with stimulus level over 20--30-dB ranges. However, these ranges overlapped only a 10-dB range. Under certain conditions the average rate of response varied without accompanying variation in the synchronized response. The physiological implications of these observations are discussed.

Representation of speech-like sounds in the discharge patterns of auditory-nerve fibers.

It is now possible to study how acoustic characteristics important for speech discrimination are represented in the discharge patterns of auditory-nerve fibers. The electrical activity of single auditory-nerve fibers in response to speech-like sounds (single and two formant synthetic stimuli and "fricative" noise bursts) was recorded with microelectrodes in anesthetized cats. Results demonstrate that a conceptualization of some basic properties of responses to simple acoustic stimuli is useful in interpreting qualitatively how certain characteristics of speech-like sounds can be coded. Specific examples are given for (1) the rapid changes in amplitude and spectrum that occur for each syllable, (2) the fundamental frequency of voiced sounds, and (3) the spectral envelope of fricative consonants. It is also shown that the presence of background noise can have qualitatively different effects on how various characteristics of speech-like sounds are represented.