Auditory nerve representation of vowels in background noise

Abstract

Responses of auditory nerve fibers to steady-state vowels presented alone and in the presence of background noise were obtained from anesthetized cats. Representation of vowels based on average discharge rate and representation based primarily on phase-locked properties of responses are considered. Profiles of average discharge rate versus characteristic frequency (CF) ("rate-place" representation) can show peaks of discharge rate in the vicinity of formant frequencies when vowels are presented alone. These profiles change drastically in the presence of background noise, however. At moderate vowel and noise levels and signal/noise ratios of +9 dB, there are not peaks of rate near the second and third formant frequencies. In fact, because of two-tone suppression, rate to vowels plus noise is less than rate to noise alone for fibers with CFs above the first formant. Rate profiles measured over 5-ms intervals near stimulus onset show clear formant-related peaks at higher sound levels than do profiles measured over intervals later in the stimulus (i.e., in the steady state). However, in background noise, rate profiles at onset are similar to those in the steady state. Specifically, for fibers with CFs above the first formant, response rates to the noise are suppressed by the addition of the vowel at both vowel onset and steady state. When rate profiles are plotted for low spontaneous rate fibers, formant-related peaks appear at stimulus levels higher than those at which peaks disappear for high spontaneous fibers. In the presence of background noise, however, the low spontaneous fibers do not preserve formant peaks better than do the high spontaneous fibers. In fact, the suppression of noise-evoked rate mentioned above is greater for the low spontaneous fibers than for high. Representations that reflect phase-locked properties as well as discharge rate ("temporal-place" representations) are much less affected by background noise. We have used synchronized discharge rate averaged over fibers with CFs near (+/- 0.25 octave) a stimulus component as a measure of the population temporal response to that component. Plots of this average localized synchronized rate (ALSR) versus frequency show clear first and second formant peaks at all vowel and noise levels used. Except at the highest level (vowel at 85 dB sound pressure level (SPL), signal/noise = +9 dB), there is also a clear third formant peak. At signal-to-noise ratios where there are no second formant peaks in rate profiles, human observers are able to discriminate second formant shifts of less than 112 Hz. ALSR plots show clear second formant peaks at these signal/noise ratios.

Predictions of formant-frequency discrimination in noise based on model auditory-nerve responses

“…Although low-spontaneous-rate (LSR) fibers may provide a wider dynamic range based on rate, it is clear from our results that there is ample information in the timing of small populations of HSR fibers to explain the behavioral results. The importance of the temporal responses of small populations of fibers tuned near the frequency of interest is consistent with previous studies Sachs et al, 1983;Delgutte and Kiang, 1984a, b;Miller et al, 1987;Tan and Carney, 2005).…”

“…However, intense noise maskers and the relatively high levels of harmonics near formant peaks tend to dominate the rates of AN fibers, reducing the sensitivity of rate to small changes in the levels of neighboring harmonics. Hienz et al (1996) concluded that behavioral results in cat for vowel discrimination in the presence of high-level maskers could not be explained by the rates of high-spontaneous-rate (HSR) AN fibers, which are severely degraded in this stimulus condition Delgutte and Kiang, 1984b;Sachs et al, 1983). Although low-spontaneous-rate (LSR) fibers may provide a wider dynamic range based on rate, it is clear from our results that there is ample information in the timing of small populations of HSR fibers to explain the behavioral results.…”

Effects of Signal Level and Background Noise on Spectral Representations in the Auditory Nerve of the Domestic Cat

“…12). Previous studies of vowel representations in background noise suggest that the noise response is suppressed by the HRTF signal (Sachs et al 1983). The effect is exacerbated when a fiber is tuned to a spectral notch in the HRTF because the sharp spectral contrast minimizes energy at BF and maximizes energy in surrounding FIG.…”

“…Consequently, formant frequencies are revealed by the BFs of fibers displaying relatively high discharge rates (Le Prell et al 1996;Sachs and Young 1979). The quality of this distributed representation depends on fiber type, sound level, and the presence of background noise (Delgutte and Kiang 1984;May et al 1998;Sachs et al 1983).…”

Midbrain Synchrony to Envelope Structure Supports Behavioral Sensitivity to Single-Formant Vowel-Like Sounds in Noise

“…While auditory nerve fibers show differences in average rate across vowels at low sound levels that could potentially support behavioral discrimination (Sachs and Young 1980), this average-rate representation deteriorates at moderate-to-high sound levels and in noise due to rate saturation. In contrast, response synchrony to individual frequency components (Young and Sachs 1979;Sachs et al 1983;Delgutte and Kiang 1984) and to F0-related envelope structure ) provides a robust code for vowel discrimination across sound levels and in noise. Synchrony-based coding of vowels might be less effective in central nuclei, given lower-frequency synchronization limits compared to the auditory nerve (Joris et al 2004).…”