Perceptual and modeling estimates of frequency selectivity suggest that acoustic stimulation reduces cochlear gain

Skyler G Jennings, Purdue University

Abstract

Auditory perception covers a wide range of intensities. This range is called the “dynamic range,” and spans roughly 120 dB. A fundamental question in auditory science relates to how the perceptual dynamic range is coded in the auditory system given that the dynamic range of individual neurons is about 20-40 dB. The medial olivocochlear (MOC) reflex may be one mechanism contributing to the perceptual dynamic range. This reflex acts as an automatic gain control by adjusting the gain of the cochlear amplifier. A perceptual technique thought to be sensitive to cochlear gain has been used to study the MOC reflex. This technique is referred to as the “gain reduction” technique. A complementary approach to studying the MOC reflex involves computational modeling where perceptual thresholds are predicted from a physiologically realistic model of the auditory periphery. The experiments described in this document involve a multifaceted approach of perceptual and computational modeling techniques to study how a reduction in gain (hypothetically via the MOC reflex) influences detection in an auditory masking task. In the perceptual studies, a standard masking condition was compared with a condition where a precursor precedes the masker (“precursor condition”). This precursor was assumed to reduce cochlear gain. The standard and precursor conditions were compared using various psychophysical techniques sensitive to cochlear frequency selectivity and gain. These techniques included psychophysical tuning curves, notched noise tuning characteristics and growth of masking functions (Chapters 2 and 3). The psychophysical experiments motivated the computational modeling simulations (Chapter 4). The modeling approach involved combining two established models; one related to the auditory periphery and the other related to masking. Model predictions were obtained for two hypotheses regarding the effect of the precursor on masking threshold. The additivity of masking hypothesis assumed that the masking effect of the precursor and masker add at the output of the cochlea. Implicit in this hypothesis is the notion that cochlear gain remains fixed over time. The gain reduction hypothesis assumed the precursor reduces the gain of the cochlea, resulting in an adapted cochlear input/ output function. The results support the gain reduction hypothesis and suggest that frequency selectivity is a dynamic process. Implications for this finding suggest that psychophysical techniques and previous psychophysical data may need to be reinterpreted by considering the potential effects of the MOC reflex. In addition, the results suggest that using a short masker with a short signal-masker delay may provide a way to avoid MOC-related effects. Finally, these results are consistent with the hypothesis that the MOC reflex may be responsible for improving speech understanding in noise. This improvement may come from increased contrast within the speech stimulus (i.e. improved envelope coding) and between the speech and the noise (i.e. improved signal-to-noise ratio).

Degree

Ph.D.

Advisors

Strickland, Purdue University.

Subject Area

Audiology

COinS