Previous work suggests that due to saturation of auditory nerve discharges, coding schemes based solely on average rate may not be adequate for the representation of complex stimuli, whereas schemes based on both neural phase-locking and place coding provide exquisite representations of these stimuli. The following fundamental questions arise from these studies. 1) How are the phase-locked discharges of auditory nerve fibers used by the CNS for the internal representation of acoustic stimuli? 2) Do strategies for the estimation of stimulus components change for narrow and broadband stimuli? 3) What is the role of cochlear filtering in these representations? The proposed research addresses these questions by determining how the statistics of neural response determines the coding of spectral information in the auditory nerve. A statistical description of the phase-locked responses of auditory-nerve fibers to speech-like stimuli will be generated. The source of neural data are entire populations of auditory nerve fibers recorded in the cat, measured in response to tones in noise and consonant-vowel syllables. The statistical descriptions involve the sample mean and variances of temporal histograms as well as frequency domain coefficients derived from these histograms. From these statistics, optimum representations of stimulus features will be determined and compared to average rate and temporal coding schemes. The function of cochlear filtering in temporal representations of broadband stimuli will be explored. Objectives of this study are to evaluate the role of bandpass filtering and synchrony suppression in limiting the statistical variation of stimulus components coded on the auditory-nerve channel. This work will provide a conceptual framework for the prediction of performance limits for the discrimination of narrow and braodband acoustic stimuli. Improvements in the understanding of the fundamental coding of these stimuli is obviously essential for the design of speech coding algorithms and cochlear prostheses. It is also expected that the mathematical modeling results which are developed for the removal of neural refractoriness from post-stimulus-time histograms will be improtant for studies which formulate physical models of intracochlear mechanisms based on neural discharges.

Agency
National Institute of Health (NIH)
Institute
National Institute of Neurological Disorders and Stroke (NINDS)
Type
Research Project (R01)
Project #
5R01NS023007-02
Application #
3405941
Study Section
Hearing Research Study Section (HAR)
Project Start
1985-12-01
Project End
1988-11-30
Budget Start
1986-12-01
Budget End
1987-11-30
Support Year
2
Fiscal Year
1987
Total Cost
Indirect Cost
Name
Washington University
Department
Type
Schools of Medicine
DUNS #
062761671
City
Saint Louis
State
MO
Country
United States
Zip Code
63130