Development and dissemination of computational models for physiology and psychophysical predictions

Abstract

This talk will present lessons from our experience in publishing and sharing computational models for physiological and psychophysical responses. Our physiological models have included rather comprehensive and nonlinear phenomenological descriptions of auditory-nerve responses to complex sounds, and simpler linear models for brainstem and midbrain responses. Using ensembles of single-neuron models to estimate population responses enables psychophysical predictions based on different aspects of sub-cortical representations. Examples of psychophysical models that we have pursued include level discrimination and diotic and dichotic masked detection of tones. Some of the challenges inherent in this type of work will be discussed. [Work supported by NIH-R01-001641 & -010813.]