Implementing an efferent signal into an auditory pathway model for tone-in-noise detection

Abstract

Simultaneous tone-in-noise detection has been studied extensively, but typically without consideration of the medial olivocochlear (MOC) efferents. We are testing hypotheses for masked detection using a central auditory model with a signal from midbrain to MOC. Masked tones are encoded in the rate profile of band-enhanced (BE) inferior colliculus (IC) neurons, which are excited by a range of modulation frequencies. Peripheral responses to noise are characterized by large fluctuations, an effective stimulus for BE IC neurons. In contrast, peripheral channels tuned near the tone have smaller fluctuations: addition of the tone flattens the signal envelope and also pushes the inner hair cell (IHC) transduction nonlinearity further into saturation. Excitatory projections to MOC from noise-driven BE IC cells would decrease cochlear gain, reducing IHC saturation, and resulting in larger fluctuations and IC rates. In contrast, tone-plus-noise-driven channels would reduce MOC excitation, resulting in relatively higher cochlear gain, more saturation, and ultimately lower IC rates. Thus, the descending signal from IC BE cells to MOC is hypothesized to enhance contrast in the IC rate profile. Because efferents have slow dynamics, timing is an important factor. Therefore, we focus on model sensitivity to masked tones of different durations for comparison to psychophysical trends.