Linear and Nonlinear Coding of Sound Spectra by Discharge Rate in Neurons Comprising the Ascending Pathway Through the Lateral Superior Olive

Abstract

Sound source location is computed at central levels in the auditory ­system based on the neural representations of the spectral and temporal characteristics of the sounds arriving at the ears. Interaural level differences, an acoustical cue to location, are first encoded by neurons in the lateral superior olive (LSO). The sound spectra at the left and right ears must be encoded accurately by the afferents to the LSO in order for the LSO neurons to extract the ILD. Here, we use a systems approach to examine spectral coding capabilities of the afferent inputs to the LSO - globular (GBC) and spherical (SBC) bushy cells of the cochlear nucleus and the medial nucleus of the trapezoid body (MNTB) - and LSO neurons themselves. Extracellular recordings were made in the trapezoid body and were classified according to their responses to short tone bursts: primary-like (PL), PL-with-notch (PLN) and chopper, which correspond to SBC, GBC, and stellate cell types, respectively. The Random Spectral Shape (RSS) method was used to estimate how spectral level is weighted by neurons, both linearly and non- linearly, across a wide band of frequencies. The first (linear) and second (non-linear) order spectral weighting functions were measured for 22 PL, 21 PLN, 16 chopper, 42 MNTB, and 15 LSO neurons. The validity of the estimated weighting functions was tested for each neuron by predicting the rate responses to arbitrary RSS stimuli and head related transfer function-filtered broadband noise. The fraction of explained variance, ranging from 1 (perfect fit) to 0 or less (poor fit), increased significantly for all neuron types by including the second order, non-linear terms. The results demonstrate that the neural pathways through the LSO can accurately encode sound spectra based on discharge rate. The psychophysical implications (sound localization, spectral shape discrimination, etc.) of these findings will be discussed.