Non-linearities and the representation of auditory spectra.

Publisher Summary This chapter describes the effects of nonlinearities on frequency analysis, using neurons in the dorsal cochlear nucleus (DCN) as an example. The physiological characteristics that make the principal cells of the DCN nonlinear are described, along with a weight–function model that allows description of nonlinearities in the DCN to second-order. The model is based on responses to noise stimuli with a wide variety of pseudorandom spectral shapes and explicitly separates frequency analysis from temporal response characteristics. The quality of the model is evaluated by its ability to predict the responses of DCN neurons to spectral shapes other than those used to derive the model. A second-order model performs well at this task in ventral cochlear nucleus (VCN) and comparatively weaker in DCN. Similar to fibers, VCN neurons are dominated by a bandpass filtering operation, with the difference of showing more inhibitory response. These properties suggest that DCN neurons behave like feature detectors as opposed to AN fibers and VCN neurons, which are generalized spectrum analyzers. Despite their nonlinearity, DCN neurons encode approximately the same amount of information about the stimulus spectrum that VCN neurons do.