Multiple sounds degrade the frequency representation in monkey inferior colliculus

Abstract

AbstractHow we distinguish multiple simultaneous stimuli is uncertain, particularly given that such stimuli sometimes recruit largely overlapping populations of neurons. One commonly proposed hypothesis is that the sharpness of tuning curves might change to limit the number of stimuli driving any given neuron when multiple stimuli are present. To test this hypothesis, we recorded the activity of neurons in the inferior colliculus while monkeys made saccades to either one or two simultaneous sounds differing in frequency and spatial location. Although monkeys easily distinguished simultaneous sounds (∼90% correct performance), the frequency selectivity of inferior colliculus neurons on dual sound trials did not improve in any obvious way. Frequency selectivity was degraded on dual sound trials compared to single sound trials: neural response functions broadened, and frequency accounted for less of the variance in firing rate. These changes in neural firing led a maximum-likelihood decoder to perform worse on dual sound trials than on single sound trials. These results fail to support the hypothesis that changes in frequency response functions serve to reduce the overlap in the representation of simultaneous sounds. Instead, these results suggest that alternative possibilities, such as recent evidence of alternations in firing rate between the rates corresponding to each of the two stimuli, offer a more promising approach.Graphic AbstractHow sensory representations encode multiple stimuli despite coarse coding is unknown. Using a maximum likelihood decoder operating on the spike count response patterns of monkey inferior colliculus neurons, we show a marked reduction in decoding accuracy when two sounds are presented compared to one. The decoding was inferior to the behavioral performance of the animals, and thus suggests the presence of alternative coding strategies.