A neural network approach to auditory streaming

Abstract

This paper suggests a neural network architecture for the streaming ofan acoustic input using fundamental grouping cues of the mammalianauditory system. The auditory periphery is simulated using biologi-cally relevant models, and the primary auditory cortex is representedby a neural network layer, in which three dierent features of the inputare mapped and streamed. Finally, an attentional searchlight that isdriven by the cortical network neurons, focuses on a selected corticalarea representing a listener who focuses their attention on a distinctacoustic input.