Spectro-Temporal Response Fields in Primary Auditory Cortex (AI)
Thursday, March 11, 1999 - 4:00pm - 5:15pm
Shihab Shamma (University of Maryland), Jonathan Simon (University of Maryland)
Spectra of complex sounds such as speech and music are broadband and dynamic. To understand the encoding of these signals in the primary auditory cortex (AI), we characterize the responses of units in AI to dynamically modulated noise, producing the Spectro-Temporal Response Field (STRF). The STRF provides both a quantitative and intuitive measure of how a unit encodes dyanimical sound features into its resultant spike train. The STRF is a linear response characteristic, but for a reasonably broad class of sounds, the STRF predicts the response of units to novel stimuli. Non-linear corrections further improve the predictions. Constraints on the structure of STRFs, such as whether they are separable (the vast majority appear to be quadrant-separable), are also constraints on the connectivity of the neuronal inputs.