The project seeks to determine the neural mechanisms for encoding species-typical communication sounds and the neural correlates of auditory scene analysis in primates. The squirrel monkey is used because of its highly stereotyped repertoire of calls. Computer generated model calls, natural calls, and their components are used as auditory stimuli to test the responses of neurons in the medial geniculate nucleus (MGN) to frequency ratios and time intervals among call components. Neural sensitivity to acoustic transients was tested with noise bursts. Previous annual reports described the discovery in the MGN of combination-sensitive neurons that respond selectively to particular calls out of a set of acoustically related calls. These combination-sensitive neurons resemble those described for lower vertebrates, indicating that these neurons may be of general importance in processing species-typical communication sounds. Call selectivity was found to depend on a sensitivity to time delays between call components. This delay sensitivity is a mechanism by which spectrally similar calls may be discriminated. Similar critical delays are known to be important in phoneme identification by humans. Thus, our results may provide insight on the neural mechanisms of human speech perception. In the past year, the focus has been on the amplitude-rate sensitivity of MGN neurons. The majority of MGN neurons were found to be sensitive to the rate of rise of linear amplitude modulation, whether presented at stimulus onset or delayed from stimulus onset. Rate of rise affords a complete description of neural response probability to broadband noise. The sensitivity of these neurons to rate of rise closely matches the performance of human listeners in a auditory segmentation task. The selectivity of these neurons for auditory events, based on rate of rise, is analogous to the selectivity of visual neurons for edges, based on visual contrast. These results are expected to be important to our understanding of hearing disorders of central etiology.
