Behavioral and neural selectivity for acoustic signatures of vocalizations

Vocal communication relies on the ability of listeners to identify, process, and respond to vocal sounds produced by others in complex environments. In order to accurately recognize these signals, animals’ auditory systems must robustly represent acoustic features that distinguish vocal sounds from other environmental sounds. In this dissertation, I describe experiments combining acoustic, behavioral, and neurophysiological approaches to identify behaviorally relevant vocalization features and understand how they are represented in the brain. First, I show that vocal responses to communication sounds in songbirds depend on the presence of specific spectral signatures of vocalizations. Second, I identify an anatomically localized neural population in the auditory cortex that shows selective responses for behaviorally relevant sounds. Third, I show that these neurons’ spectral selectivity is robust to acoustic context, indicating that they could function as spectral signature detectors in a variety of listening conditions. Last, I deconstruct neural selectivity for behaviorally relevant sounds and show that it is driven by a sensitivity to deep fluctuations in power along the sound frequency spectrum. Together, these results show that the processing of behaviorally relevant spectral features engages a specialized neural population in the auditory cortex, and elucidate an acoustic driver of vocalization selectivity.