CONICET | Buscador de Institutos y Recursos Humanos

Songbirdsare a widely studied example of vocal learning that allows integrating neuraland peripheral recordings with a precisely quantifiable behavior. Althoughneural activity has been related to song acoustics in auditory playbackexperiments, it remains unresolved whether neural activity is related to songspectral structure during singing. To address this issue, we worked with aminimal physical model for birdsong production, having as an output a syntheticsong. Each syllable was coded in terms of parameters related to air sacpressure and tension of the syringeal labia, defining motor gestures. The modelfor birdsong production allowed making predictions about motor control that wevalidated developing a novel technique for dynamic control of subsyringealpressure in zebra finches. To further validate this model, we assessedresponses of the premotor forebrain nucleus HVC to song playback in sleepingbirds, as HVC neurons exhibit highly selective responses to the bird´s own song(BOS). Remarkably, the mathematical model was able to elicit responsesstrikingly similar to those for BOS. The analysis of the neural activity interms of the model for song production allowed proposing a novel model formotor control in songbirds. We confirmed this result with HVC recordings insinging birds, showing the pertinence of working with a low dimensional modelthat represents an approximation of peripheral mechanics but captures relevantfeatures of behavior.