INVESTIGADORES
MINDLIN Bernardo Gabriel
congresos y reuniones científicas
Título:
Using a song production model to study tuning properties of selective neurons in zebra finches
Autor/es:
A. AMADOR; SANZ PERL, YONATAN; G. B. MINDLIN; MARGOLIASH DANIEL
Lugar:
washington dc
Reunión:
Congreso; Neuroscience 2011 - Meeting of the society for neuroscience; 2011
Institución organizadora:
Society for Neuroscience
Resumen:
Neurons in the birdsong system exhibit strong selective responses to acoustic broadcast of the bird's own song (BOS), exhibiting stronger responses to BOS than tones, noises, and even conspecific songs or slightly modified BOS. These responses have been intensively studied as a window into, and part of the mechanism of, sensorimotor vocal learning. BOS responses are strongest in sleeping birds, emerge early in sensorimotor learning, and their prevalence may vary with species-specific patterns of learning. Despite their potential importance, the extreme response selectivity of song system neurons have made them difficult to study with traditional sensory physiological approaches.  In species with narrowband songs such as white-crown sparrow it has been possible to effectively reproduce and alter the songs by manipulating the zero-crossings, and broadband songs in species such as zebra finches have been deconstructed and manipulated with a multiple narrowband filtering approach. Song system neurons are sensitive to such manipulations, but it remains completely unknown whether such manipulations are relevant to the internal representations birds use to produce, and perceive, their song vocalizations.  To address this fundamental limitation and explore song system neuronal tuning properties, we have worked with a low dimensional model for zebra finch song production that includes a description of the sound source and vocal tract where some mathematical parameters can be linked to physiological properties observed during singing. Model output is a synthetic song. We propose the hypothesis that changes in parameters in the model correspond to changes in motor control parameters birds actually use to control song output. To date we have seen that complete models elicit neuronal responses in the HVC (a sensorimotor nucleus) strikingly similar to BOS responses, eliciting the same phasic-tonic features and somewhat lower magnitude of response. Progressively including the oropharyngeal cavity into the model, by changing its dissipation, allowing it to progressively include its filtering influence into the sound, or increasing the intrinsic noise in the activity of the syringeal muscles results in systematic increase of response magnitude but not a change in phasic/tonic activity patterns. These results demonstrate that a low dimensional model representing an approximation of peripheral mechanics is sufficient to capture behaviorally relevant features of song.