The proposed research aims to gain a comprehensive understanding of the motor mechanisms involved in the generation of complex sounds. Particular emphasis is placed on the most widely used animal model system for learned vocal behavior, the zebra finch, whose song is characterized by a broad range of acoustic features from tonal sounds to complex spectral content. A combination and interaction of theoretical and experimental approaches is used to unravel the peripheral mechanisms of generation of this range of sound characteristics. The interactive approach involves recording of physiological data (electromyograms and air sac pressure) to drive the models and syringeal muscle stimulation experiments to test predictions of the modeling work. Based on these refinements of the computational approach, a prototype of an electronic syrinx, implementing the differential equations of the models, will be modified to reproduce zebra finch song. This electronic syrinx will be controlled by the neural instructions for song production (muscle activation patterns and respiratory pressure), which can be monitored in the singing bird and used for on-line generation of sound. This subject-controlled vocal prosthesis is then used to generate acoustic output in muted birds, with the potential for experimentally manipulating song characteristics, to test which acoustic features are required as information from auditory feedback for maintenance of song.
This research will contribute to our knowledge of the interplay between neural control, peripheral systems and acoustic behavior, which will enhance our understanding of speech production in humans. The electronic syrinx will provide an excellent test device for developing a similar system for the human larynx.
Showing the most recent 10 out of 49 publications
