ORAL DEFENSE OF DOCTORAL DISSERTATION Tatsuo Okubo B.S., University of Tokyo, 2006 M.S., University of Tokyo, 2008 Neural Mechanisms Underlying the Emergence of Rhythmic and Stereotyped Vocalizations in Juvenile Songbirds Monday, September 21, 2015 2:30 PM 46-3189 Thesis Supervisor Michale Fee https://stellar.mit.edu/S/project/graduatethesis/courseMaterial/topics/topic10/readings/Okubo_thesis/Okubo_thesis.pdf Abstract: Complex motor behaviors in humans, such as speech and athletic performances, are not innate, but instead learned through a trial-and-error process. What is the neural mechanism that allows the brain to learn such behaviors? Here, I use songbird as a model to understand the neural mechanisms underlying vocal learning, an example of a learned complex motor behavior. Previous studies have shown that a premotor area HVC (a proper name) generate sequences of bursts responsible for the stereotyped adult song. However, the activity of HVC neurons during juvenile vocalizations remains unknown. This thesis provides a comprehensive characterization of HVC activity during the entire song learning process. HVC activity exhibited dramatic change in activity during the emergence of the first stereotyped component in the juvenile song. During subsong, the earliest vocalization of juvenile birds, roughly half of HVC neurons exhibited syllable-locked activity. Over several days, more neurons started exhibiting rhythmic bursts at 5-10 Hz that were locked to syllables. As a population, different neurons were active at different latencies thus forming a rhythmic neural sequence, which was often associated with prototype syllables (‘protosyllables’). Thus, growth of a rhythmic neural sequence in HVC underlies the transition from highly variable subsong to the emergence of protosyllables. This rhythmic neural sequence was then split to give rise to multiple distinct sequences corresponding to multiple distinct syllable types. During the emergence of two syllable types from a protosyllable, there were neurons that were active during both of these emerging syllable types (‘shared neurons’). Over development, the fraction of shared neurons decreased suggesting that splitting of a neural sequence in HVC underlies the emergence of a new syllable type. Moreover, this sequence splitting was used in different song learning strategies, indicating this is a fundamental neural mechanism for song learning. In summary, this work demonstrates how a growth of a rhythmic neural sequence and its subsequence splitting can give rise to the emergence of complex stereotyped vocalizations in songbirds. Given the ubiquity of neural sequence in various brain areas of different animals, it is possible that this mechanism applies to learning of complex motor behaviors in general.