PROJECT SUMMARY Motor skills that we learn early in life (e.g. talking or reaching) are essential to everyday function. Learning these motor skills involves building a repertoire of distinct motor actions (“motor gestures”) followed by refining these gestures. Through repeated practice, the brain integrates sensorimotor feedback to refine behavior until consistent high levels of performance is achieved. Prior studies have used anatomical, genetic, and molecular approaches to characterize the changes that occur during motor learning. However, since the successful execution of motor skills rely on the precise relaying of signals from the brain to coordinate precise activations of ensembles of muscle, a critically unsolved problem is the neurophysiological understanding of how changes in the brain’s activity patterns enable skill learning. Thus, the overall goal of our research is to understand how the brain’s motor activity changes during learning to enable animals to create new gestures and finely tune each gesture to high levels of reliability and precision. I propose to study the neural control of behavior during skill acquisition by studying birdsong, a highly quantifiable behavior that is learned during development through imitation of an adult song “tutor”, a process for which the neuroanatomical and behavior architecture has been well-studied and characterized. During song learning, juvenile birds acquire new units of vocalizations (“syllables”) which are refined until high levels of acoustic precision is reached in adulthood. My objective is to examine whether and how the neural control of song vocal behavior changes during the process of syllable creation and refinement. The central hypothesis of this proposal is that the nervous system learns motor gestures by first creating assemblies of co-active neurons specific to each gesture and then by refining the precisely timed spike patterns within that ensemble. I will combine advanced electrophysiological methods with novel mathematical (information-theoretic) tools to quantify the relationship between neural activity and behavior (“neural code”) across skill learning. I will test our central hypothesis by examining changes in the activity patterns of individual cortical neurons across song learning in juvenile Bengalese finches (Aim 1), and in the activity patterns across neural ensembles in the cerebellum during the learning of reaching tasks in mice (Aim 2). These studies will reveal new insights into how the brain implements precise patterns of activity to learn motor skills. Broadly, these results will provide a framework for investigating how transformations in spike patterns drive learning across different species.