Project Summary Complex motor sequences are fundamental to many highly skilled behaviors, ranging from athletics to musical and vocal expression. Learning such complex movements requires both motor variability, to facilitate exploration important to reinforcement learning, and motor flexibility, to enable the adaptive modification of behavior in response to reinforcement signals. Studies of relatively simple skill learning, such as lever pressing or licking, emphasize that interactions between the motor cortex and the striatum are central to both exploration and reinforcement. However, how corticostriatal circuits facilitate motor exploration and reinforcement of complex behaviors such as speech remains unknown. The goal of this proposal is to monitor and manipulate the activity of corticostriatal circuits dedicated to learning complex vocal behaviors. Integrating these circuit studies with predictive computational models and novel methods of vocal analysis will allow us to test the hypothesis that corticostriatal circuits provide a source of directed variability that enables the systematic exploration of high-dimensional vocal “space.” The Specific Aims are to: 1) Use high- resolution imaging methods to determine how variability in the vocalization-related activity of cortical and striatal neurons relates to vocal variability; 2) Use cell-type specific optogenetic perturbations targeted to small neural ensembles to determine how cortical and striatal neuron activity contributes to vocal variability; and 3) Use closed-loop methods to determine how reinforcement signals modulate corticostriatal ensemble activity and how these changes in activity affect vocal performance. Collectively, the proposed experiments will help achieve our long term goal of causally testing how corticostriatal circuits contribute to motor exploration and reinforcement needed to learn complex skills.

Project Narrative Learning complex skills such as speech or musical performance requires trial and error motor “exploration” coupled with reinforcement of “better” outcomes. This proposal seeks to build predictive models of how the brain enables variations in behavioral performance to be generated and selectively reinforced to enable complex motor learning.