Project Summary The ability to learn from experience is one of the most fundamental features of neural circuits. Changes in synaptic connections in specific circuits underlie experience-dependent circuit modifications essential for learning. A detailed understanding of this process is important, not just to understand the mechanisms of learning, but also to better diagnose and treat conditions that affect memory abilities, such as Alzheimer's disease, aging-related dementia, and Parkinson's disease. Our ultimate goal is to understand the precise, fine- scale circuit modifications that support learning. One of the fundamental forms of learning is motor learning in which animals adjust the way they move their bodies to fit their behavioral goals. Among a number of brain areas involved in motor learning, the primary motor cortex (M1) is a major locus where changes take place during motor learning. Many types of changes in M1 have been described that accompany motor learning, including changes of the somatotopic map, neural population activity changes, and synaptic plasticity. However, it is unclear whether M1 is always involved in the control of movements throughout learning and overtraining. Furthermore, the precise functional reorganization of synaptic inputs in M1 during motor learning is only beginning to be understood. We will address these two questions using cutting-edge technologies in mice. Mice under head-fixation will be trained in a forelimb-based motor learning task daily over weeks. In Aim 1, we will perform longitudinal recording of M1 neural populations during months of motor learning and overtraining. Combined with optogenetic perturbation of M1 activity at various phases of training, we test the hypothesis that a movement that is dependent on M1 early in learning can become M1-independent with long-term overtraining. This will also define the period during which the particular motor task we use in the proposal depends critically on M1. Focusing on this period when M1 is critical for motor performance, we will study precise functional reorganization of synapses in M1. We will do this using longitudinal functional imaging at synaptic resolution. In particular, we will test the hypothesis that motor learning induces functional clustering of synaptic inputs related to the learned movements. Such functional clustering would allow the learning-related information to robustly drive circuit activation. These experiments will contribute fundamental neural circuit mechanisms underlying motor learning. Such knowledge could ultimately contribute to a better diagnosis and treatment of motor disorders such as Parkinson's disease and stroke.