Generating adaptive body movements is a fundamental function of the brain. The primary motor cortex (M1) is a central locus for motor learning and execution. M1 receives long-range inputs from several brain areas and processes this information through local recurrent connections. However, how various inputs and local connections work together to shape M1 activity is unknown. In this proposal, we will investigate the contributions of long-range inputs and local connections to M1 activity and how they are shaped during motor learning. We will do this using a well-established motor learning paradigm in mice. Our central hypothesis is that the functional properties of individual M1 neurons are defined by their responses to long-range inputs and local neurons. We further hypothesize that learning induces a reorganization of the responses of M1 neurons to specific inputs. This proposal will investigate the nature of the information that M1 receives from various upstream regions, the unique activity patterns of M1 neurons that receive these inputs, and how the local M1 network contributes to their activity patterns.