Talk Title:

Distributed neural computations supporting motor intelligence and dexterity

Talk Abstract:

Smooth, coordinated movement is essential to human expression. By studying neural responses accompanying isolated movements, such as point-to-point reaches, we are beginning to understand how the motor cortex generates output signals that ultimately activate muscles and move the body. In addition to generating simple, decontextualized movements, our versatile motor faculties enable us to interact skillfully with hundreds of objects, tools, and engineered devices. Each of these motor skills must engage tailored neural computations to perform skill-specific pattern generation, predictive feedforward coordination, and reactive feedback control. 

In this seminar, I will discuss two avenues of my previous work that aim to elucidate how the macaque motor cortex supports motor skills. First, I will explore the neural population mechanisms underlying motor learning during a force field adaptation task. When a set of distinct force fields was learned in sequence, we observed field-specific shifts in neural activity that separated the associated motor memories in the neural state space. The precise, predictable geometry of these shifts in preparatory activity suggested they serve to index motor memories, facilitating the acquisition, retention, and retrieval of a broad motor repertoire.

Second, I will present results relating to direct neural perturbations of the motor cortex, using optogenetic and electrical stimulation alongside Neuropixels recordings. We develop a novel analytic approach that relates measured activity to theoretically tractable, dynamical models of the cortical circuit; this allows these perturbations to reveal the dynamical mechanisms that shape patterns of neural population activity in ways that passive observation of activity cannot. I will demonstrate that the motor cortex isolates neural computations needed for a specific behavioral context within a “self-contained” neural subspace. This result suggests a possible general mechanism for robust computation, as well as a neural basis for compartmentalizing neural computations associated with specific motor skills.

Overall, these results highlight a considerably richer portrait of the computations performed by neural population dynamics supporting skilled movement. Understanding how the motor system orchestrates skill-specific control will be essential to establishing a neural computational basis for motor intelligence, enabling us to solve functional problems via motor behavior in a creative and efficient manner.