CAREER: NEURAL MECHANISMS OF SKILL LEARNING

Carnegie Mellon University, Pittsburgh PA

Investigators

Abstract

Practice makes perfect. While most people can swing a hammer, it takes thousands of hours of practice to become a master carpenter who can swing a hammer with fluid speed and precision. This process of long-term improvement in movement accuracy is known as skill acquisition, and is presumably driven by coordinated changes in our brain's neural representation of the movement. Yet despite the bedrock importance of skill learning in our daily lives, the link between neural reorganization and skill acquisition is still largely unknown. This CAREER proposal lays the foundation for a long-term integrated research and educational program that will discover the links between cortical reorganization and skill acquisition and determine the behavioral factors that drive skill learning. An improved understanding of the science behind skill learning will have long-term impact on our clinical understanding of the progression of various motor control disorders, such as Parkinson's disease and stroke, and may inform the design of targeted rehabilitation paradigms and brain-computer interface systems for those patient groups. Learning is a fundamental principle of brain operation that impacts every aspect of neural function. Long-term practice has been accompanied by a reorganization of cortical activity in a number of brain areas, but it is typically only imaged at the level of millions of neurons. The research proposed here will explicitly probe how changes in the properties of individual neurons are linked to skill acquisition. To accomplish this goal, this proposal leverages a brain-computer interface (BCI) skill learning task that enables precise investigation into the links between neural activity and behavioral improvement. With a BCI, it is possible to create a mapping between the activity of neurons and the movement of a computer cursor. To achieve dexterous control of the cursor, BCI subjects must practice with a particular neuron-to-cursor mapping to learn how to structure their neural activity to achieve desired movements. We will train Rhesus macaques to operate BCIs and give them new skills to master by presenting them with new BCI mappings. We will then track the reorganization in neural activity that occurs over days and weeks as they gain skilled control of the device. Critically, these mappings will be tailored to test various hypotheses about the behavioral drivers (i.e., visuomotor error, visuomotor bias, and movement efficiency) of cortical reorganization. The results of this study will thus provide an unparalleled view into the neural instantiation of skill development over its entire time course.