Dissecting Motor Adaptation in Visually Guided Reaching Movements

Abstract

Movement is essential to human life because it provides us with the freedom of mobility and the power to affect our surroundings. Moreover, movements are vital to communication: from hand and finger movements when writing, mouth and throat movements when speaking, to painting, dancing, and other forms of artistic self expression. As people grow and experience new environments, adaptively maintaining the accuracy of movements is a critical function of the motor system. In this dissertation, I explore the key mechanisms that underlie the adaptability of simple visually guided reaching movements. I specifically focus on two key facets of this adaptability: how motor learning rate can be predicted by motor variability and how motor learning affects the mechanisms which underlie movement planning. Inspired by reinforcement learning, I hypothesized that greater amounts of motor variability aligned with a task will produce more effective exploration, leading to faster learning rates. I discovered that this relationship predicts person-to-person and task-to-task differences in learning rate for both reward-based and error-based learning tasks. Moreover, I found that the motor system actively and enduringly reshapes motor output variability, aligning it with a task to improve learning. These results indicate that the structure of motor variability is an activelyregulated, critical feature of the motor system which plays a fundamental role in determining motor learning ability. Combining prominent theories in motor control, I created a model which describes the planning of visually guided reaching movements. This model computes a weighted average of two independent feature-based motor plans: one based on the goal location of a movement, and the other based on an intended movement vector. Employing this model to characterize the generalization of adaptation to movements and movement sequences, I find that both features, movement vector and goal location, contribute significantly to movement planning, and that each feature is remapped by motor adaptation. My results show that multiple features contribute to the planning of both point-to-point and sequential reaching movements. Moreover, a computational model which is based on the remapping of multiple features accurately predicts how visuomotor adaptation affects the planning of movement sequences.