Computational approaches to motor control and their potential role for interpreting motor dysfunction

Abstract

Purpose of review Computational frameworks, notably internal models and optimal control theory, have led to rapid advances in our understanding of how the brain plans and controls movement. The purpose of this review is to provide an overview of these theoretical ideas, how they have been used to interpret motor control, as well as their potential role for interpreting motor dysfunction. Recent findings There are two general types of internal models, neural processes that mimic the mechanical properties of the limb (and environment). Forward internal models parallel the normal causal flow of the motor periphery and estimate limb motion from motor commands. Inverse internal models perform the reverse process by estimating motor commands from signals related to intended limb motion and/or spatial targets. This framework has led to several important behavioural observations on motor planning, control and learning, and has also been influential for interpreting neural activity in awake, behaving non-human primates. A more recent framework for interpreting motor function is optimal control theory, which recognizes that noise or errors are an inherent feature of the motor system and may influence strategies to plan and control movement. Summary Internal models and optimal feedback control both provide frameworks for interpreting motor performance, and may be of value for interpreting many motor dysfunctions associated with neurological injuries. Advanced technologies such as robots that have played a key role in these frameworks may be also of considerable value for motor assessment and rehabilitation.