The Role of Motor Behaviour in Value-Based Decisions
Speaker
Neil DundonLocation
Sage Center, Psychology 1312Info
Value-based decisions often involve appraisal of a reward outcome relative to an absorbed motor cost, however much theory describing human learning and decision-making stems from tasks where motor performance in action execution is largely trivial. Our goal was to explore an expanded model of learning and decision-making, under a control theory framework, where a controller integrates signals from action outcomes, while also appraising the current state of motor performance. We present a novel task that combines value-based decision-making with motor-planning; the task drives both an upward trajectory in decision execution in context, and also driving an upward trajectory in motor performance. We first demonstrate that outcome (reward) and motor performance (spatial and temporal skill) variables are relevant in the value-based decision process. We next test three learning models that explore the dynamics of motor performance in value-based decisions, i.e., do action outcome and motor performance (i) operate in parallel as distinct signals of action value, (ii) gate one another during the learning process, or (iii) combine at the point of the decision to probabilistically predict outcomes.
