Contributions of Implicit and Explicit Memories to Sensorimotor Adaptation of Movement Extent During Goal-Directed Reaching

We examined how explicit memory of kinematic performance contributes to motor adaptation during goal-directed reaching. Twenty subjects grasped the handle of a horizontal planar robot that rendered spring-like resistance to movement. Subjects made rapid "out-and-back" reaches to capture a remembered visual target at the point of maximal reach extent. The robot9s resistance changed unpredictably between reaches, inducing target capture errors that subjects attempted to correct from one trial to the next. The subjects performed four sets of trials. Three were performed without concurrent visual cursor feedback of hand motion. Two of these required self-assessment of performance between trials, whereby subjects reported peak reach extent on the most recent trial. This was done by either moving a cursor on a horizontal display (visual self-assessment), or by moving the robot9s handle back to the recalled location (proprioceptive self-assessment). Two control conditions were performed either without or with concurrent visual cursor feedback of hand motion. We analyzed movement kinematics and used regression analyses to quantify the extent to which prior reach errors and explicit memories of prior performance contribute to subsequent reach performance. Consistent with prior reports, providing concurrent visual feedback of hand motion increased reach accuracy and reduced the impact of past performance errors on future performance, relative to the no-vision control condition. By contrast, we found no impact of interposed self-assessment on reach kinematics or on how prior target capture errors influence subsequent reach performance. Self-assessments were biased toward the remembered target location, differing markedly from actual reach performances. Including self-assessments as explicit memory of performance in the regression analyses did not improve model predictive accuracy. Therefore, we conclude that explicit memories of prior kinematic performance do not contribute to implicit sensorimotor adaptation of movement extent.