Robotic guidance with variability of practice can improve the learning of a golf putting task


Although robotic guidance has yielded limited effectiveness in improving motor functions in neurologically intact and patient populations (e.g., Krishnan et al., 2012; Kummel et al., 2014), it has typically employed constant practice. Employing variability of practice principles, our lab has previously employed robotic guidance to acutely improve movement smoothness of a discrete trajectory (see Manson et al., 2014). The purpose of the current study was to investigate the impact of physical guidance involving variability of practice on the learning of a sequential movement, namely a golf putt. The current study employed a pre-test, a training phase, followed by an immediate and a 24-hr post-test. During the pre-test, the kinematic data from the club head was collected and converted into robotic coordinates to be executed using a robot arm, which is highly accurate, consistent, and smooth (see Manson et al., 2014). During training, three groups of novice participants performed putts towards 3 targets (i.e., 192, 213, & 234 cm amplitudes), benefiting from robot guidance on 0%, 50% or 100% of training trials. Only the group that trained with the robot 50% of the trials significantly reduced the endpoint distance and variability between the pre-test and the immediate and/or 24-hr post-test. This study demonstrates that-following a single acquisition session-the combination of unassisted and robot assisted practice represents the most optimal approach to facilitating short-term learning of a sequential movement. Such work could be relevant to improving putting performance and other sport skills in addition to other practical areas (e.g., rehabilitation).

Acknowledgments: Natural Sciences and Engineering Research Council of Canada (NSERC), Canada Foundation for Innovation (CFI), Ontario Research Fund (ORF), University of Toronto