The effect of robotic guidance on the use of visual information during a pointing task

Abstract

Robotic guidance has been shown to be effective for rehabilitation although fundamental research suggests guidance can be detrimental to performance due to the decreased need for an efferent command. However, physical guidance does provide added proprioceptive feedback which could influence how visual information is used. We sought to compare the influence of robot-guided vs. active upper limb pointing on the use of visual feedback. Participants completed a training phase comprised of 210 trials with vision to 3 target amplitudes (18, 20, 22 cm) during which participants were either guided by a robot, or actively aimed to each target (Control). We also included pre- and post-tests of 20 trials each (10 vision, 10 no vision) using the 20 cm amplitude. A 2 group (robot, control) by 2 phase (pre, post) by 2 vision (vision, no vision) mixed ANOVA was done on all accuracy and movement time variables. Overall, participants exhibited lower variability and were more accurate with vision compared to no vision. A group by phase interaction demonstrated that the control group exhibited a shorter reaching amplitude than the robot group after training, regardless of vision condition. This suggest that the altered efferent requirements, and the additional proprioceptive feedback presented during robotic guidance does not negatively impact the sensorimotor representation of a reaching task.

Acknowledgments: This research was supported by NSERC