Robot therapist versus human therapist: Evaluating the effect of corrective feedback on human motor performance

Abstract

Cerebral Palsy (CP) is the most common motor disability in childhood, affecting nearly 1 in 323 children in the United States. Repetitive therapeutic exercises play a key role in upper-body rehabilitation interventions during which a therapist provides corrective feedback to a patient based on the patient's motor skill performance. Recently, an innovative system combining a serious game with an interactive robot has emerged as a powerful tool in enhancing upper-body rehabilitation and intervention outcomes. Although several studies have shown that integrating robots into physical therapy sessions can encourage engagement and improve the efficacy of the rehabilitation protocol, most studies have not directly compared outcomes when using a robot therapist versus a human therapist. The present study aims to evaluate whether a therapy intervention coupled with a robot agent is as effective as an intervention coupled with a human agent. We evaluate this effectiveness in terms of human motor performance and intrinsic motivation. A between-subject experiment was performed with twenty participants. All participants were randomly assigned to one of the following groups: 1) participants received corrective feedback from a robot agent or 2) participants received corrective feedback from a human agent. Results showed that participants in the robot therapy group improved faster than participants in the human therapy group, but the effect from the corrective feedback lasted longer in the human therapy group than the robot therapy group. The Intrinsic Motivation Inventory (IMI) survey indicated comparable results between the two groups. The experimental results provide further evidence towards supporting the efficacy of a robotic therapy coach for children with disabilities, and motivate future studies in this domain.