Enhancing Engagement during Robot-Assisted Rehabilitation Integrated with Motor Imagery Task

Abstract

Stroke remains the most common cause of motor deficits for adults. Enhancing engagement has become the focus of recent research with the aim of improving the efficiency of robot-assisted rehabilitation. Since motor imagery (MI) has the potential to engage the subject, the objective of this study is to explore the influence of complementing robot-assisted rehabilitation with MI during training exercises. An experiment was designed and conducted in which 10 healthy subjects were recruited to participate in two separate sessions. An acoustic-cue-based experimental paradigm was applied in both sessions. In the first session, each patient was required to imagine moving arm after the cue, then the robot device drove the arm during the MI process; while in the second session, the robotic device drove the user to move without requiring the MI tasks. Each session consisted of 20 trails, in which electroencephalogram (EEG) was recorded to analyze the activated brain regions. Analyses showed that the activation of sensorimotor cortex (SM1) was the strongest during passive movement (PM) integrated with MI than either PM or MI alone. The results indicated that robot-assisted training integrated with MI task can enhance the subject's engagement as shown by a stronger event related desynchronization (ERD), which can lead to a stronger stimulation on SM1. This indication can explain why only passive movement driven by robotic device has a low rehabilitation efficiency during clinical practice. The result can also contribute to the understanding of the mechanism underlying the brain computer interface (BCI) supported rehabilitation therapy, which can improve rehabilitation efficiency by closing the loop between the motor intention and sensorimotor feedback.