Virtual reality-based myoelectric prosthetic control training: Effects of action observation and motor imagery with visual feedback of electromyographic signals.

Abstract

Background: Conventional myoelectric prostheses (myo-prostheses) training involves repetitive grasping and manipulation training, which requires considerable training time. It is necessary to develop a short and efficient myo-prostheses training. This study aimed to verify the immediate and sustained effects of action observation and motor imagery (AOMI) using virtual reality (VR) on myo-prostheses control and clarify the effect of visual feedback of electromyogram (EMG) signals during AOMI using VR.

Study design and methods: We evaluated 24 healthy right-handed individuals wearing a myo-prostheses simulator in their dominant hands. We divided participants into 3 groups: VR video observation with EMG presentation during manipulation (VR+), VR video observation without EMG presentation (VR), and control group. We evaluated prosthetic control skills using the Grasp Force Adjustment Test (GFAT) and Bowknot task immediately before and after AOMI and 1 week later. In addition, we evaluated the level of immersion during AOMI.

Results: The rate of change in the GFAT 1 week after the intervention was significantly greater in the VR+ (P < 0.05, d = 1.32) and VR (P < 0.01, d = 2.34) groups than in the control group. Immersion was significantly higher in the VR+ and VR groups than in the control group. The condition and time required for GFAT had significant effects, although the post-hoc test showed no significant difference between VR+ and VR groups.

Conclusions: AOMI using VR had sustained effects on motor learning of myo-prosthetic control despite EMG presentation. Therefore, AOMI, using VR, manipulates prostheses once learned, and it might be used for future training of myo-prosthetic control.