EMG pattern recognition control of multifunctional prostheses by transradial amputees.

Electromyogram (EMG) pattern recognition approach has been investigated widely with able-bodied subjects for control of multifunctional prostheses and verified with high performance in identifying different movements. However, it remains unclear whether transradial amputees can achieve similar performance. In this study, we investigated the performance of EMG pattern recognition control of multifunctional transradial prostheses in five subjects with unilateral below-elbow amputation. Testing results on both residual and intact arms showed that the average classification error (21%) of amputated arms for ten motion classes (four wrist movements, six hand grasps) and a 'no movement' class over all five subjects was about 15% higher than that of intact arms. For six basic motion classes (wrist flexion/extension, wrist pronation/supination, and hand open/close), the average classification error over all five subjects was about 7% from residual arms, which was similar to the result from intact arms (6%). Only six optimal electrode channels might be needed to provide an excellent myoelectric control system for the six basic movements. These results suggest that the muscles in the residual forearm may produce sufficient myoelectric information to allow the six basic motion control, but insufficient information for more hand functions with fine finger movements.