Intelligent human computer interaction based on non redundant EMG signal

The function of current artificial arms is limited by inadequate control methods. We developed a technique that used nerve transfers to muscle to develop new electromyogram control signals and nerve transfers to skin, to provide a pathway for cutaneous sensory feedback to the missing hand. We did targeted reinnervation surgery on a woman with a left arm amputation at the humeral neck. The ulnar, median, musculocutaneous, and distal radial nerves were transferred to separate segments of her pectoral and serratus muscles. Two sensory nerves were cut and the distal ends were anastomosed to the ulnar and median nerves. After full recovery the patient was fit with a new prosthesis using the additional targeted muscle reinnervation sites. Functional testing was done and sensation in the reinnervated skin was quantified. The patient described the control as intuitive; she thought about using her hand or elbow and the prosthesis responded appropriately. Functional testing showed substantial improvement: mean scores in the blocks and box test increased from 4.0 (SD 1.0) with the conventional prosthesis to 15.6 (1.5) with the new prosthesis. Assessment of Motor and Process Skills test scores increased from 0.30 to 1.98 for motor skills and from 0.90 to 1.98 for process skills. The denervated anterior chest skin was reinnervated by both the ulnar and median nerves; the patient felt that her hand was being touched when this chest skin was touched, with near-normal thresholds in all sensory modalities. Targeted reinnervation improved prosthetic function and ease of use in this patient. Targeted sensory reinnervation provides a potential pathway for meaningful sensory feedback.