Chronic Peripheral Nerve Interfacing For Closed-loop Sensorimotor Control Of Robotic Hands
Jonathan Cheng, MD1, Edward Keefer, PhD2.
1University of Texas Southwestern Medical Center, Dallas, TX, USA, 2Nerves Incorporated, Dallas, TX, USA.
PURPOSE We are developing a closed-loop sensorimotor control solution to provide hand dexterity in amputees with robotic limbs. Our approach is based on surgical targeting of fascicles within the residual peripheral nerves of the amputated hand. METHODS We implanted 1 transradial and 3 partial hand amputees with chronic intraneural electrodes in selected fascicles of the median/ulnar nerves, for 90 days. Weekly electrophysiology sessions include: sensory stimulation with quantitation of tactile and kinesthetic percepts, proprioceptive maps, and functional tasks; motor recording during imagined hand movements and stereotyped positions. RESULTS Individual prosthetic finger control and functionally relevant sensations are reliably obtained. Sensory stimulation produces a range of stable percepts (e.g., light pressure, pulsing, squeezing, individual finger movements), and can be tuned using modulation of pulses/trains and multielectrode stimulation. Subjects can reliably discriminate up to 7 stimulus intensities within a single anatomical location, and detect stimulus trains as short as 2 pulses. When mapped to the subject's robotic hands, percepts can be used to identify contact events on multiple digits and joint motion during grasping. Recorded data can be decoded to identify imagined movement of specific amputated fingers with high accuracy. CONCLUSIONS Surgical targeting of nerve fascicles enables stimulation of modality-specific sensory percepts in hand amputees. Motor signals can be decoded to identify subjects' imagined finger movements. When fully implemented, we believe these motor decode algorithms and sensory restoration protocols will provide dexterous control of prosthetic hands with restoration of both tactile and proprioceptive sensations.
Back to 2018 Program