Proprioceptive Sensorimotor Integration With Neural Interfaces for Hand Prostheses

Hands are the primary means of interacting with the world, and Veterans who have lost a hand must cope with diminished functional and sensory capabilities. While myoelectric prostheses can restore the basic grasping function of the lost hand, many amputees abandon use of prostheses because current prosthetic options do not meet their needs. In addition, many amputees struggle with phantom pain and psychosocial deficits following limb loss that are not fully mitigated by current treatments.

In the normal sensorimotor system, proprioception, which is the sense of limb position and movement, is critical for informing motor plans and correcting errors in movement. In addition, proprioception plays a key role in body ownership, agency, and phantom pain. Despite its importance, proprioception is absent in current prosthetic options.

The goal of the proposed project is to make upper limb prosthetic devices more natural and useful for Veterans who have lost a hand by providing proprioceptive feedback about prosthesis postures and movements. The central hypothesis is that effectively closing the loop between a prosthesis user and their device requires that the sensory inputs and control outputs closely mimic the underlying sensorimotor neural processes of the body schema. The investigators will examine integration of artificial proprioception with prosthesis control in eight trans-radial or four trans-humeral amputees using the investigators' implanted neural-interfacing system and intramuscular EMG.

Aim 1. Characterize proprioceptive sensations elicited by peripheral nerve stimulation. The investigators will test time-varying Peripheral Nerve Stimulation (PNS) patterns and examine the impact on the kinematics of the evoked proprioceptive percept using position matching with the contralateral intact limb and multivariate regression analyses. The investigators will investigate paired agonist-antagonist stimulation strategies and compare the discriminability between agonist-only and paired agonist-antagonist strategies using psychometric tests. As well as assess the psychometric properties of PNS-evoked proprioceptive percepts and examine the interaction of proprioception with concurrent touch feedback and voluntary muscle contraction.

Aim 2. Investigate the mechanisms of stimulation-evoked proprioception. The neurophysiological basis of proprioception from PNS is unknown. To investigate whether proprioceptive percepts originate from direct afferent activation or indirect recruitment via direct muscle activation, the investigators will perform a motor block of the residual forearm of trans-radial participants and examine changes to the proprioceptive percepts. In trans-humeral participants, Targeted Muscle Reinnervation will be performed at the time of system implant, and percepts will be compared before and after muscle ingrowth.

Aim 3. Determine the integration of proprioceptive stimulation with motor control during posture matching. To restore proprioception to Veterans during active prosthesis use, PNS-evoked percepts must be integrated with voluntary limb control. The investigators will determine the impact of proprioception on motor control through a virtual reality posture matching task and compare performance with velocity and position-based prosthetic hand controllers with and without proprioceptive PNS. The investigators will determine the impact of voluntary control on the perception of PNS-evoked proprioception through a psychometric dissimilarity rating task. The investigators will assess the impact of artificial proprioception on embodiment via surveys. The investigators will also explore how control of multi-DOF movements with intramuscular EMG signals can be decoded via machine learning (ML)-based controllers by comparing the integration of proprioception with machine learning-based velocity control vs machine learning-based position control in virtual hand posture matching tasks. Investigators will access this impact by having the participants perform functional tasks in the laboratory with the best performing controller and a multi-DOF prosthetic hand with joint position sensors.

The project will be the first to investigate the sensorimotor integration of proprioception from PNS in a broad population of upper-limb amputees and with both position and velocity controllers. The study will provide important information about the kinematics of artificial proprioception and the role of direct muscle activation in forming proprioceptive percepts. The investigators expect that knowledge gained in this proposal will improve prosthesis utility and acceptance for Veterans with limb loss and will advance prosthetic hand technology and standard of care in the neurorehabilitation field.