(27) High-density Intraspinal Neural Interface for the Reanimation of the Hand After Spinal Cord Injury

Introduction: Restoring hand function is the top priority for patients with hand paralysis following cervical spinal cord injury (SCI), yet no existing therapy is capable of recovering dexterous hand control in this population. In recent years, our group has utilized epidural spinal cord stimulation (SCS) to successfully improve upper limb strength and gross motor function. Unfortunately, SCS has not been able to produce fine hand motor control due to fundamental biophysical constraints. Specifically, because the various hand muscle groups share common spinal roots (C8–T1), you cannot selectively activate the afferents to target a specific muscle post-synaptically. We hypothesize that by directly targeting specific motor pools, we can selectively hand and digit musculature. To test this hypothesis, we retrospectively analyzed data from cervical intraspinal microstimulation (ISMS) collected by Neuralink as well as performed in silico analysis of activation properties for local neurons using intraspinally implanted electrodes.

Methods: We used data previously collected by Neuralink applying cervical ISMS in an anesthetized macaque using a high-count Neuralink system. We identified relationships between stimulation locations and parameters (e.g., electrode location, stimulation frequency, and stimulation amplitude) and muscle recruitment measured via electromyography (EMG). Finally, using micro-CT, we developed a subject-specific volume conductor model of the monkey spinal cord to analyze current flow and neural recruitment patterns for the various intraspinal electrodes, as well as test novel stimulation protocols.

Results: Low frequency ISMS ( < 100 Hz) was able to selectively recruit muscles to an unprecedented degree and in agreement with the expected rostro-caudal distribution of spinal motor pools (including hand and digit muscles), whereas higher stimulation frequencies co-activated numerous muscle groups. Computational modeling revealed that conventional microstimulation preferentially activated axons and axon terminals from non-local cells at lower amplitudes than neurons with somata near the stimulating electrode. Modeling results indicate that novel stimulation paradigms could have the opposite effect, allowing for preferential recruitment of local neurons and more focal regions of activation.

Conclusion: Our results show that low frequency high-density ISMS has the potential to engage spinal motoneurons with high-selectivity that could be further increased by novel stimulation protocols. These results offer a path towards clinical justification of a high-density human ISMS system.