(50) Model-based design of electrodes for restoring limb sensation via spinal cord stimulation

Introduction: Current prosthetic devices lack the capability to provide somatosensory feedback to lower-limb amputees, limiting their effectiveness in restoring balance control and gait. Prior studies have demonstrated that dorsal root stimulation can evoke somatosensory percepts in the feet; however, these percepts often extend to more proximal regions of the residual limb, reducing their functional utility. This computational modeling study aims to design novel epidural electrode arrays capable of selectively stimulating spinal rootlets to generate focal sensations in the absent limb. By optimizing neural selectivity through customized electrode configurations, we seek to improve gait control and reduce phantom limb pain.

Methods: We developed a comprehensive computational modeling framework to evaluate how system design parameters influence neural selectivity in epidural stimulation. We constructed a three-dimensional volume conductor model of the human lumbo-sacral spinal cord (L4-S2) using high-resolution magnetic resonance imaging data from volunteers and anatomical data from cadavers. We incorporated multi-compartment models of sensory neurons, with trajectories reconstructed from imaging data, and designed custom 32-contact epidural electrode arrays positioned over the lateral spinal cord. We computed the extracellular potential fields generated by various monopolar and bipolar stimulation configurations and simulated neural responses to these fields. Additionally, we employed an optimization algorithm to determine the stimulation parameters and electrode configurations that maximize selective activation of spinal rootlets, which is crucial for evoking focal sensations in the limb.

Results: Our results indicate that while monopolar stimulation enables highly selective activation, bipolar configurations provide selective recruitment over a broader stimulation amplitude range, potentially increasing the therapeutic window. Furthermore, our optimized stimulation paradigms demonstrated the ability to target specific rootlets at each spinal level, enhancing the precision of sensory restoration.

Conclusion: This study provides key insights into optimizing epidural stimulation strategies for restoring somatosensory feedback in lower-limb amputees. By leveraging computational modeling and optimization algorithms, we designed electrode arrays that enhance neural selectivity, enabling more precise and targeted sensory percepts. These findings have significant implications for improving gait control, reducing phantom limb pain, and ultimately enhancing the quality of life for lower-limb amputees.