The number of applications using neural prosthetic interfaces is expanding. Computer models are a valuable tool to evaluate stimulation techniques and electrode designs. Although our understanding of neural anatomy has improved, its impact on the effects of neural stimulation is not well understood. This study evaluated the effects of fascicle perineurial thickness, diameter, and position on axonal excitation thresholds and population recruitment using finite element models and NEURON simulations. The perineurial thickness of human fascicles was found to be 3.0% ± 1.0% of the fascicle diameter. Increased perineurial thickness and fascicle diameter increased activation thresholds. The presence of a large neighboring fascicle caused a significant change in activation of a smaller target fascicle by as much as 80% ± 11% of the total axon population. Smaller fascicles were recruited at lower amplitudes than neighboring larger fascicles. These effects were further illustrated in a realistic model of a human femoral nerve surrounded by a nerve cuff electrode. The data suggest that fascicular selectivity is strongly dependent upon the anatomy of the nerve being stimulated. Therefore, accurate representations of nerve anatomy are required to develop more accurate computer models to evaluate and optimize nerve electrode designs for neural prosthesis applications.

Abstract-The design of neural prostheses to restore standing balance, prevent foot drop, or provide active propulsion during ambulation requires detailed knowledge of the distal sciatic nerve anatomy. Three complete sciatic nerves and branches were dissected from the piriformis to each muscle entry point to characterize the branching patterns and diameters. Fascicle maps were created from serial sections of each distal terminus below the knee through the anastomosis of the tibial and common fibular nerves above the knee. Similar branching patterns and fascicle maps were observed across specimens. Fascicles innervating primary plantar flexors, dorsiflexors, invertors, and evertors were distinctly separate and functionally organized in the proximal tibial, common fibular, and distal sciatic nerves; however, fascicles from individual muscles were not apparent at these levels. The fascicular organization is conducive to selective stimulation for isolated and/or balanced dorsiflexion, plantar flexion, eversion, and inversion through a single multicontact nerve-cuff electrode. These neuroanatomical data are being used to design nerve-cuff electrodes for selective control of ankle movement and improve current lower-limb neural prostheses.