(36) Lateral Spinal Cord Stimulation Parameters to Optimize Selective Recruitment of Sensory Nerve Branches in Cats.

Introduction: Restoring somatosensory feedback in lower-limb amputees improves mobility, reduces fall risk and reduces phantom limb pain. Using clinical electrodes, our group restored sensation from the missing foot of prosthetic users via lateral spinal cord stimulation (SCS). However, subjects simultaneously report sensation in their residual limb. To limit these sensations, we evaluated the ability of high-density paddle electrodes (16 × 2 array), with smaller contact sizes to evoke sensations in focal regions of the limb.

Methods: In six acute feline experiments, we performed an L4-L7 laminectomy to expose the epidural surface of the spinal cord. The SCS electrodes were placed over the lateral surface of the cord and secured under the laminectomy border. We delivered stimulation to the cord (15–500µA; symmetric biphasic pulses - 66µs pulse width; 33µs interphase interval) and recorded antidromic compound action potentials in the sciatic and femoral nerve branches of the ipsilateral hindlimb. We compared selective nerve recruitment at threshold (the minimum current that elicited nerve responses) and the range of amplitudes where selectivity is maintained (dynamic range) across three stimulation configurations (bipolar, lateral monopolar, and medial monopolar) and electrode diameters (1.0mm, 0.5mm, and 0.15mm). We measured conduction velocities of recruited afferents to classify their fiber types.

Results: Thresholds were lower with bipolar stimulation than with monopolar stimulation. The 1.0mm diameter electrodes had higher thresholds, whereas we observed no significant difference between the thresholds for the 0.5 mm and 0.15 mm diameter electrodes. Stimulation using lateral monopolar electrodes was more selective (74%) than using bipolar (62%) or medial monopolar (55%) electrodes. The 0.5 mm diameter electrodes also produced the most selective nerve recruitment (68%) compared to 0.15 mm (62%) and 1.0 mm (62%) diameter electrodes. L4 and L5 stimulation selectively recruited femoral branches, whereas L6 and L7 stimulation primarily recruited sciatic branches. Conduction velocities at threshold were mostly greater than 60m/s, suggesting that we mostly recruited large diameter afferents. Lastly, lateral monopolar stimulation had larger dynamic ranges of selectivity than the other configurations.

Conclusion: While we observed statistical differences in threshold distributions, the median thresholds are less than 5µA apart, across the different configurations and sizes. Overall, bipolar and lateral monopolar stimulation are more selective at threshold. Similarly, the 0.5 mm diameter electrodes were the most selective at threshold. These results demonstrate the ability of our custom electrodes to recruit specific nerves branches, suggesting that we can target focal regions of the leg for somatosensory neuroprostheses.