Natural versus artificial sensors applied in peroneal nerve stimulation

We have attempted to quantify the performance of natural versus artificial sensors when used in a closed-loop functional electrical stimulation system. Peroneal nerve stimulation was applied during gait to a multiple sclerosis subject with a drop foot. Stimulation was applied only during the swing phase to provide artificially induced dorsiflexion of the foot. Detection of the onset of the stance phase was accomplished using a standard heel contact switch mounted inside the subject's shoe (the artificial sensor) and using processed nerve signals derived from an implanted nerve-cuff electrode (the natural sensor). A detection percentage of at least 85% was achieved using the afferent nerve signal information only. When muscle activity (also recorded in the cuff) and additional information about the gait cycle were incorporated, functional detection ratios approaching 100% were achieved.