Testing peripheral somatosensory neuroprostheses by recording from raccoon cortex

A topologically organized representation of the body surface exists within the mammalian somatosensory cortical areas such that stimulation of a part of the body surface will produce a response in a localized region of the contralateral somatosensory cortex. Because of this topography, we propose that the selectivity of a peripheral somatosensory neuroprosthetic electrode array can be tested by noting whether the locus of maximum activation in the cortex moves in a consistent manner when differing portions of the array are stimulated. We further propose that the raccoon might well be the ideal animal in which to test this hypothesis, since the raccoon has a rather unique cortical somatosensory area where each digit is represented in individual subgyri around the tri-radiate sulcus. To demonstrate the feasibility of this concept, a pilot study was carried out in one raccoon under barbiturate anesthesia. The median nerve was stimulated via selective quadrants of a nerve cuff array of four tripolar electrodes implanted around the nerve. Cuff stimulation produced short-latency evoked surface potentials in the digit areas of the raccoon first somatosensory cortex. Response selectivity could be demonstrated, as could a separation between thresholds for producing movement or producing cortical evoked potentials. The sensory and motor responses elicited were consistent with the orientation of the median nerve within the cuff as determined by a postmortem identification of the muscle innervation pattern of the nerve.