Spinal Cord Microstimulation Generates Functional Limb Movements in Chronically Implanted Cats Vivian K. Mushahwar, David F. Collins, 1 and Arthur Prochazka Division of Neuroscience, University of Alberta, Edmonton, Alberta T6G 2S2, Canada Received September 3, 1999; accepted February 4, 2000 Spinal cord injuries disrupt the communication be- tween the brain and peripheral nerves, but leave mo- toneurons and networks of interneurons below the level of the lesion intact. It is therefore possible to restore some function following injury by providing an artificial stimulus to the surviving neurons below the level of the lesion. We report here on a novel ap- proach for generating functional movements by elec- trically stimulating the spinal cord through chroni- cally implanted ultrafine, hair-like electrodes. Six to 12 microwires were implanted in the lumbar enlarge- ment of intact cats for 6 months. Twice a week, trains of stimuli were delivered through each microwire and the evoked electromyographic and torque responses were recorded. Strong coordinated hindlimb move- ments were obtained by stimulating through individ- ual electrodes. The joint torques elicited were capable of supporting the animals’ hindquarters. The re- sponses were stable over time and the contractions caused no apparent discomfort to the animals. No ob- vious motor deficits were seen throughout the 6-month duration of implantation. The results demon- strate that microwires implanted in the spinal cord remain stably in place and stimulation through these electrodes produces strong, controllable movements. This provides a promising basis for the use of spinal cord neuroprostheses in restoring mobility following spinal cord injury. © 2000 Academic Press Key Words: spinal cord injury; electrical stimulation; neuroprostheses; microelectrode stability; control of movement; chronic implants. INTRODUCTION Recovery of function following spinal cord injury re- mains a daunting medical challenge. The difficulty in achieving regeneration of functional neural connec- tions in the central nervous system has prompted the development of systems that use electrical stimulation of muscles or nerves to improve respiration, micturi- tion, and motor function (16). Intramuscular, epimy- sial, or nerve cuff electrodes are used to activate mus- cles. Though some neural prostheses have been suc- cessful (e.g., diaphragm pacers (3) and foot-drop stimulators (20)), the electrical restoration of whole limb movements remains problematic (11, 16). Spinal cord microstimulation (SCstim) may provide an alter- native. Ultrafine electrodes placed in the ventral horn can activate limb muscles through ensembles of inter- neurons and motoneurons. The spinal cord is distant from moving muscles, so electrodes are less likely to be dislodged or damaged. The lumbar enlargement of the cord is relatively short (5 cm in adult humans), al- lowing activation of the main limb muscles by elec- trodes implanted in a small, protected region. Previous experiments in sodium pentobarbital-anesthetized cats showed that muscles can be selectively activated by SCstim to produce smooth, graded contractions with little fatigue (7–9, 18). However, the question remained whether the same selectivity would pertain in the nonanesthetized spinal cord and whether im- planted electrodes would remain stably in place. This is the first full report showing that these prerequisites of a viable spinal neuroprosthesis can be met. METHODS Arrays of 6 to 12 microwires were implanted in the lumbosacral region of the spinal cord of healthy, adult cats. The techniques of implantation and electrode sta- bilization were based on those developed for chronic recordings from single neurons in spinal dorsal roots (10). Placement of the microwires was based on maps established in acute experiments (8, 9). The wires were precut and bent to an appropriate angle and length so once they were inserted in the spinal cord (to a depth of 3.5 to 4.5 mm), their epidural portions lay flat on the dura mater (Fig. 1). Animals were maintained from 2 to 24 weeks (mean 9 weeks) following surgery. Intact animals were used to test the stability of intraspinal microstimulation under the most demanding condi- tions: free, normal bodily movement. We were partic- 1 Present address: Prince of Wales Medical Research Institute, High Street, Randwick, Sydney, NSW 2031, Australia. Experimental Neurology 163, 422– 429 (2000) doi:10.1006/exnr.2000.7381, available online at http://www.idealibrary.com on 422 0014-4886/00 $35.00 Copyright © 2000 by Academic Press All rights of reproduction in any form reserved.