More than 600,000 people in the U.S. have severely impaired motor function from disorders including amyotrophic lateral sclerosis, brainstem stroke, and spinal cord injury. The cost of caring for these patients runs into the billions of dollars. Brain machine interfaces (BMIs) have the potential to free patients from the shackles of these paralyzing disorders by decoding signals recorded from the brain into commands to control the movements of a computer cursor, a prosthetic limb, or even a person’s own limb. These signals have been recorded from different levels of the brain, using noninvasive electrodes on the scalp to electrodes placed on top of the brain or even inside the brain itself. In general, there is a tradeoff between invasiveness and the signal quality. Recently, several research groups have had success at allowing non-paralyzed patients, already undergoing brain surgery for different reasons, to use intermediate-level (subdural) signals to control a cursor movement in two-dimensions. Our research group proposes to investigate a less-invasive (epidural) level of signal to decode not only cursor movement, but also the movements of the fingers and muscle activity while the patient performs one of several grasps. If we can decode individual finger movements, or even decode between one of several grasps, we could then use these signals to control a device that would electrically stimulate a paralyzed person’s arm muscles to restore these hand and finger movements. Also, the mathematical techniques that we will use to decode these signals rely on brain signals that are related to the way people normally move. This is different from the BMIs that currently are used with subdural signals.
Marc Slutzky, M.D., Ph.D.