This year, a monkey named Thor picked up a ball and put it into a cup while all the nerves in its hand were turned off. Rather than send signals through the spinal cord, the monkey’s brain was talking to the muscles through tiny electrodes embedded in its wrist and hand. Suddenly the lifeless limb moved with control and intention. This experimental therapy, called brain-controlled functional electrical stimulation, or FES, bypasses the spinal cord altogether, and could one day provide the means for re-animating paralyzed limbs.
Spinal cord injury is a communication problem. Our brains do all the planning and scheming, but they can’t send orders to the body without the spinal cord, which shuttles electrical messages into the muscles. When this data stream gets shut down, the limbs are as isolated as an astronaut passing behind the moon—completely cut off from their command center. The body has no way of repairing the lines of communication, but neuroscientists have come up with a way to patch it externally.
While there has been considerable excitement this week about technology that allows people who are paralyzed control external robot arms, the technology Thor demonstrated could allow patients to control their own damaged bodies once again.
Lee Miller, a neuroscientist at Northwestern University calls it “eavesdropping” on the brain. Miller opens a tiny window in the skull and implants electrodes right into the primary motor cortex. “We’re talking about a small amount of pinpricks on the very surface of the brain,” he says. With this proverbial glass on the door of the brain, Miller can listen as it chatters away about when and where it wants to move. Fortunately, the brain continues to mutter to itself even after the spine has been irreparably injured. In a recent paper in the journal Nature, Miller and his colleagues proved that he could take these electrical recordings and use them to activate electrodes in a paralyzed limb, essentially building a data detour around the spine.
“It could give these patients essentially voluntary control of their limbs—of course not as good as normal, but a reasonable approximation of that,” he says.
Miller calls it “eavesdropping,” but listening is really the easy part. Decoding the language of the brain is where it starts to get tricky. To do this, Miller had to build an algorithm that could guess the brain’s intention based on patterns of electrical activity. He began his experiments by taking simultaneous recordings from the brain and from muscles in the hand while allowing Thor to move his arm freely. Over time, patterns emerged and the algorithm began to pin down the specific neural signatures of all the commands being sent to the muscles.
In the next phase of the experiment, Miller paralyzed Thor’s hand with a local anesthetic, dulling sensation and movement. The algorithm continued to decode the intention of the brain, but now this signal got sent down to electrodes in the monkey’s wrist and hand and controlled the timing of their electrical output. The electrodes pulsed in tiny bursts, resulting in smooth, continuous muscle contractions.
Many patients with spinal cord injury already have these electrodes embedded in their muscles, but they control them other muscles in their bodies, rather than with their brains. Those that still have enough movement can push a button to activate electrodes in their trunk, bladder, or legs. More impaired patients may have to wiggle their ears or tilt their heads. Each option is more like activating the electrodes by remote control than by intuition. Until now, this has been the best technology available.
Such an anemic signal makes it especially difficult for patients to manipulate intricate hand movements, according to Dawn Taylor, a neuroscientist at the Cleveland Clinic who oversees FES clinical research at the VA Medical Center in Cleveland, Ohio.
