He Lost His Sense of Touch in a Spinal Cord Injury. New Brain Implants Helped Restore It.

Keith Thomas remembers only fragments of the July 2020 accident: diving into a shallow swimming pool, a helicopter landing on the lawn, and waking the next day in a hospital unable to move. The impact fractured his neck and severely damaged his spinal cord, leaving him paralyzed from the chest down. Thomas, a Manhattan finance professional, could no longer raise his arms, hold a cup, feed himself, or even scratch his nose.

About a year after his injury, Thomas became the first patient in a pioneering long-term study at New York’s Feinstein Institutes for Medical Research that used a complex brain-computer interface (BCI) system to test whether they could restore some of that lost function. In a study published in July, researchers reported that their system could do more than translate Thomas’s thoughts into movement: By stimulating both the brain and spinal cord, it also helped him regain some movement and sensation below his injury—including a sense of touch that persisted after the devices were switched off.

Chad Bouton, who heads the Feinstein team, describes the technology as “a hybrid system that stimulates both the spinal cord and the brain to restore movement and sensation.” He calls it a brain-body interface, or BBI, because a brain-computer interface is only one part of it. “We’re creating the full bridge from the brain to the body,” he says.

Thomas can now control his wheelchair and, with the full system operating, feed himself, drink from a cup, and grasp something as delicate as an egg without crushing it. He says his returned sensation even allows him to feel his dog’s fur.  “It’s about independence—and dignity, too, not always needing someone to help wipe your mouth,” Bouton says. “Those are things able-bodied people don’t think about and take for granted.”

The technology has not restored all the physical function Thomas lost. He still requires round-the-clock care, and the gains followed years of intense rehabilitation and what he calls “a lot of trial and error”—not just a single surgery. But for a quadriplegic, he says, “even 10 percent... is a big deal.”

Thomas’s experience points to a potentially larger shift in neurotechnology. Brain-computer interfaces have traditionally bypassed damaged pathways to operate a computer or limb. By reconnecting the brain with the body, this system may also help the nervous system recover some ability to move and feel on its own. How it produced those gains—and whether they can be replicated in anyone beyond a single participant—are now the central questions.

How the “double neural bypass” works

In 2023, as part of the ongoing clinical trial, surgeons implanted five microelectrode arrays into Thomas’s brain—two in the primary motor cortex, which controls voluntary movement, and three in the somatosensory cortex, which processes touch. The procedure made him the first person to undergo what the team calls a double neural bypass.

The “double” refers to the system’s two branches, Bouton says: one connects the brain to the body, while the other connects the brain to the spinal cord. Together, they let information travel both ways—decoding Thomas’s intention to move and delivering sensory feedback to his brain.

On the outgoing side, the two motor arrays pick up the electrical firing of individual neurons as Thomas simply thinks about moving his hand. An AI-driven decoder translates those patterns into an intended action, such as opening or closing his hand. That command can then activate electrical stimulation patches worn over his forearm muscles and a 3D-printed hand brace fitted with an artificial-tendon system. The brace closes his index and middle fingers, providing grip strength that his hand muscles cannot generate on their own.

(New spinal cord therapy helps paralyzed patients walk again.)

A separate stimulation patch, worn on the back of his neck and targeted at specific spinal nerve roots, does something entirely different. Rather than moving his hand on command, this spinal stimulation paired with physical therapy, after about 35 weeks, increased the force Thomas could generate with his biceps by 86 percent in his right arm and 62 percent in his left. This allowed him to lift his own hands to his face for the first time since the accident—a gain that has persisted well after the stimulator was turned off.

On the incoming side, electrodes in the sensory cortex send information back into the brain. Force sensors built into the hand brace detect the instant Thomas’s fingers close around something. The system converts that pressure into electrical pulses delivered directly into the patch of cortex that represents his fingertips—recreating, in real time, the feeling of an object in his grip. In one demonstration, the feedback allowed Thomas to feel his sister’s hand in his.

But Bouton considers the real breakthrough came later, when the team devised what they call a “cortical mirroring” technique to recreate lasting sensation in Thomas’s hand—not just moment-to-moment feedback, but a change that even outlasts the device being switched off. 

After months of spinal-cord stimulation alone had produced no change in Thomas’s sense of touch, Bouton’s team recorded the distinct micro-patterns of brain activity that appear when a person imagines different parts of a hand being touched. Then they played those patterns back into Thomas’s brain using electrical stimulation of his sensory cortex, layered on top of spinal cord and skin-level stimulation. “When we did that, it worked,” Bouton says.

Sensation at Thomas’s wrist improved measurably—within several weeks, roughly tenfold, according to Bouton. “You don’t normally see that three or four years past injury,” he says, particularly below the level of the injury, where Thomas’s improvement occurred. For Thomas, the return of touch has mattered as much socially as physically—the ability to register another person’s hand in his again. “It’s unreal,” he has said of the feeling, “because I haven’t felt that in years.”

How close is this technology to wider use?

Outside researchers see real significance in these results, but they also urge caution. Daniel Rubin, a critical care neurologist at the Mass General Brigham Neuroscience Institute, calls the study “a tour de force”—not because any single element is new, but because three techniques that had each worked in isolation had never been combined before. More striking to Rubin is the suggestion that stimulation can durably change the nervous system itself—evidence, he says, of “neuro-recovery, restoring a person’s own ability to move,” rather than a device simply standing in for lost function.

“I think a lot of us had a sense all along that this had to be true,” Rubin says. “We know the brain maintains plasticity. The fact that we remember what happened yesterday means something changed in our brains.”  What the result adds up to, in his view, is functional independence—being able to do things for yourself that you could do before your injury.

John Downey, a BCI researcher at the University of Chicago, is more skeptical about what, precisely, produced Thomas’s gains. The recent study reveals the cortical sensory electrodes were placed in a region that maps more naturally to the face than to the wrist, he notes, making it difficult to attribute the lasting sensation to plasticity in that specific circuit. He suspects the spinal-cord stimulation, paired with stimulation of nerves near the skin, may be doing more work than the brain implant does. “It's not clear that they're being used together in the most productive way,” he observes.

(Scientists are learning how to interrupt pain before it forms.)

Both researchers, and Bouton himself, return to the same caveat: this is a study of one person. Thomas was in his 40s and otherwise healthy when he enrolled—younger, and with an injury more recent, than many people living with paralysis, which may make him unusually well positioned to benefit. Larger, industry-sponsored trials now underway will determine how broadly the approach generalizes. “You have to start somewhere,” Bouton says, “but we don’t want people to think we’ve cured paralysis.”

One component of the broader approach is closer to commercial use. Bouton is the founder and CEO of Neuvotion, which plans to launch NeuStim at a limited number of clinical sites this fall. Unlike the system implanted in Thomas, NeuStim does not read signals directly from the brain. Instead, the noninvasive wearable uses sensors and AI to infer intended movement from subtle body-language cues, then stimulates the appropriate muscles—for example, helping a hand open to grasp an object. The FDA cleared NeuStim as an external functional neuromuscular stimulator. Bouton says the initial clinical systems will cost roughly $25,000, with a less expensive home version planned for late 2027.

Several other companies—Neuralink, Paradromics, and Synchron among them—are running early human trials of implanted brain-computer interfaces; most aim to restore communication for people who cannot speak, a goal researchers consider nearest to market. Rubin’s own lab recently helped a paralyzed man type 22 words a minute using only his thoughts, faster than his pre-injury typing speed—a result he attributes to the man’s brain learning an entirely new skill, the way one learns a piece of music.

For Thomas, the science ahead matters less than what he’s already achieved. “A lot of people don’t know it’s out there,” he says of the technology that gave him back parts of his arms. “I’d never even heard of BCI before this.” But asked whether it was worth it—the surgeries, the years of rehabilitation—he doesn’t hesitate. “Yeah,” he says. “It’s been worth it.”