His Entire Body Was Paralyzed. Then They Put AI-Powered Chips in His Brain

"It was a Sunday afternoon," Keith Thomas, a 45-year-old Long Island native, told us of his accident. "I dove into the wrong side of the pool, and I blacked out."

The next thing he knew, Thomas says, he was being airlifted to a nearby hospital; it was July 2020, just a few months into the pandemic, and he'd badly broken his neck at the C4 and C5 vertebrae of his spine. He's been paralyzed from the neck down since, unable to move or feel his limbs — until a few months ago, that is, when a first-of-its-kind clinical trial brought both movement and feeling back to his arms and hands for the first time in three years.

Thomas, who lives with quadriplegia, was the first patient to receive what his doctors are calling a double neural bypass, a new bioelectrical therapy pioneered at Northwell Health's Feinstein Institutes for Medical Research. Led by Chad Bouton, a professor at Northwell's Institute of Bioelectronic Medicine, the experimental new procedure involves a combination of AI, brain-computer interface (BCI) implants, external computers, and non-invasive wearable tech.

Like a coronary bypass surgery creates a detour for your heart to pump blood around an obstacle, a neural bypass uses a combination of machine learning and electrical signaling to reroute an individual's neural signals, avoiding whatever barrier is preventing them from making it where they're supposed to go. A double neural bypass, then, reroutes the signal in just one but two places: in this case, the areas responsible for movement and touch.

The goal? To answer an elusive question: how do you restore the communication between the brain and the body, when the two can no longer speak?

"It's a very challenging problem," said Bouton, who's also the founder and CEO of a biotech firm called Neuvotion, over a video call. "You're looking at these complex electrical patterns in the brain, and you're trying to make sense out of the patterns and extract information from them. We want to know when someone's thinking about moving their hand, or moving their fingers, and we want to be able to then channel those thoughts into something useful."

Bouton and his team refer to this approach as "thought-driven therapy," in which chips embedded in the patient's brain use machine learning to interpret the complex language of neurons. Does it sound like sci-fi? Absolutely. But so far, it's showing unmistakable promise — and the implications for the millions worldwide who suffer from paralysis or movement impairment could be significant.

"It's frustrating when someone looks at their limb, and they can't make the movement they want to make," Bouton said. "They're trying, and the brain knows they're trying, but things aren't happening. It's super frustrating, and it can be depressing."

The professor and his team performed the world's first single neural bypass surgery back in 2016, successfully restoring movement in the arms of a patient who had broken his neck on a family vacation six years prior. But while that procedure was able to reestablish the ability to move — when hooked up to a computer, that is — it didn't bring back the patient's sense of feeling.

Now, seven years later, the double neural bypass has been designed to do both: bring back movement and sensation.

In Thomas' case, he first had to spend months staring at simulated arm and hand movements on a computer screen, urging his brain — unsuccessfully, at the time — to mimic the motions. The doctors and engineers, meanwhile, took detailed MRIs of his brain, mapping the areas responsible for arm movement and hand touch. (Like searching for a needle in an extremely delicate, blood vessel-laden haystack, Bouton told us.)

Armed with this data, the doctors then hatched a plan to implant a total of five BCI chips: two at the area of the brain that presides over movement, and three at the region responsible for touch and feeling in the fingers. The chips pass decoded bioelectrical messages to the computer, which then sends electric signals to a series of electrode-laden patches placed across Thomas' spine and forearms. Finally, a handful of infinitesimal sensors placed on Thomas' fingertips and palms send touch and pressure data back up to the sensory region of Thomas' brain.

"Every time he thinks about moving and feeling, we actually send another signal to the spinal cord, and that supercharges the spinal cord," said Bouton. "It tries to strengthen connections."

Installing the chips was no small feat. Thomas underwent a 15-hour open brain surgery back in March, and as if that wasn't enough on its own, the Long Islander was awake for large portions of the procedure, verbally relaying the sensations he was feeling back to Bouton and his surgeons, a team led by Northwell neurosurgeons Ashesh Mehta and Netanel Ben-Shalom.

But Thomas "didn't really have any reservations" about the surgery, he recalled, before conceding: "until the night before."

Fortunately, the procedure was a resounding success. The BCI install went off without a hitch, and for the first time since his accident, Thomas was able to hold — and feel — his sister's hand.

"It was incredible," Bouton recalled. "It still makes me tear up."

In the four months since the procedure, Thomas has regained full strength in both arms, even experiencing a 110 percent recovery in his right arm. But most excitingly, Thomas has started to experience natural recovery in his forearm and wrist — meaning that the therapy might have kickstarted his nervous system's innate healing processes.

"Only several months into the study, he's making huge gains," Bouton said, "doubling his arm strength, and starting to feel new sensations in his forearm and even wrist even after he goes home outside the lab, even when we turn [the computer] off."

When we reached out to experts in the field, enthusiasm for the procedure's success — and AI's role in it — was palpable.

The surgery is a "novel and exciting advance in the field of both BCI and spinal cord neuroprosthetic interfaces," Dr. Wilson Zachary Ray, Executive Vice-Chair of the Department of Neurosurgery and chief of spine surgery at the Washington School of Medicine in St. Louis, who wasn't involved in the study, said over email. "I suspect this sort of AI and ML innovation will see a massive growth in clinical applications over the next 3 to 5 years."

"At some point in the not too distant future," Ray added, "implantable 'smart technology' will be integrated into the fabric of our daily life, similar to how all view our smartphones today."

But as remarkable as these results are, they're not without caveats. Although Thomas has experienced new sensations outside the lab, the computer needs to be turned on in order for him to be able to move. And as Bouton told us in our interview, the contraption itself isn't exactly minimalist.

"It's kind of like the early heart and lung machine," the professor told us of the contraption. "We've got some parts that are in the body, some parts that are on the laboratory table, and some wearables."

But over time, he says, the goal is to condense the device's size, ideally to the point that it's portable. His company, Neuvotion, is working on a number of non-invasive treatments and devices seeking to restore autonomy to those suffering movement impairment and paralysis, among other applications.

"In the more challenging cases, like Keith's," he added, "combining brain-interface technology with non-invasive devices is powerful."

The recovery also requires a lot of effort for patients — hours-long therapy sessions, visits with specialists — as they relearn how to move and strengthen those movements, one day at a time.

"You have to be really patient, and really dedicated, to want to do this," said Thomas. "It's a lot of work." Recounting his many weekly therapy sessions and visits to specialists, he added: "It's pretty much a full-time job, being quadriplegic."

But Thomas doesn't mind. The "stars aligned" for him to meet Bouton, he says, and seeing the tangible results of his effort has been extraordinary. If his role in this research helps others down the line, according to Thomas, it's all worth it.

"All of the effort that I'm putting in is paying off," he told us. "I realize it's not going to happen overnight, but the little things — reaching up to my chin, being able to touch my other hand, rub my cheek when I have to, call people." He quieted for a second. "It's the little things."

More on paralysis technology: Paralyzed People Successfully Test Brain-controlled Electric Wheelchairs