AI and a brain implant restored a paralysed man's movement and touch
TNW
A paralysed man regained hand movement and touch using a brain implant paired with AI and electrical stimulation. The gains stuck around for years, suggesting his nervous system actually rewired itself.
Based on reporting by TNW — read the original for the full story.
Summary, retelling and take written by AI under human oversight; images are AI-generated illustrations. How we work · Report an error
Keith Thomas broke his neck diving in 2020 and ended up with complete tetraplegia, unable to even lift his hands to his own face. Thirteen months after the accident, he signed up for a three-year trial at the Feinstein Institutes for Medical Research, Northwell Health's research arm, testing something the team calls a double neural bypass. It's a mashup of a brain-computer interface, AI decoding software, and electrical stimulation aimed at both the spinal cord and the brain itself.
The results, published in Nature Medicine, are hard to shrug off. Over 35 weeks of training, Thomas's right arm got 86% stronger and his left 62% stronger. He can feed himself, drink from a cup, scratch his nose, wipe his mouth. None of that was possible before. A companion technique called cortical mirroring went after his sense of touch, and after roughly 25 weeks he started feeling sensation again in a wrist that had been numb since the injury.
What makes this stand out isn't just the movement or the feeling coming back. It's that much of it stuck around after the stimulation was switched off. On a follow-up more than two years later, the gains were still there. Chad Bouton, the study's corresponding author, put it plainly: they're not simply routing around the injury, they're rewiring the nervous system. That's neuroplasticity doing real work, not a temporary crutch.
The mechanics behind it are wild. Surgeons spent 15 hours implanting five microelectrode arrays in Thomas's brain. AI reads his intended movements and fires stimulation into his forearm muscles to move his actual hand. A 3D-printed brace loaded with sensors feeds signals back into his sensory cortex, recreating the sensation of touch. The decoder kept 84.6% accuracy across five months without needing retraining, and Thomas managed to lift empty eggshells without cracking them 87% of the time — even mid-conversation, which says something about how automatic this has become for him.
This lands in a crowded, fast-moving field. Other groups are chasing speech restoration through implants, others are betting on wearable or non-invasive setups, and China has already cleared its first commercial brain implant. With roughly 15 million people worldwide living with spinal cord injuries, and hand function topping the wish list for most people with tetraplegia, it's easy to see why every advance here gets attention. The Feinstein team says bigger trials are next, along with tests of the system for other conditions like stroke.
My take — AI-written commentary, not fact-checked reporting
The two-year durability is the actual headline here, not the eggshell trick — a lot of neurotech demos look great on stage and vanish the moment the stimulation stops. If this rewiring effect replicates in a larger trial, it changes the conversation from assistive gadget to genuine therapy, which is a much bigger deal than another flashy BCI headline.
Read more about this at: TNW