Alex Smith He was 11 years old when he lost his right arm in 2003. A drunk driver driving a boat crashed into a family’s boat on Lake Austin, sending him overboard. He was hit by a propeller and his arm fell into the water.
A year later, he received a myoelectric arm, a type of prosthesis powered by electrical signals from the residual limb muscles. However, Smith rarely used it because it was “very, very slow” and had limited range of motion. He could open and close his hands, but couldn’t do much else. He’s tried other robotic arms over the years but had similar problems.
“They’re just not very functional,” he says. “There is a huge delay between executing a function and then having the prosthesis actually perform that function. It’s faster to find different ways to do things in everyday life.”
Recently, he’s been testing a new system from Phantom Neuro, an Austin-based startup, that could provide more realistic control of prosthetic limbs. The company is creating thin, flexible muscle implants that allow amputees a wider, more natural range of motion simply by thinking about what they want to do.
“Not many people use robotic limbs because the control systems are so poor,” says Connor Glass, CEO and co-founder of Phantom Neuro.
In data shared exclusively with WIRED, 10 participants in a study conducted by Phantom used a wearable version of the company’s sensors to control a robotic arm already on the market, achieving an average accuracy of 93.8% across 11 hand and wrist gestures. I did it. Smith was one of the participants, and the other nine were healthy volunteers, a common occurrence in early prosthetics research. The success of this research paves the way for future testing of Phantom’s implantable sensors.
Current myopotential prostheses, like the one Smith attempted, read electrical impulses from surface electrodes placed on the amputated stump. Most robotic prostheses have two electrodes or recording channels. When a person bends their hand, the arm muscles contract. These muscle contractions still occur when upper limb amputees bend. The electrodes pick up the electrical signals from the contraction, interpret them, and initiate movement of the prosthesis. However, surface electrodes do not always capture stable signals because they can slip or move, which reduces accuracy in real-world environments.
