A tiny, paperclip-sized implant lets paralyzed people control exoskeletons and prosthetic limbs, just by thinking about it. The new device comes from the University of Melbourne, in partnership with DARPA, and is set begin medical trials next year.
The implant, called a stentrode (stent-based electrode), doesn’t even need to be placed into your brain to do its work. It sits in a vein next to the brain’s motor cortex and detects neural activity. In fact, this lack of open brain surgery is the device’s main selling point, says Thomas Oxley of Melbourne University’s bionics lab, the paper’s lead author.
The trick is recording the brain activity and translating it so that it can be used for controlling limbs and exoskeletons. Existing research into interpreting brainwaves has focused on how they look to electrodes jabbed into the brain itself. One possible complication is that blood rushing through the vein could interfere with the readings, “but we showed this isn’t true,” says Oxley. After implanting the stentrode, it is absorbed into the vein’s wall, where it can start picking up the brain’s signals at the same strength as signals from implanted electrodes.
“The fact that our device can record signals up to 190 hertz is the most exciting finding in our Nature Biotechnology paper,” Oxley told Melbourne University’s Pursuit blog. “The data between 70 to 200 hertz is the most useful for brain machine interfacing.”
Fitting the device is a little like squeezing a ship into a bottle. The basket-like outer layer is made from nitinol, a flexible alloy of nickel and titanium, also used in bra underwires and spectacle frames. The basket collapses and is delivered using a catheter. Once in place, the catheter is removed and the basket unfurls like the sails of the now-bottled ship.
The challenge now is in interpreting the electrical signals from the roughly 10,000 neurons that fire close to the stentrode. While the interfaces will be designed to take these signals and translate them, much of the work will need to be done by the wearer. They’ll have to learn what kind of thoughts will make the machines move. Oxley likens it to learning to play the piano.
“You know that your hands are physically capable of playing it,” he says, “but you don’t understand the sequence in which the keys have to be struck. It will take time to use your hands to learn how to play the piano. With our device, you’ve essentially connected an electronic limb to the patient’s brain, but they have to learn how to use it.”
In 2017, the first implants will be given to paraplegic and quadriplegic patients at the Royal Melbourne Hospital, and if successful, the earliest we could expect to see these in the wild would be 2022. But if they work, then the non-invasive nature of implanting the stentrodes could make them even easier to fit than installing one of Oxley’s pianos into your apartment, giving many more people the chance to learn how to “play.”