UCSF Brain Implant Lets Paralysis Patients Speak and Gesture Through Mind-Controlled Avatar

On September 14, Nature Neuroscience published a study showcasing a brain-computer interface that translates speech and movement in two individuals with severe paralysis. The device, roughly the size of an iPhone, is equipped with 253 electrodes and is positioned on the sensorimotor cortex. One participant suffered from a stroke, while the other had amyotrophic lateral sclerosis (ALS). As they attempted to produce phrases and gestures, computer models decoded their neural activity, which was then represented by a digital avatar. The participant with ALS achieved 85% accuracy for gestures and 75% for phrases, whereas the stroke survivor attained 100% accuracy. Co-author Samantha Brosler emphasized the significance of gestures in communication, with future goals including expanding vocabulary and enhancing control for amputees.