A single brain implant lets paralysed patients speak and gesture through a digital avatar

Researchers at UCSF demonstrated a brain-computer interface that decodes both speech and hand gestures from a single brain implant and animates a digital avatar to represent the user. In a proof-of-concept with two paralysed participants, a small set of gestures and spoken words were accurately decoded using machine learning, with the team aiming to expand vocabulary and continuous movements. The work could eventually help millions with speech loss from stroke, ALS and other brain conditions, but ethical, surgical, and generalisation challenges remain, as the approach currently relies on patient-specific models and invasive procedures. The study adds to a growing field alongside companies like Neuralink and Echo Neurotechnologies.