Researchers at the University of California developed a BCI that allows people with severe paralysis to communicate using both speech and upper-body gestures through a digital avatar. The system uses one implant containing 253 electrodes positioned on the surface of the sensorimotor cortex to capture neural activity associated with attempted speech and movements. Machine-learning algorithms decode these signals in real time, enabling an animated avatar to speak while simultaneously performing gestures such as waving or nodding. The study involved two participants with severe paralysis and demonstrated that neural signals for speech and movement can be extracted from the same cortical region and translated into coordinated communication.
The research represents a step toward BCIs that reproduce the richness of natural human communication rather than simply generating text, speech, or cursor movements. Because conversation relies heavily on non-verbal cues, combining vocal expression with gestures could make interactions more intuitive and socially expressive for people affected by conditions such as ALS or brainstem stroke. Importantly, the findings suggest that a single neural interface can support multiple functions simultaneously, potentially reducing the need for separate implants dedicated to different tasks. The researchers ultimately envision multifunctional systems capable of restoring a broader range of communication and interaction abilities for people with severe paralysis.
More information: