Scientists at the University of California, Davis fitted a 45-year-old man with ALS with a neuroprosthesis, achieving a speech accuracy of 97.5%.
Five years after the onset of the disease, he received four microelectrode arrays implanted in the left ventral precentral gyrus. These record neuronal activity from 256 intracranial electrodes. The system decodes signals from the brain’s cortex during attempted speech and sends data to a screen. Speech is then vocalized using text-to-speech conversion. The program is designed to replicate the patient’s voice as it was before ALS.
On the first day of use—25 days post-surgery—the neuroprosthesis demonstrated 99.6% accuracy with a 50-word vocabulary. System calibration took 30 minutes. On the second day, after 1.4 hours of training, accuracy reached 90.2% with a vocabulary of 125,000 words. Over the following 8.4 months post-implantation, accuracy improved to 97.5%.
“In a person with ALS and severe dysarthria, the intracortical speech neuroprosthesis achieved a level of effectiveness that allows conversational speech recovery after brief training,” the study results stated.
Amyotrophic lateral sclerosis is a progressive, incurable degenerative disease of the central nervous system, affecting motor neurons and leading to paralysis and subsequent muscle atrophy.
Neurosurgery startups are striving to combat ALS and paralysis by integrating brain-computer interface (BCI) systems. The most notable example is Elon Musk’s Neuralink, which recently implanted a neurodevice in a second patient.
In June, Synchron integrated OpenAI’s AI into its BCI, allowing patients with severe motor impairments to send more contextual messages using thoughts.
Later, the company connected the brain implant to Apple’s mixed reality headset, Vision Pro.