The UC Davis neurointerface allowed a patient with ALS to transmit nearly 2 million words over 19 months of home use
Researchers from the University of California, Davis have presented the results of long-term home use of a speech neurointerface by a patient with amyotrophic lateral sclerosis (ALS). Over 19 months, Casey Harrell transmitted 183,060 sentences, equivalent to 1,960,163 words, at an average speed of 56 words per minute. This is not a laboratory demonstration, but a real, everyday application of the system without the presence of researchers.
Harrell used the neurointerface at home for over 3,800 hours. The system allowed him to communicate with family, friends, colleagues, and doctors, send messages, participate in video calls, use the internet, and maintain full-time employment despite paralysis. According to the patient's assessment, 92% of sentences were decoded by the system as at least "mostly correct." In formal tests with words displayed on a screen, accuracy exceeded 99% with a vocabulary of 125,000 English words, and the peak rate reached 99.2%.
Technical Details of the System
In 2023, Harrell had four microelectrode arrays implanted in the left precentral gyrus—a brain region associated with speech coordination. The system reads signals from 256 cortical electrodes when the patient attempts to speak. An algorithm converts neural activity into phoneme probabilities every 80 ms, and a language model selects the most likely word sequence from a vocabulary of approximately 125,000 words. The text is displayed on a screen in real time, and after completing a phrase, the system can vocalize it using a synthesized voice tuned to Harrell's voice before his illness. A cursor decoder was also used to control the computer, demonstrating the feasibility of implementing both modes through signals from the speech motor cortex.
From Laboratory Tests to Everyday Life
The team had previously demonstrated high accuracy in laboratory sessions. In 2024, during the first 30-minute training session, the system achieved 99.6% accuracy with a vocabulary of 50 words, and on the second day, 90.2% with 125,000 words. The new work shifts the focus to long-term use in daily life. Co-author of the study, Sergey Stavisky, noted that 3,800 hours of recorded brain activity constitute the largest individual dataset with resolution at the level of individual neurons. However, the authors emphasize that this is a single clinical case, and it is not yet known how transferable the results are to other patients. The system remains experimental, uses wired connections, and requires daily setup by trained assistants.
My comment: This breakthrough demonstrates that neurointerfaces are transitioning from conceptual demonstrations to real tools that change people's lives. However, the key challenge remains scaling the technology—the need for individual calibration and limitations on implant types currently hinder widespread adoption. Nevertheless, the pace of progress is impressive, and we will likely see commercial solutions within 5-7 years.