Researchers at the University of California, Davis published in the journal Nature Medicine results of long-term home use of a speech neuron interface by a patient with amyotrophic lateral sclerosis (ASL). Over the course of 19 months, Casey Harrell, paralyzed by the disease, transmitted 1,960,163 words at an average rate of 56 words per minute using the implant.

The system, which reads signals from 256 electrodes in the talking motor cortex, converts neural activity into real-time text and voices it with a synthesized voice similar to that of the patient before illness. Decoding accuracy in formal tests exceeded 99% for a dictionary of 125,000 words. Harrell used the device for more than 3,800 hours, which allowed him to stay fully occupied, communicate with loved ones, and use the internet.

The authors point out that this is one of the most important steps from laboratory demonstrations to practical application of BCI. However, the study describes only one clinical case, and the device itself remains experimental: it requires a wired connection and daily assistance from assistants. Against this background, competitors such as Neuralink are also actively developing their BCI devices, demonstrating success in computer control and music creation by the power of thought.