A brain-computer interface enables speaking and using a computer after paralysis

A man paralyzed and with severe speech impairments due to amyotrophic lateral sclerosis was able to regain fluid communication and autonomous use of his computer thanks to an innovative brain-computer interface. For nearly two years, he used this system at home, without researcher assistance, accumulating more than 3,800 hours of use. This device allowed him to exchange over 183,000 sentences—nearly two million words—at an average speed of 56 words per minute. He was also able to control his personal computer, send messages, browse the internet, and maintain a full-time professional activity.

The system relies on electrodes implanted in the brain area responsible for speech. These electrodes capture neural activity related to attempts at speech or movement, then translate it into text or commands for the computer cursor. A major innovation lies in the use of a decoder based on a transformer model, an artificial intelligence technology that improves accuracy and reduces the need for daily recalibration. This decoder achieved a 99% accuracy rate for word recognition in structured tests, with a vocabulary of 125,000 words.

For cursor control, another decoder was developed, allowing the user to move the mouse and click simply by imagining hand movements. This system was optimized to reduce calibration time, from several minutes to less than one minute per day, while maintaining high performance. The user could thus switch between eye-tracking control and neural control, depending on their preferences.

The stability of neural signals was a major challenge. Despite the prolonged implant, more than 90% of the electrodes continued to detect reliable neural activity throughout the study. The neural representations related to speech remained stable for at least 19 months, enabling precise and consistent decoding.

The user gradually adopted a “silent speech” strategy—that is, moving the facial muscles without producing sound—which proved less tiring and faster than vocalized speech. This adaptation allowed him to increase his communication speed, from 30 to nearly 50 words per minute, while maintaining high accuracy.

This system was designed to integrate seamlessly into daily life. The initialization and shutdown of the device were automated, allowing the user and their caregivers to manage it with just a few simple steps. Once launched, the system could run continuously for up to 19 hours without intervention.

The user interface was designed to be intuitive and customizable. It allowed for correcting misdecoded sentences, adjusting cursor or click sensitivity, and even activating a privacy mode to disable data recording. Dedicated software was also developed to connect the system to the personal computer, enabling the user to perform actions such as copying and pasting decoded text, browsing the web, or participating in video calls.

This breakthrough demonstrates that it is now possible to restore natural communication and full digital access for people with severe paralysis, paving the way for practical and sustainable assistive technologies.

Documentary Sources / Document Base

Reference Report

DOI: https://doi.org/10.1038/s41591-026-04414-6

Title: Long-term independent use of an intracortical brain–computer interface for speech and cursor control

Journal: Nature Medicine

Publisher: Springer Science and Business Media LLC

Authors: Nicholas S. Card; Tyler Singer-Clark; Hamza Peracha; Carrina Iacobacci; Xianda Hou; Maitreyee Wairagkar; Zachery Fogg; Elena C. Offenberg; Leigh R. Hochberg; Sergey D. Stavisky; David M. Brandman