en.Wedoany.com Reported - Scientists at the University of California, Davis, have successfully helped a patient with amyotrophic lateral sclerosis (ALS) return to full-time work by implanting a neural interface that synthesizes speech from brain activity.
The implant was installed in patient Casey Harrell in 2023 and remains operational. In controlled tests, the system achieved 99% accuracy in reconstructing sentences from brain activity, with daily use accuracy around 92%. Harrell controls a computer cursor with his thoughts, having used the device for over 3,800 hours in the past few years, averaging more than 5 hours per day.
A key difference from previous studies is the system's full autonomy. Earlier research required researchers to accompany patients at home or for patients to visit a laboratory. Harrell's implant can be used in a home environment without the involvement of scientists.
Study participant and neurosurgeon David Brandman stated that this achievement marks a new threshold for neural interface technology. The scientist noted that the patient has returned to full-time work and can hold complete conversations with his daughter, who has never heard his real voice.
Brandman explained that the study did not use specially designed equipment; Harrell interacts with the world through an existing "brain-computer" interface developed by Blackrock Neurotech. The team of scientists upgraded the system using machine learning technology, creating a platform that manages the neural interface through neural decoding. The algorithm converts activity in brain regions responsible for facial, oral, and jaw motor functions into English phonemes. The program then matches phonemes to words and converts words into sentences.
Ultimately, the system enables the patient to perform daily work in the field of environmental protection. Harrell himself described the implant through the system as a technology that allows him to communicate "in a natural way," something no previous device could achieve.
This article is compiled by Wedoany. All AI citations must indicate the source as "Wedoany". If there is any infringement or other issues, please notify us promptly, and we will modify or delete it accordingly. Email: news@wedoany.com
