Unique Brain Implant Expands Communication Prospects

Scientists from the University of Duke have developed a speech prosthesis that can turn brain signals into words, opening up new opportunities for people with neurological disorders such as lateral amyotrophic sclerosis or prisoner syndrome. The existing methods of communication for such patients are limited and ineffective, with current methods of decoding speech reaching only 78 words per minute, compared to the average speaking speed of about 150 words per minute.

To address this issue, Professor of Neurology Gregory Kogan and specialist in biomedical engineering Jonathan Viventan created a new implant. The implant consists of 256 microscopic sensors on a flexible medical plate, similar in size to a postage stamp, enabling accurate differentiation between signals from neighboring brain cells.

In clinical trials, participants were instructed to repeat meaningless words while their brain activity was recorded by the implant. The data was then processed by a machine learning algorithm, which determined the corresponding sound based solely on the recorded brain activity. The decoder achieved an accuracy rate of 40%, a significant accomplishment considering similar technologies typically require much more time to collect data.

The study’s leading author, Sushendrakumar Duraiil, a graduate student in biomedical engineering, expressed excitement about the wireless version of the device, which has already received $2.4 million in funding from national healthcare institutions.

Despite the researchers’ confidence in their progress, the commercial use of the speech prosthesis is still a distant prospect. The device currently operates slower than natural speech, but there is a visible trend of improvement, according to Viventan.

/Reports, release notes, official announcements.