Researchers from Stanford University (USA) have just developed a brain-computer interface (BCI) device capable of decoding sentences that are thought but not spoken by the user.
In the new experiment, researchers implanted tiny microelectrode arrays in the motor cortex – the area responsible for speech – to record neural signals. Volunteers were then asked to either try to speak or imagine saying a series of words. Both actions activated overlapping brain regions and produced similar brain activity, albeit to different degrees.
Next, they trained artificial intelligence (AI) models to interpret the words that volunteers thought but did not say. The results showed that the brain chip could interpret the sentences that volunteers imagined in their heads with up to 74% accuracy.
Researchers from Stanford University (USA) are testing microelectrodes Photo: STANFORD UNIVERSITY
The team says the technology could one day help people who can't speak communicate more easily. "This research really gives hope that speech BCI could one day restore the ability to communicate as fluently, naturally and comfortably as normal speech," said Frank Willett, a neurosurgeon at Stanford University and a member of the team, as reported by Euro News.
BCIs work by connecting a person’s nervous system to devices that can decode their brain activity. This allows the user to perform actions, such as controlling a computer or moving a prosthetic hand, using only their thoughts. One of the most well-known devices is the Neuralink brain chip developed by American billionaire Elon Musk and his colleagues.
Source: https://nld.com.vn/them-thanh-tuu-ve-giao-dien-nao-may-tinh-196250818210702909.htm
Comment (0)