(Neurology Today) – Researchers have successfully used brain signals to decode the speech of a man with quadriparesis and anarthria, a feat considered a major breakthrough in the use of computer-brain interfaces.
The man is not cognitively impaired but cannot physically produce speech. This is the first time signals directly from the brain, using an electrode array placed on the brain surface, have been used to decode speech as words from such a person.
Researchers from the University of California, San Francisco used an algorithm to identify and decode words from the brain signals, along with language modeling in which words were identified using the context in sentences. They were able to decode speech with a word error rate of 25.6 percent when decoding sentences. CONTINUE