Researchers at University of California San Francisco (UCSF) have developed a system that translates signals from a speech-impaired person’s brain to the vocal tract into words that appear as text on a screen. This has helped a man with severe paralysis communicate in sentences.
(Subscribe to our Today's Cache newsletter for a quick snapshot of top 5 tech stories. Click here to subscribe for free.)
“This is the first successful demonstration of direct decoding of full words from the brain activity of someone who is paralysed and cannot speak,” said Edward Chang, MD and neurosurgeon in UCSF.
The team tested the system on a man in his late 30s who suffered a brainstem stroke 15 years ago that damaged the connection between his brain and vocal tract and limbs. They worked with him to create a 50-word vocabulary from his brain activity using advanced computer algorithms.
They surgically implanted a high-density electrode array over the participant's speech motor cortex. When he attempted to say each of the 50 words the electrodes recorded brain signals from his speech cortex.
Also Read | A brain implant in paralysed person helps control computer cursor
The team used custom AI models to translate the recorded neural activity into words. When the participant attempted to speak, these models could detect which words he was trying to say. These words were decoded from his brain activity, one by one, on a screen.
The system was able to decode up to 18 words per minute with up to 93% accuracy.
However, the new system translates signals intended to control muscles of the vocal system for speaking, rather than signals to move the arm or hand to enable typing.
The new method taps into the natural and fluidic aspects of speech and promises more rapid and organic communication, Chang said.
Also Read | Musk says Neuralink will enable people with paralysis to use smartphone with their mind
The team estimates that with speech, information is communicated at a very high rate, up to 150 or 200 words per minute.