Scientists have taken a step closer to translating brain activity into speech, which could allow patients who can no longer speak to communicate again.
- American researchers have succeeded in developing an artificial intelligence capable of decoding the speech produced by our brain.
- This new artificial intelligence can transcribe sentences in our brains in real time, with 97% success.
Cerebral vascular accident (CVA), Charcot’s disease, confinement syndrome… Many illnesses can deprive patients of speech, despite intact cerebral capacities. To help them, researchers have been working for years on a “speech decoder”, which would translate thoughts into words using a computer connected to brain implants.
Translate thoughts into words
“Dix years after speech was first decoded from human brain signals, the accuracy and speed of artificial speech still remains far below that of natural speech”, explain Joseph Makin, David Moses and Edward Chang, of the University of California at San Francisco (USCF) in the United States. Their work, presented in the journal Nature Neuroscience, aim to streamline the process. Thanks to the new artificial intelligence (AI) system they have developed, sentences in the brain can now be decoded in real time, with 97% accuracy. “The average word error rate does not exceed 3%”, welcome the scientists. “We’re not there yet, but we think it could be the basis for a speech prosthesis.”summarizes Dr. Joseph Makin.
To achieve these results, his team recruited four Americans and then fitted them with brain implants originally intended to monitor epileptic seizures. The patients then repeated several times series of 30 to 50 sentences aloud, containing up to 250 different words, for about forty minutes. The data collected for each sentence was fed into a machine learning algorithm which transformed it into a string of numbers, which in turn were converted into English sentences.
Eventually, some also aim to equip healthy humans with this type of brain implant. The famous American Elon Musk has invested 150 million dollars for this purpose to develop his Neuralink device, while the European Commission is massively funding the BrainCom project.
.