California scientists announced on Wednesday that they have developed and successfully tested an experimental brain implant that translates neural signals into words that appear on a computer screen.

In a report published in the American Wall Street Journal (WSJ), writer Rolf Winkler said that this achievement - whose various details were described in a research paper published in the New England Journal of Medicine - represents a step towards a technology that would To enable people who have lost the ability to speak to speak through thinking, and it provides a glimmer of hope to thousands of people who annually lose the ability to speak as a result of injury or disease.

The scientists used "speech neuroprosthesis", a technology that allows the conversion of infinitesimal electrical signals from the brain into actions in the sensory world, such as through speech, writing or moving a computer cursor.

However, this technology is still in its infancy, and in recent years has attracted the interest of academic scientists and technology companies, including Elon Musk's Neuralink Corp., kernel, and Facebook.

neural implant

The writer mentioned that Facebook is among the sponsors of the new research, and he said in a blog post that it was eager to develop wearable devices that do not require surgical intervention that would help people write through thinking.

To test this neural implant, researchers at the University of California, San Francisco, used a man in his late thirties who lost the ability to speak more than 15 years ago as a result of a stroke that left him paralyzed.

This man - who communicates using a pointer attached to a hat on his head to indicate letters appearing on a screen - has agreed to undergo surgery to attach small rectangular electrodes to the outer cortex of the brain.

Over a period of 81 weeks and 50 separate sessions, the researchers attached a computer to the electrodes to record the man's brain activity while he observed individual words displayed on a screen and imagined himself saying them out loud.

In the paper, the researchers said they were able to pinpoint every word the man was thinking with 47% accuracy each time.

The accuracy increased to 76% when the scientists added word prediction algorithms to the system, which are similar to the automatic suggestion feature in e-mail and word processing programs. The study was limited to a vocabulary of 50 words.

Dr Eddie Chang, a neurosurgeon at the university and lead author of the study, said: 'To our knowledge, this is the first successful trial of a direct decoding process of whole words from the brain activity of a paralyzed and aphasic person. ".

And the author quoted Amy Orsborn, an assistant professor of bioengineering at the University of Washington - who was not involved in the study - as saying that in addition to proving that the area responsible for speech in the brain continues to function even after years of aphasia, the new study shows that computers can You learn to encode whole words from your brain activity and not just letters.

Orsborn noted that devices capable of doing this could one day speed up communication for people who have lost the ability to speak, many of whom write words by tapping each letter on assistive devices as the man in the study does.

margin of error

However, Dr. Orsborn and other experts caution that the system has a high margin of error, is limited to a limited set of vocabulary, and the significant time it takes to train the system to recognize imagined words is among the reasons this technology has a long way to go before A practical device is available.

This new study comes two months after researchers at Stanford announced the development and successful testing of a similar system that enabled a paralyzed man to "type" 90 letters per minute with an accuracy of 94% and 99% when using word prediction algorithms. Implanted in the brain instead of fixed in the outer cortex - in a research paper published last May in the journal "Nature".

According to the University of California, the man who participated in the study wants his identity to remain anonymous, and it is noteworthy that the device he used in the framework of the study is experimental and cannot be used on a daily basis.