Story at a glance:
- An experimental brain implant turns brain signals into text-to-speech words.
- It is being backed by tech heavyweight investors like Elon Musk’s Neuralink Corp., Kernel and Mark Zuckerberg's Facebook Inc..
- The study was limited to 50 vocabulary words that elementary school students would be familiar with.
Researchers from the University of California, San Francisco (UCSF) unveiled on Wednesday their development of an experimental brain implant that turns brain signals into text-to-speech words on a computer.
First published in the New England Journal of Medicine, the technology is supposed to help people unable to speak regain verbal communication, The Wall Street Journal reported.
The speech neuroprosthesis is being backed by tech heavyweight investors like Elon Musk’s Neuralink Corp., Kernel and Mark Zuckerberg's Facebook Inc.
The brain-computer technology sends tiny electrical signals from the brain to be transfigured into speaking words or text.
The UCSF researchers tested their neuroprosthesis on a man in his 30s who was not able to speak due to a massive stroke that paralyzed him 15 years ago.
Placing a cap on his head attached to his brain, the man was able to talk again using his brain to type individual letters on a screen.
More than 81 weeks and 50 separate sessions later, UCSF researchers observed the man’s brain activity with several computers as he watched individual words displayed on a screen and imagined uttering them aloud, The Journal reported.
The researchers were able to accurately determine what words the man was saying 47 percent of the time.
Accuracy rose to 76 percent when the scientists used a word-prediction algorithm similar to common features on email and word-processing programs.
The study was limited to 50 vocabulary words that elementary school students would be familiar with.
“To our knowledge, this is the first successful demonstration of direct decoding of full words from the brain activity of someone who is paralyzed and cannot speak,” Eddie Chang, chairman of the Department of Neurological Surgery at UCSF, said. “It shows strong promise to restore communication by tapping into the brain’s natural speech machinery.”
However, Amy Orsborn, an assistant professor of bioengineering at the University of Washington, and other experts say the systems’ predisposition to errors is reason enough to hold off, depending on the technology being made public until it is trained to recognize the imagined words from the subjects.
In addition to demonstrating that the brain region responsible for speech continues to function even years after the ability to speak has been lost, the new research shows that computers can be taught to decode full words from brain activity and not just letters, said Orsborn, who was not involved in the research.
READ MORE STORIES FROM CHANGING AMERICA