A computer helped a woman find her voice after 20 years of paralysis
[ad_1]
Almost 20 years ago, a paralyzed woman lost her voice. But she was able to find him through a virtual avatar thanks to a brain implant and a mind-reading algorithm. At the same time, another implant helped another woman. She lost her voice due to a rare neurological disease, but was able to communicate via text messages at a speed close to normal speech. Both studies were published in the journal Nature, writes The Wall Street Journal. They point to significant progress in systems that allow people to control devices using brain signals. In the first case, Eddie Chang, a neurosurgeon at the University of California, San Francisco Neuroscience Institute, and colleagues implanted a paper-thin sheet of silicon with 253 electrodes on the surface of the brain of a paralyzed woman. Photo: SEBASTIAN KAULITZKI/Gettyimages.com A 48-year-old patient was unable to speak or use her arms or legs after suffering a brain stem stroke about 18 years ago. Normally, the brains of people with paralysis can emit electrical signals – even when the connections between the brain and muscles are broken. The woman’s implant did just that, trying to pick up signals intended to control the muscles associated with speech: the tongue, jaws, larynx, etc. The electrodes were connected to computers via a cable connected to a port attached to her skull. The scientists trained the algorithm to recognize her brain’s signals about speech and facial expressions. They were asked to silently say words from the screen and imagine facial expressions with emotions. The researchers then tested the algorithm’s ability to translate brain signals into text and speech. The patient was given new sentences to read silently. And the system had to decode brain signals. Read also: Can you lose weight if you think a lot? How our brain burns calories The computer was able to compose sentences at a speed of 78 words per minute. It is 5 times faster than previous interfaces. And with a vocabulary of 1,000 words, the system was accurate 75% of the time. Typically, the patient uses a head-tracking assistive device that types about 14 words per minute. The “brain-computer” interface allowed the woman to communicate with the help of a talking avatar. For personalization, a recording of a woman’s speech at a wedding was used. In the second case, a 68-year-old patient was diagnosed in 2012 with a progressive neurodegenerative disease – Lou Gehrig’s disease. She could not speak clearly. Stanford University researchers connected sensors in the brain to computers that taught them to recognize the correspondence of brain signals to speech. During 4 months of training, the system translated brain signals into text twice a week at a speed of 62 words per minute. With a reserve of 125,000 words, the system was accurate 76% of the time. “This could allow for truly free speech and a real restoration of the ability to communicate with others,” said Dr. Jamie Henderson, a professor of neurosurgery at Stanford University and co-author of the study. He also added that the sensors could produce better results by reading signals from individual neurons. But the brain moves, so there could have been scars around the sensors – this displaced the devices. Earlier we told that in Ukraine, for the first time, a child underwent brain surgery with awakening. Read also: He was paralyzed for 12 years: a man from the Netherlands got back on his feet thanks to a brain implant
[ad_2]
Original Source Link
