2023-08-24 09:32:13
pte20230824017 technology/digitization, medicine/wellness
Scientists turn imaginary words into a conversation – but the system is still error-prone
How the signals are captured in the brain (Image: Ken Probst, ucsf.edu)
San Francisco/Berkeley/Stanford (pte017/24.08.2023/11:30)
Using computer-brain interfaces and artificial intelligence (AI), researchers at the University of California San Francisco (UCSF) and the University of California Berkeley enables two women with neurodegenerative disorders to communicate once more. One had suffered a massive stroke that paralyzed all of her muscles, and the other had progressive brain disease. Stroke patient Ann was able to speak and express emotions through a digital avatar. And researchers of Stanford University convert the ALS patient’s brain activity into text.
AI converts signals into words
UCSF neurosurgeon Edward Chang implanted Ann 253 electrodes on the surface of the brain, right where language is produced. They were connected to a computer by cable. The brain signals that the electrodes picked up when Ann was talking in her head were mapped to words using AI. To do this, Ann had to complete a training program so that the AI might interpret the signals accordingly. In total, she learned 1,024 words.
Next, the researchers developed an avatar and used machine learning to teach it facial expressions controlled by Ann’s mind. The avatar’s face moves according to Ann’s thoughts and can express emotions such as joy, sadness, and surprise. The system converts thoughts into spoken text at an average rate of 78 words per minute. The rate of natural conversation among English speakers is around 160 words per minute. However, the error rate is still 25 percent.
Thought words become text
Pat, who got her communication skills back from Stanford scientists, had two tiny sensor arrays implanted on the surface of her brain. Each array contains 64 electrodes. The subsequent process was similar to that of the researchers who treated Ann. Pat manages 62 words per minute with the system.
“This is a proof of concept, not an actual device that people can use in everyday life. But it’s a major step forward in restoring the ability to quickly communicate to people who are unable to speak due to paralysis,” says Francis Willett, neurosurgeon in Stanford.
(End)
1692869964
#ALS #stroke #brings #language
