Brain Signals Converted Into Words 'Speak' For Person With Paralysis (sciencemag.org) 24
sciencehabit shares a report from Science Magazine: A man unable to speak after a stroke has produced sentences through a system that reads electrical signals from speech production areas of his brain, researchers report this week. The approach has previously been used in non-disabled volunteers to reconstruct spoken or imagined sentences, but this is the first demonstration of its potential in the type of patient it's intended to help. The participant had a stroke more than 10 years ago that left him with anarthria -- an inability to control the muscles involved in speech. Researchers used a computational model known as a deep-learning algorithm to interpret patterns of brain activity in the sensorimotor cortex, a brain region involved in producing speech, and 'decoded' sentences he attempted to read aloud.
In the new study, [the researchers] temporarily removed a portion of the participant's skull and laid a thin sheet of electrodes smaller than a credit card directly over his sensorimotor cortex. To "train" a computer algorithm to associate brain activity patterns with the onset of speech and with particular words, the team needed reliable information about what the man intended to say and when. So the researchers repeatedly presented one of 50 words on a screen and asked the man to attempt to say it on cue. Once the algorithm was trained with data from the individual word task, the man tried to read sentences built from the same set of 50 words, such as "Bring my glasses, please." To improve the algorithm's guesses, the researchers added a processing component called a natural language model, which uses common word sequences to predict the likely next word in a sentence. With that approach, the system only got about 25% of the words in a sentence wrong, they report today in The New England Journal of Medicine. With the new approach, the man could produce sentences at a rate of up to 18 words per minute.
In the new study, [the researchers] temporarily removed a portion of the participant's skull and laid a thin sheet of electrodes smaller than a credit card directly over his sensorimotor cortex. To "train" a computer algorithm to associate brain activity patterns with the onset of speech and with particular words, the team needed reliable information about what the man intended to say and when. So the researchers repeatedly presented one of 50 words on a screen and asked the man to attempt to say it on cue. Once the algorithm was trained with data from the individual word task, the man tried to read sentences built from the same set of 50 words, such as "Bring my glasses, please." To improve the algorithm's guesses, the researchers added a processing component called a natural language model, which uses common word sequences to predict the likely next word in a sentence. With that approach, the system only got about 25% of the words in a sentence wrong, they report today in The New England Journal of Medicine. With the new approach, the man could produce sentences at a rate of up to 18 words per minute.
Spazzies (Score:2)
Grandmas (Score:2)
Re: (Score:2)
Well if you want to cut your grandma in half and implant a quad leg robot so she ends up being a part of a dystopian cyberpunk world.
Or should she just be more careful in her old age.
Re: Spazzies (Score:2)
Weird question.
Obviously the key to helping millions of people avoiding slip and fall injuries is to be able to make the tech solution affordable. Just making something that works at all - which has now been achieved - is the first step.
Adapting that working tech into simpler and cheaper versions comes next.
First message received (Score:2, Funny)
"My balls itch!"
Re: (Score:1)
Do you have mental issues?
Re: (Score:2)
I think he does. Unless he is quoting some obscure internet joke.
Re: First message received (Score:1)
Re: (Score:2)
Second message received:
DUPE!!!!
"This sucks" (Score:2)
We have ways of making you talk (Score:4, Interesting)
Not wanting to rain on this remarkable achievement, but part of me shudders to think how all these new neuro brain-to-text technologies could one day be abused. We've seen less invasive varieties of the technology come by Slashdot as well lately. In this case they've tapped into the speech area, but what if you could tap into someone's inner monologue?
Re: (Score:3)
, but what if you could tap into someone's inner monologue?
A few tips are described in the Chaos Walking series [wikipedia.org], like counting all the time.
Re: (Score:2)
...the researchers added a processing component called a natural language model, which uses common word sequences to predict the likely next word in a sentence. With that approach, the system only got about 25% of the words in a sentence wrong,...
The double-edged sword of being given the gift of speech, but with Clippy as your translator..
they are syphoning results of RF Brainscan... (Score:2)
into various "breakthroughs", startups etc:
https://www.reddit.com/r/consp... [reddit.com]
it's just one of the assets.
This is how the game on global level looks ATM:
https://www.reddit.com/r/consp... [reddit.com]
Would it work on people in a coma? (Score:2)
There is ongoing debate about the degree of consciousness of patients in a coma and how to even establish if they are conscious at all.
Could this not be used to determine consciousness then? Even more importantly we might be able to get insights in to that state eg. are they in pain or do they feel pain in the same way?
Re: Would it work on people in a coma? (Score:1)
25% correct? (Score:2)
Man: More mustard!
Computer: Excuse me sir, would you be so kind as to apply a tad more mustard to my ham?
Ah yes (Score:2)
Byte Magazine Circuit Cellar, Steve Ciarcia Cite (Score:2)
I remember reading about this a while back.
Ciarcia, Steve. Mind over matter: Add biofeedback input to your computer. In Ciarcia's Circuit Cellar Volume II. BYTE Publications, Inc., 1981, pp. 74
https://dl.acm.org/doi/abs/10.... [acm.org]
Also June, 1988 Byte p273.
https://vintageapple.org/byte/... [vintageapple.org] Page 273
I'd prefer death (Score:3)
Telepathy tech? (Score:2)