An ALS Patient Set a Record For Communicating Via a Brain Implant: 62 Words Per Minute (technologyreview.com) 17
An anonymous reader quotes a report from MIT Technology Review: Eight years ago, a patient lost her power of speech because of ALS, or Lou Gehrig's disease, which causes progressive paralysis. She can still make sounds, but her words have become unintelligible, leaving her reliant on a writing board or iPad to communicate. Now, after volunteering to receive a brain implant, the woman has been able to rapidly communicate phrases like "I don't own my home" and "It's just tough" at a rate approaching normal speech. That is the claim in a paper published over the weekend on the website bioRxiv by a team at Stanford University. The study has not been formally reviewed by other researchers. The scientists say their volunteer, identified only as "subject T12," smashed previous records by using the brain-reading implant to communicate at a rate of 62 words a minute, three times the previous best. [...] People without speech deficits typically talk at a rate of about 160 words a minute. Even in an era of keyboards, thumb-typing, emojis, and internet abbreviations, speech remains the fastest form of human-to-human communication.
The brain-computer interfaces that [co-lead author Krishna Sehnoy's] team works with involve a small pad of sharp electrodes embedded in a person's motor cortex, the brain region most involved in movement. This allows researchers to record activity from a few dozen neurons at once and find patterns that reflect what motions someone is thinking of, even if the person is paralyzed. In previous work, paralyzed volunteers have been asked to imagine making hand movements. By "decoding" their neural signals in real time, implants have let them steer a cursor around a screen, pick out letters on a virtual keyboard, play video games, or even control a robotic arm. In the new research, the Stanford team wanted to know if neurons in the motor cortex contained useful information about speech movements, too. That is, could they detect how "subject T12" was trying to move her mouth, tongue, and vocal cords as she attempted to talk?
These are small, subtle movements, and according to Sabes, one big discovery is that just a few neurons contained enough information to let a computer program predict, with good accuracy, what words the patient was trying to say. That information was conveyed by Shenoy's team to a computer screen, where the patient's words appeared as they were spoken by the computer. [...] The current system already uses a couple of types of machine learning programs. To improve its accuracy, the Stanford team employed software that predicts what word typically comes next in a sentence. "I" is more often followed by "am" than "ham," even though these words sound similar and could produce similar patterns in someone's brain. Adding the word prediction system increased how quickly the subject could speak without mistakes.
The brain-computer interfaces that [co-lead author Krishna Sehnoy's] team works with involve a small pad of sharp electrodes embedded in a person's motor cortex, the brain region most involved in movement. This allows researchers to record activity from a few dozen neurons at once and find patterns that reflect what motions someone is thinking of, even if the person is paralyzed. In previous work, paralyzed volunteers have been asked to imagine making hand movements. By "decoding" their neural signals in real time, implants have let them steer a cursor around a screen, pick out letters on a virtual keyboard, play video games, or even control a robotic arm. In the new research, the Stanford team wanted to know if neurons in the motor cortex contained useful information about speech movements, too. That is, could they detect how "subject T12" was trying to move her mouth, tongue, and vocal cords as she attempted to talk?
These are small, subtle movements, and according to Sabes, one big discovery is that just a few neurons contained enough information to let a computer program predict, with good accuracy, what words the patient was trying to say. That information was conveyed by Shenoy's team to a computer screen, where the patient's words appeared as they were spoken by the computer. [...] The current system already uses a couple of types of machine learning programs. To improve its accuracy, the Stanford team employed software that predicts what word typically comes next in a sentence. "I" is more often followed by "am" than "ham," even though these words sound similar and could produce similar patterns in someone's brain. Adding the word prediction system increased how quickly the subject could speak without mistakes.
RIP, Dr. Shenoy (Score:2, Informative)
The release of this preprint that will no doubt become a landmark paper should not be done without recognition that Dr. Shenoy very recently passed away, far too young, after a career of significant contributions to science that outpaced his peers.
RIP, Dr. Shenoy.
Re: (Score:3)
Speaking of which, many of his inventions are licensed by Neuralink -many of its employees come from his lab, and he was an advisor to neuralink .. but Musk didn't even acknowledge the guy dying. Reference: https://npsl.sites.stanford.ed... [stanford.edu] also https://twitter.com/shenoystan... [twitter.com]
Today is dunk on Elon day? (Score:4, Funny)
First Mercedes announces they have a Level 3 certified car, then these dudes at Stanford announce they have a brain implant that beat Neuralink to brain-controlled-typing.
Re: (Score:3)
To complete the trifecta, we would need something like spinlaunch actually managing to deliver a satellite or similar
Re: (Score:1)
Almost like the guy got lucky by selling off companies and hiring yes men. Elon didn't create PayPal or Tesla either.
Re: (Score:2)
Reference: https://twitter.com/ChrisMayfi... [twitter.com]
This is incredibly cool (Score:5, Interesting)
My dad died of ALS back in the 90s. For the first several years after he fell ill, he could only communicate with the help of a person holding a spelling board. The helper would slide her finger along the spelling board until Dad blinked - first picking the row of letters, then signifying the specific letter he wanted. You can imagine how slow and at times frustrating that process could be for everyone involved.
Later, the VA bought him a personal computer running some specialized software he could operate with a chin switch. It was basically a digitalized version of the old spelling board, hooked up to some software that drove a voice synthesizer. Despite all the problems (still really slow, and the chin switch was incredibly finicky), it seemed like a huge leap forward - but this new solution could be absolutely game changing.
Let me say - non-ironically, for once - truly, we live in the future. I wish dad could have heard about this.
Re: (Score:3)
Indeed it is cool My late wife also died from ALS. By the time she fell to it, they had eye-gaze computers. But it was still a pain keeping it calibrated, and typing with her eyes was slow as heck.
I wish this had been available even just 10 years ago.
Re: (Score:3)
My father also died with ALS in the early 90s. Most of his issues were physical muscle control, no voice issues etc (which is good I guess since both parents are/were speech pathologists...) I guess dad got lucky - all of the men on his side of the family had massive heart attacks in their early 70s, and I'm fairly sure that is what actually killed him...
Anyway, came here to post a hearty "Fuck ALS" for the rest of us that have lost a loved one to this nasty disease....
Control an exosuit (Score:3)
Next what we need is to be able to control a mech or an exosuit with this. If it can do 62 words per minute that easily ought to be enough to control an exosuit with simple commands and some AI.
How long does it last? (Score:1)
I've seen stories similar to this for about 15 years. There is some early success, but the "small pad of sharp electrodes embedded in a person's motor cortex" eventually cause the neurons they come in contact with to die and the device stops working.
Re: (Score:2)
Well, the guys who made the array for this research claim to have it working for at least 7 years in a bunch of patients. Reference: https://www.globenewswire.com/... [globenewswire.com]
They also recently announced a more advanced version that should be a lot more durable too.
Even if it is just 7 years, that ought to be good enough, minor brain surgery every 7 to 10 years vs. being almost locked in .. choice seems to be a no brainer.
Re: (Score:1)
Interesting. I looked at their website, more details there;
https://www.medicaldesignandou... [medicaldes...urcing.com]
The devices have 100 electrodes, and they typically will stick 4-6 of them into your brain. The electrodes are 0.5-1.5 mm long.
fastest human to human communication (Score:1)
Is speech really "the fastest form of human to human communication". A photograph is significantly faster then speech at describing something. If keyboards are allowed in the context, so is a camera.