News: 0176899943

  ARM Give a man a fire and he's warm for a day, but set fire to him and he's warm for the rest of his life (Terry Pratchett, Jingo)

Brain Interface Speaks Your Thoughts In Near Real-time

(Wednesday April 02, 2025 @11:20AM (BeauHD) from the life-changing-tech dept.)


Longtime Slashdot reader [1]backslashdot writes:

> [2]Commentary , [3]video , and a [4]publication in this week's Nature Neuroscience herald a significant advance in brain-computer interface (BCI) technology, enabling speech by decoding electrical activity in the brain's sensorimotor cortex in real-time. Researchers from UC Berkeley and UCSF employed deep learning recurrent neural network transducer models to decode neural signals in 80-millisecond intervals, generating fluent, intelligible speech tailored to each participant's pre-injury voice. Unlike earlier methods that synthesized speech only after a full sentence was completed, this system can detect and vocalize words within just three seconds. It is accomplished via a 253-electrode array chip implant on the brain. Code and the dataset to replicate the main findings of this study are available in the Chang Lab's public [5]GitHub repository .



[1] https://slashdot.org/~backslashdot

[2] https://engineering.berkeley.edu/news/2025/03/brain-to-voice-neuroprosthesis-restores-naturalistic-speech/

[3] https://www.youtube.com/watch?v=MGSoKGGbbXk

[4] https://www.nature.com/articles/s41593-025-01905-6?error=cookies_not_supported&code=a7145d68-856c-4f7e-bc24-01c85674de85

[5] https://github.com/cheoljun95/streaming.braindecoder



Yay (Score:4, Insightful)

by vbdasc ( 146051 )

This heralds some advanced new methods of interrogation.

Re: (Score:2)

by znrt ( 2424692 )

methinks this needs the subject's collaboration as it doesn't scan thoughts but mental vocalization efforts. if the decoding isn't universal, you'd first need some primitive simple methods to enforce proper training, which would be a fairly verifiable process, but then you'd still need collaboration in the proper interrogation which could be easily defeated; back to uncertainty unless ... those simple primitive methods were in place.

ofc i wouldn't rule out that the same principle can be applied to actual th

Re: (Score:2)

by Guignol ( 159087 )

Indeed, the thing does not in fact understand thoughts at their innermost level where they would be very abstractly shaped and difficult if possible at all to decode, but instead it's 'wiretaping' the speech production of the end result.

This is why with this interface, if you are to drive this vehicle or country at ludicrous speed, you will need to think in Russian like Clint Eastwood or Donald Trump.

Oh great (Score:2)

by quonset ( 4839537 )

More voices in my head.

Re:Oh great (Score:4, Funny)

by Targon ( 17348 )

You lock the door, throw away the key, there's someone in my head, and it's not me....

Do not want (Score:3)

by Ol Olsoc ( 1175323 )

I triple parse every word I utter internally before speaking. I suspect that would confuse the shit out of this system. People definitely do not want to hear the first thing I think of before speaking. Hopefully there is a way for the people who need this to filter the speech.

Otherwise, there will be a lot of Sophia Petrillos running about. (obscure Golden Girls reference)

Re:Do not want (Score:5, Interesting)

by backslashdot ( 95548 )

That is not how this works, not unless you physically mutter your thoughts. It detects neurons activating in the sensorimotor cortex, which is the signal that normally would propagate to your vocal cords and muscles. Your brain would have already decided to sound out the words.

Re: Do not want (Score:2)

by Jeremi ( 14640 )

Seems like that's where things are headed though. It's going to be tough to keep anything private once the government can strap you to a machine and observe your thoughts directly. Don't think of an elephant!

Re: (Score:3)

by gtall ( 79522 )

As the GP said, that is not how that device works. If you think about it, you realize that they found a choke point for voice and then worked on synthesizing voice from that. That's a far cry from siphoning thoughts in your brain to what comes out of your mouth, unless you are one of those who utters whatever zephyr is wafting about up there. In that case, though, you could just strap a mic the person's mouth....just think of a politician in front of the Klieg lights.

Re: (Score:2)

by burtosis ( 1124179 )

> As the GP said, that is not how that device works. If you think about it, you realize that they found a choke point for voice and then worked on synthesizing voice from that. That's a far cry from siphoning thoughts in your brain to what comes out of your mouth, unless you are one of those who utters whatever zephyr is wafting about up there. In that case, though, you could just strap a mic the person's mouth....just think of a politician in front of the Klieg lights.

A model T isn’t a Bugatti Chiron, but I think they meant if things keep going this way for another hundred years or two we may have some black mirror like crap to deal with, and I also share that concern.

Re: (Score:2)

by omnichad ( 1198475 )

I'm just guessing here. They would have to train it on your own self-organized brain first. It's highly unlikely that your motor control system is set up the same as the next person. You would have to voluntarily read something out loud to train it. And then because you have working muscles, you would have to actually speak out loud for it to read your thought-words.

And either way, it doesn't prevent lying.

Re: Do not want (Score:2)

by devslash0 ( 4203435 )

If everyone's thoughts were public, the world would be a much more peaceful, direct and logical place. You wouldn't need to second-guess people. You'd know if you're dealing with a wanker or not, and what are their true intentions within seconds.

Re: (Score:2)

by larryjoe ( 135075 )

> That is not how this works, not unless you physically mutter your thoughts. It detects neurons activating in the sensorimotor cortex, which is the signal that normally would propagate to your vocal cords and muscles. Your brain would have already decided to sound out the words.

Like any new tools, the user would have to get used to a new way of doing things. Probably the thing that would take some training and practice is the delayed radio effect. We're used to hearing our speech instantaneously (or close enough). Hearing that delayed speech 3 seconds (or even a fraction of a second) delayed would be disturbing enough to make speech difficult. It would take some training and concentration to ignore our own delayed speech while we're in the middle of speaking.

Tourette Device (Score:1)

by Bart 5020663 ( 10135293 )

Give it a couple of decades and also this great technology will fit in your pocket.

Re: (Score:2)

by HiThere ( 15173 )

Not clear. This isn't just technology, it's an electrode array implanted in the brain. So unless you put your brain in your pocket, this won't fit in your pocket.

N.B.: We can fly to the moon, but it still takes just as long to walk across the room.

Re: (Score:2)

by GoTeam ( 5042081 )

Maybe in a few decades they'll get the contactless version going...

Re: (Score:3)

by SchroedingersCat ( 583063 )

Give it a couple of decades and everyone will have bunch of tiny holes in their skulls.

I will save you some time (Score:4, Funny)

by zawarski ( 1381571 )

boobiesâ¦boobiesâ¦boobiesâ¦

Make every politician wear one (Score:1)

by mwalker59 ( 8811201 )

... and pipe the output directly to Twitter.

Re: (Score:2)

by GoTeam ( 5042081 )

> ... and pipe the output directly to Twitter.

I think that would turn everyone into nihilists. Most politicians are vomit educing without knowing their thoughts.

Methinks... (Score:2)

by nospam007 ( 722110 ) *

..it will need a Tourette-blocker.

The street preacher looked so baffled
When I asked him why he dressed
With forty pounds of headlines
Stapled to his chest.
But he cursed me when I proved to him
I said, "Not even you can hide.
You see, you're just like me.
I hope you're satisfied."
-- Bob Dylan