Major Breakthrough in Direct Neural Interface

jd writes "In a major breakthrough, neurologists are reporting that they can decypher neurological impulses into speech with an 80% accuracy. A paralyzed man who is incapable of speech has electrodes implanted in his brain which detect the electrical pulses in the brain relating to speech. These signals are then fed into computers which covert these pulses into signals suitable for speech synthesis. As a biotech marvel, this is astonishing. Depending on the rate of development it is possible to imagine Professor Hawking migrating to this, as it would be immune to any further loss of body movement and would vastly accelerate his ability to talk. On the flip-side, direct brain I/O is also a major step towards William Gibson's Neuromancer and other cyberpunk dark futures."

what if

[rucs_hack (784150)]

The subject turns out to have Tourettes syndrome?

OI! [redacted] will you [redacted] [redacted] [redacted] make me a [redacted][redacted][redacted] cup of [redacted] coffee?

Brain obscenity filters for teh wins....

Sadly more likely...

[nick_davison (217681)]

My wife was in a massive car accident, a decade ago. She was in a coma for a month, suffered brain injuries, a collapsed lung, shattered arm, cracked eye socket, multiply broken jaw, etc. A national merit scholarship winner before the accident, her parents were told that, if she survived, she'd likely never walk much or be able to look after herself again.

As it happened, she was sufficiently beaten up at the time that she had no concept of how bad her injuries were. She got out of the wheelchair simply because it frustrated her. She went back to working part time simply because she didn't realize she wasn't supposed to be able to. By the time she comprehended what had happened, she'd improved enough that setting impossible goals like "become a personal trainer" weren't quite so impossible. We taught her to read again (yes, even that got messed up) and even managed to get her back in to school - initially only able to pull a 2.0 average but improved each semester.

In her case, she had an amazing recovery. Yet she, herself, says, "If I'm ever like that again, turn me off." She didn't realize how hurt she was and got lucky with recovering before she did. Understanding now, she has absolutely no desire to try that fight again. She'd rather just call it a day.

So, sadly, there's a real likelihood that his first words, upon realizing he can finally communicate, after years of being unable to and stuck in a totally paralyzed body, will be, "Kill me." Probably not ideal to have the family in the room for.

And yes, that entire story was just so I could "drop" that I have a wife in a slashdot post. Cunning, huh?

Re:Sadly more likely...

[tjstork (137384)]

And yes, that entire story was just so I could "drop" that I have a wife in a slashdot post. Cunning, huh?

Your wife's recovery and you staying with her, through all of that, is the most poignant thing I have read on Slashdot, ever.

A story like yours deserves to be told, and demands that we listen.

May the winds always be at your back.

Re:

[papvf (725513)]

As a slashdot.dot reader it goes without saying that I love to revel in the latest tech but, stories like this one prove that it is people like you and your wife that are the true inspirations in the world. All the tech and science is wasted if it can't benefit people with "real lives" like yours. Like tjstork said: "A story like yours deserves to be told, and demands that we listen." Any that don't listen, cut them selves off to reality and lose out on more than they can dream of. -papvf

More info

[niceone (992278)]

The BBC article is pretty light on detail, and the New Scientist one is subscribers only, but there is more stuff here [eurekalert.org].

They have hooked up to 41 neurons and:

For now, the team is focusing on the building blocks of words. In a series of experiments over the last few years, Ramsey has imagined saying three vowel sounds: "oh", "ee" and "oo". By watching his brain activity, the researchers have been able to identify distinct patterns associated with the different sounds. Although the data is still being analysed, they believe that they can correctly identify the sound Ramsey is imagining around 80 per cent of the time

Re:

[WombatDeath (681651)]

Cool! With a bit more work he'll be able to join in the chorus of Old MacDonald.

Re:

[sseaman (931799)]

Thanks, that's quite helpful. I could find no details about this on my own, lacking a New Scientist subscription. He isn't "imagining" these sounds - he's trying to produce them. I suspect they've tapped into the motor cortex, where one of the last stages of motor processing. They're not tapping into "speech" centers - it's simply a motor area associated with articulatory muscles. Not that it isn't impressive, but it's not a step towards mind-reading or better computer-human interfaces unless you suffe

Re:

[fair_n_hite_451 (712393)]

"oh-ee-oh"

Well, he can already do a voice over part of the Flying Monkey Chorus if they ever remake the Wizard of Oz.

This tech is so cool it's not funny.

As Fleet Captain Pike said..

[Rob T Firefly (844560)]

"Beeeeeep." [memory-alpha.org]

What drives modern science?

[InvisblePinkUnicorn (1126837)]

What drives the advances of the last couple decades?

Two desires:

1. To restore Stephen Hawking's physical body to its former fully-functional form.

2. To turn Stephen Hawking into a mobile, indestructible cyborg of incomprehensible power.

Been done!

[Sqweegee (968985)]

http://www.theonion.com/content/node/39133 [theonion.com]

"With the new exoskeleton, Stephen will be able to safely handle radioactive isotopes in the high-radiation area of the new supercollider particle accelerator. And his new robo-arms are capable of ripping open enemy tanks like they were nutshells,"

Wait-- they haven't actually done this yet

[raddan (519638)]

Read carefully

Although the data is still being analysed, researchers at Boston University believe they can correctly identify the sound Mr Ramsay's brain is imagining some 80% of the time.

In the next few weeks, a computer will start the task of translating his thoughts into sounds.

"We hope it will be a breakthrough," says Joe Wright of Neural Signals, which has helped develop the technology.

While this is indeed promising, and I hope that this 'unlocks' this poor fellow, this 'unlocking' has not happened yet. Hopefully, when they are able to decipher these signals, he's not saying, "Kill me" over and over again.

Re:

[noidentity (188756)]

It's OK, the editors are only able to decipher what TFA says some 80% of the time.

What?

[Wellington Grey (942717)]

Electrodes have been implanted in the brain of Eric Ramsay, who has been "locked in" - conscious but paralysed - since a car crash eight years ago.

What do you do for eight years as a locked in? Wouldn't that drive a normal person insane or dull the mind beyond all recognition? Does anyone know about the mental state of these people?

-Grey [silverclipboard.com]

can still communicate

[circletimessquare (444983)]

http://en.wikipedia.org/wiki/The_Diving_Bell_and_the_Butterfly [wikipedia.org]

he could blink. that's it. yes or no. and with that ability, letter by letter, he wrote a book (with the help of some very patient nurses/ assistants)

it's coming out as a movie soon too i think

Re:What?

[klenwell (960296)]

I believe Antonio Damasio addresses this question in one of his books. Apparently, a fortunate side-effect of this condition is it impairs the part of your brain that would normally find this horrific and intolerable and leaves you with a weird sense of acceptance and well-being (IIRC). Otherwise, I guess you just blink a lot and hope they keep the feeding tube hooked up.

Re:

[Wellington Grey (942717)]

Apparently, a fortunate side-effect of this condition is it impairs the part of your brain that would normally find this horrific and intolerable and leaves you with a weird sense of acceptance and well-being

Really? I hope so, but that just seems like too much of a coincidence -- like something the caregivers tell themselves so they don't have to deal with the horror of the situation.

-Grey [luminiferous-aether.net]

Mr. Gibson's dark future is a human failure ...

[2TecTom (311314)]

and not a 'techno-biological' failure. The future's darkness comes from a tyrannical plutocracy which misuses the technology, which could have just as easily been used to save mankind. It is in fact an outgrowth of current economics and politics, not technology. Please, get your stories straight.

Mmmmyeeaaah, but ...

[anticlimate (1093749)]

But your kind of reasoning could also be used inside out, eg: "Mr. Gibson's dark future is a technological failure and not an economical/political one. That nasty future comes from a tyrannical group of technologists who misuse the social system."

What I want to say is technology and politics/economics are all a creature of humans. It's just as misleading blaming "economics" and "politics" instead of the people misusing the system (who are basically all of us), as it is to blame a particular technology for

I'll raise the BS flag on that

[conspirator57 (1123519)]

Tyranny has been around since before the stone age. What has technology got to do with it other than increasing the tyrant to subject ratio? The desire to oppress is inherently a human social one. Some will claim (neocons for instance) that we can use tyranny to make things better, but it doesn't work that way. Technology, on the other hand is much more legitimately separable from human motivation (there are a variety of motivations that can lead to most technologies.) Moreover, unlike tyranny, we have a chance of using a given technology only(or at least predominately) for good. Technology is a double edged sword, in part because it and its fruits are actually tools, not motivations unto themselves.

Vinge, not Gibson

[wurp (51446)]

Gibson didn't invent cyberspace. Vernor Vinge invented cyberspace (although I don't think he coined the term) in True Names.

If you haven't read it, I highly recommend it. Read True Names to get a notion of the profound visionary Vernor Vinge is. (Remember it was published in 1981).

Then read Rainbows End with your newfound respect for Vinge's powers of prognostication, and recognize that you're seeing into the near future.

This could be really embarrasing for users

[JustNiz (692889)]

how would it bea ble to differentiate between "out loud" voice and private thoughts? This could be really embarrasing for users. Imagine if a secretary (or nurse) walks by when you're in the middle of speaking or dictating a letter:

Dear sir,
I am writing wow nice tits and she has a great ass too uh oh wedding ring in order to ask if you would be interested in our new product line of neural-input word processors.

80% accuracy...

[uwbbjai (661340)]

It reads: "Dear Aunt, let's set so double the killer delete select all"

What do you want to decipher today?

That extra 20% probably wasn't important anyway...

[uhlume (597871)]

Subject's first words? "Dear Aunt, let's set so double the killer delete select all."

Just make sure the interface....

[StressGuy (472374)]

...is in the right part of the circut

I don't always want my "first throught" to be the one that gets verbalized, know what I mean?

Hi Mrs. Johnson, nice tits!....buts a little big though

Oh shit....did I say that out loud?

Research posters

[FleaPlus (6935)]

For those curious, this speech prosthesis research was presented in a number of posters at the Society for Neuroscience (SfN) conference a couple weeks ago. Their six SfN posters can be found on their website here, covering topics like the circuitry they developed, Bayesian signal analysis, and so forth:

http://migrate.speechprosthesis.org/DNN2/SpeechProsthesisHome/tabid/52/Default.aspx [speechprosthesis.org]

There's also a nice blog entry on this over at Neurophilosophy:

http://scienceblogs.com/neurophilosophy/2007/11/speech_prosthesis.php [scienceblogs.com]

Obligatory

[Guppy06 (410832)]

Dear aunt, let's so double the killer delete select all.

Who's doing the work

[mesterha (110796)]

One has to wonder who is doing the work. Is the paralyzed man adapting to the computer or is the computer learning the brain signals. Either way, it's good work, but I would bet that the way to perfect this type of technology is to "teach" the human to control his neurological impulses. I doubt the technology is directly eavesdropping on his speech.

Re:Really accurate?

[DragonWriter (970822)]

How do they know they're accurately converting the signals to sound, if they're basing this off a man who has no ability to speak?

Many people who are unable to speak are able to communicate in some other way (usually, some form of gesture, whether sign language, nodding, blinking, whatever.) It doesn't take a much to be able to indicate "right" or "wrong".

Re:Really accurate?

[sorak (246725)]

Many people who are unable to speak are able to communicate in some other way (usually, some form of gesture, whether sign language, nodding, blinking, whatever.) It doesn't take a much to be able to indicate "right" or "wrong".

Remember, it's only 80% accurate. It may be more like "rigm!" or "prong!"

Re:Really accurate?

[skoaldipper (752281)]

I have time to kill on lunch. Let's see...

Right Wrong = 10 letters.

P in Prong = 1
M in rigm = 1 (+1 letter missing)
1 missing = 1/2

So, (10 - 2.5)/10 = 0.75 ~ 80%

Your post above not only meets funny standards, but accolades for careful thought in using relevant and accurate choice of words. Well done, sir, well done!

Re:Really accurate?

[fbjon (692006)]

Reportedly, the first words spoken through this interface were:

"Frist wrods!!"

Re:Really accurate?

[rucs_hack (784150)]

I'm guessing the 80% comes from the fact that this is an issue of the linear separability of signals. Its generally hard to get reliable sensitivity/specificity measures over this that anyone is going to take seriously.

Sensitivity = percentage number of correct identifications
Specificity = corresponding percentage of incorrect identifications at each measured sensitivity.

Probably they can get up to 90%, but from experience I would say the rate of false positives at this sensitivity likely is moving towards exponential increase. It's better to stop at 80%, at least when something is in the early stages.

This is just guessing of course, I have no understanding of their research, but going from my own work on non linearly separable sets, I'd say this is what's happening.

Re:Really accurate?

[dintech (998802)]

Now you are just putting words in his mouth. :)

Re:Really accurate?

[Lord Apathy (584315)]

This is great. Now all we have to do is reverse the fucker so it figures out 80% garbage and 20% signal. Then we attach it to congress critters, lawyers, and RIAA stoges. Now we don't have to listen to their shit at all anymore.

Re:Really accurate?

[QRDeNameland (873957)]

How do they know they're accurately converting the signals to sound, if they're basing this off a man who has no ability to speak?

I can see it going something like this...

Researcher: "The machine translates his electrical pulses as 'I'd really enjoy a blowjob from your assistant, Ms. Jenkins.' Ms. Jenkins, do you mind?"

Ms. Jenkins: "Anything in the name of science!!"

Researcher: "Well, that ear-to-ear smile is conclusive proof that he is in fact enjoying it. Eureka, it works!!!"

Re:Really accurate?

[Harmonious Botch (921977)]

No need; Ms Jenkins is reputed to be quite competent. ;-)

Re:

[Thought1 (1132989)]

They know it's accurate because the voice translation told them it was! It then said something about "robotic voice translator overlords..." We're not sure about that bit. (:

Re:

[vux984 (928602)]

They know it's accurate because the voice translation told them it was! It then said something about "robotic voice translator overlords..." We're not sure about that bit. (:

Hence the 80%.

Re:

[deathy_epl+ccs (896747)]

How do they know they're accurately converting the signals to sound, if they're basing this off a man who has no ability to speak?

That was the easy part... they were able to start with the assumption that he just kept repeating "kill me" over and over again.

Re:Really accurate?

[by Anonymous Coward]

The article says the man is 'locked in', which means that he not only cannot speak, but he has no voluntary movement whatsoever, even blinking eyelids.

There was an article recently in New Scientist about this. One problem doctors studying this field have is that since it is an experimental treatment, they need consent of the patient, and how can they get consent if the patient can't communicate?

With some locked-in patients, they are able to respond based on the acidity of their saliva. They are told to either imagine eating lemons (for yes) or eating milk (for no), and their saliva sympathetically adjusts to their thoughts. Then their saliva is measured. See more here: http://www.mindhacks.com/blog/2007/08/locked_in_with_the_b.html [mindhacks.com]

Sad to say it, but I suspect the first thing the patient will say is "kill me".

Re:

[Lord Ender (156273)]

Clearly, you have never watched Star Trek. They put him in a little electric wheel chair with a big red light on it. He can make it beep once for "yes," twice for "no."

And, amazingly enough, he can somehow still get his mojo on if you beam him down to the right planet.

Re:

[Beardo the Bearded (321478)]

"Wow, it's actually picking up what I'm salmon."

"Wait, what?"

Well, there's my extra marks for funny. Now for the serious part:

I've taught a few people who couldn't speak how to work their voice. In one case, she would talk a little like Boomhauer from King of the Hiil. "Daddy, mumble mumble me mumble mumble juice mumble mumble counter?" Once she got used to the feedback and the system, she would fill in the mumbled parts with the correct conjunctions. Perhaps that's how the 80% is getting in there. The gene

Re:80% accurracy?

[$RANDOMLUSER (804576)]

80% accuracy is NOWHERE near good enough.

It's good enough to get you elected president - twice.

Re:What about the babies??

[Not_Wiggins (686627)]

Would a device like this work on someone who doesn't know how to speak english or better yet a baby that speaks no language at all

The answer is "Yes" (but not the way you intended) and "No."

It would work for a non-English speaker IFF that speaker was trying to speak his native language; what they've detected is the brain's intention to produce a SOUND; so, by extension, the interpretation is producing a phonetic representation of the sounds in the person's head.

It isn't interpreting the concept of the sound (someone isn't thinking of a cat and the word "cat" is produced). It should be possible for someone speaking any language (including a made-up one) to use this system.

For a baby (who has no word associated with the object), it wouldn't provide any use... unless your conjecture is that a baby doesn't speak because the muscles in her throat aren't strong enough to form words, but her brain knows what sounds would be made. Then... sure, it would work. 8)

Re:

[hoggoth (414195)]

> Would a device like this work on someone who doesn't know how to speak english or better yet a baby that speaks no language at all, if so then we just invented the universal translator, live long and prosper trekkies

Yes it certainly would. The device works by directly picking up the intent of the subject in a global individual-neutral format. That intent is then translated into English by dictionary lookup and standard text-to-speech software. It would be a trivial matter to subsitiute any other langua

Re:I'm skeptical at best.

[hoggoth (414195)]

> you would need to do some sort of heroic measure of training for each individual

Not be be callous, but I'm pretty sure they can find time in their busy eating, sleeping, and bedpan changing schedules in order to regain the ability to communicate with the world.

Re:Slashdot.

[Lord Ender (156273)]

Judeo-Christian values were at the core

If, by "Judeo-Christian," you mean "Western," then you are right.

The majority of Western values do not trace their roots to any of the Middle Eastern religions. They come from other places, such as Greek philosophers.

In fact, the philosophical foundations of the US are in many ways opposite to the so-called Christian values. Cruel and unusual punishment, for example, is condoned--actually commanded--by the Christian god. Slavery, and the belief that all men are NOT created equal, is a common theme in the Bible.

The statesmen/philosophers who founded this country may have been Christian, but the documents they wrote to found this country were quite the opposite.

Re:Slashdot.

[mrchaotica (681592)]

Actually, most of the founding fathers were Deist [wikipedia.org], not Christian.

You need no Creator to believe in rights

[spun (1352)]

It's not duplicity of thought. You just lack understanding. One does not need a creator to imagine a human spirit. In fact, the idea of a creator adds nothing to the idea of the spirit. It just marks an artificial stopping point in the quest for answers: What did it? Creator did it! What made Creator? Don't go there! Dumb.

Eastern religions have a better word for it: suchness. That is just so, as it is. The idea of spirit relates more to the idea that things are more than the sum of their parts (due to the i