Slashdot Mirror

← Back to Stories (view on slashdot.org)

Neuroscientists At MIT Developing DNI

Posted by Zonk on from the jack-in-and-tune-out dept.

coolphysco1010 wrote to discuss the possible development of a direct neural interface, ala 'The Matrix', that could eventually allow for instant object recognition. From the article: "Now, neuroscientists in the McGovern Institute at MIT have been able to decipher a part of the code involved in recognizing visual objects. Practically speaking, computer algorithms used in artificial vision systems might benefit from mimicking these newly uncovered codes ... In a fraction of a second, visual input about an object runs from the retina through increasingly higher levels of the visual stream, continuously reformatting the information until it reaches the highest purely visual level, the inferotemporal (IT) cortex. The IT cortex identifies and categorizes the object and sends that information to other brain regions."

36 of 126 comments (clear)

  1. Imagine the possibilities... by theufo · · Score: 3, Funny

    for adult entertainment.

    1. Re:Imagine the possibilities... by exaviger · · Score: 4, Funny

      wohoo, alter my IT neurons to think my girl friend looks like buffy :)

    2. Re:Imagine the possibilities... by Cruithne · · Score: 5, Funny

      Now were you R'ingTFA, or were you looking at the woman in the red dress?

  2. Sweet mother of brain implants. by Mecdemort · · Score: 3, Funny

    I'm going to be first in line for the new computer interface brain implants. Hopefully they don't run windows.

    1. Re:Sweet mother of brain implants. by annex1 · · Score: 2, Interesting

      Consider me second in line. I can't tell you how much of an improvement for the species this will be. Better yet, the blind may finally have a hope of actually receiving ACTUAL replacement vision, not a poor substitute. :D

    2. Re:Sweet mother of brain implants. by Mecdemort · · Score: 2, Funny

      Indeed, I was hinting at a new high of laziness that could be achieved. No longer would you have to move your fingers to use the computer. It would be possible to actually vegitate.

    3. Re:Sweet mother of brain implants. by Artifakt · · Score: 2, Interesting

      I'm thinking of Walter a John Williams novel, 'Hardwired'. The protagonist has cyber-eyes, and bought them just when the company had gotten most of the bugs out, but while they were still trying unusual features like sepia tone overlays, film nior settings, and the like to see where consumer interest lay. They then dropped all these features to make cheaper, more basic designs they could sell to the broadest possible market. I'm with Alexander Pope on this one:

      "Be not the first by whom the new are tried -
      Nor yet the last to lay the old aside."

      --
      Who is John Cabal?
    4. Re:Sweet mother of brain implants. by Anonymous Coward · · Score: 3, Informative

      They don't work well. I was involved in artificial vision implant work a few years ago, and the current spread from the electrodes is too large to stimulate the small numbers of neurons necessary to give anything other than a "light blob". The lab put a grid of 5x5 electrodes in a human eye and the human reported on what they could perceive from current applied to it: the surgery was only done on people about to have their eye removed anyway for medical reasons, to be at little risk of hurting a good eye.

      If the electrodes could be improved, with work like that of David Edell at http://www.ninds.nih.gov/funding/research/npp/sow/ N01-NS-2-2347SOW.pdf, then it might become possible to actually provide the shape of a letter to an artificial eye. But not even tapping the brain directly will get past the need to localize current for the nerves, and fascinating things happen around smaller and smaller electrodes that make it very tough. It's also delicate, expensive work, with lots of need for using lab animals and careful data and record keeping to be sure you're measuring what you think you're measuring. The concept of giving someone a neural jack that provides high-bandwidth computer data of any sort is, and remains, utter fiction.

      Surprisingly, the implants for artifical hearing work very well: having the auditory nerve laid out, low frequency to high frequency, along the bony tube of the cochlea helps localize the current to just the nerves you want to hit with each electrode.

  3. Poor Monkeys by el+americano · · Score: 3, Insightful

    I don't want to see the results when they start trying to recreate those nueral patterns in the monkeys brains. Honesty, to say that observing these kinds of patterns brings us any closer to injecting images directly into the brain, when we have so little technology to do that (knives and chemicals basically) is ludicrous. I suppose the writer, rather than the scientists, can probably take all the credit for that exaggeration.

    --
    Those are my principles. If you don't like them I have others. -Groucho Marx
  4. Chair by GloomE · · Score: 3, Funny

    I'm looking to purchase a dentist chair.
    Hole in the headrest preferable.

  5. It's about time! by Anonymous Coward · · Score: 3, Insightful

    The implications for using this technology to cure blindness (one day, obviously not immediately) are wonderful! This is the kind of thing science was really meant for - helping humans live better lives. Kudos to MIT!

    1. Re:It's about time! by Yvanhoe · · Score: 2, Informative

      Immediatly, a large range of blindness can be cured by implants, either by putting a CCD array inside the retina or, in case of damage in the optic nerve, a camera can be wired to the visual cortex. Right now, some blind people see ( with a low res, b&w image but see nonetheless) thanks to implants.

      link
      other link

      But yes, with the technology presented in the article, I suppose one could even cure blinds that have a damaged visual cortex.

      --
      The Wise adapts himself to the world. The Fool adapts the world to himself. Therefore, all progress depends on the Fool.
    2. Re:It's about time! by Anonymous Coward · · Score: 5, Informative

      Kudos to MIT!

      Every visual neuroscientist, ever, has been working on "deciphering part of the code involved in recognizing visual objects." Poggio and DiCarlo's contribution is mostly that they were able to record from a large number of neurons simultaneously in the inferotemporal cortex (IT). It's a logical (but interesting, to be sure) progression of work that has been done for decades in IT -- most of that work done elsewhere.

      Neural prosthetics and DNI are the bullshit that people trot out to make neuroscience interesting to the public. It's worth pointing out that neither of the actual named scientists in this work raise the possibility, and in fact, other than the abstract, there's nothing that even hints at the idea. These guys aren't working on a DNI. They're doing basic science. Years, decades down the road, some engineers might take the work that built on Poggio and DiCarlo's work and turn it into a DNI. Or at least, we can so hope.

      Name a university, and I can guarantee that the odds are that they'll have some basic science research underway with as much potential for the betterment of society as this stuff. So when you say "kudos to MIT" like this, remember that you're praising their PR department, not their scientists.

  6. Matrix? by Auckerman · · Score: 3, Interesting

    The article reads more like they are reverse engineering pattern recongition systems as the brain sees and interperates objects, which sounds closer to the movie Brainstorm.

    --

    Burn Hollywood Burn
    1. Re:Matrix? by Tune · · Score: 2, Informative

      Or Strange Days, which could be considered a (lose) remake of Brainstorm.

      Well, actually the article focusses on intercepting the sensoric data and making sense of it. I believe scientists have for some time been able to make sense of the basic sensoric data; stuff like using a cat's eye to produce webcam quality images. This research seems directed at interpreting the signals at a much deeper level.

      Though very interesting, it's still a one-way extraction process (ie. *not* synthesis) which is just completely unrelated to anything i saw in The Matrix, but I may have stumbled into an excuse to view that movie again ;-)

    2. Re:Matrix? by Haydn+Fenton · · Score: 2, Funny

      Sigh.
      Looser is what grammar and spelling nazi's should be. Loser is what grammar and spelling nazi's are.
      What's so damn bad about someone making a mistake so minor that anybody with an IQ higher than a banana can still understand?
      ;-)

  7. No 12 monkeys by noc_man · · Score: 4, Informative

    I read an article many years ago about them doing this to live human patients. Via a fiber cable brain wet-ware implant, a blind man was able to discern colors and rudimentary objects. He did have a short seizure during the interview; however, once the subject got passed that he immediately requested that the researchers continue.
    Unfortunately this was so long ago I cannot remember the magazine or relocate the article. But googling artificial vision shows a few parts of history and HOWSTUFFWORKS has a full set of details

    http://health.howstuffworks.com/artificial-vision. htm

    1. Re:No 12 monkeys by Timeburn · · Score: 3, Interesting

      IIRC, it was in Wired, circa about 1999 or 2000. The article covered research in South America (banned in the US), on a patient who had lost his vision, but whose optic nerve was intact. They interfaced directly into the nerve, stimulating it manually at first (This is when the seizure occurred).

      The project was apparently quite successful, as the patient was able to move about the facility, pick up a phone from a desk, and even drive a car around the parking lot. Fairly low-res input, but enough to see shapes and movement.

      Don't know what's happened with the project since, nor can I find the original article right at the moment, but it definately sounded promising.

      --
      "Not one shred of evidence points to the notion that life is serious" -- Samuel Clemens, aka Mark Twain
    2. Re:No 12 monkeys by groomed · · Score: 3, Informative

      The Vision Quest in Wired 10.09 of September 2002.

      The article, as well as the feasibility of Dr. Dobelle's (who has died in 2004) research, are sketchy at best. Apply truckload of salt.

  8. Fabulous by Centurix · · Score: 2, Funny

    Maybe they could simulate the feeling of taking a really great dump.

    --
    Task Mangler
  9. Just recordings by venicebeach · · Score: 2, Interesting

    Seems to me they are just recording from IT neurons. There's no input to the cortex. I haven't read the science paper (is it out yet?) but it really seems like they are just analyzing the firing patterns of IT neurons while the monkey looks at objects. Nothing new here technology-wise.

    1. Re:Just recordings by venicebeach · · Score: 2, Insightful

      OK I just read the Science article. What's interesting about it is that they got recordings from a large population of neurons in IT during object recognition and have some cool analyses of the kinds of information that can be extracted from the capture, e.g. how large a population of neurons you need to accurately identify the object, how well the neurons discriminated among the categories and generalized across the same image at different sizes and positions, etc.

      Important to remember that these monkeys were trained on a limited stimulus set, so its not that you can tell what the monkey is looking at by loooking at the recordings without knowing it is one of these pre-trained items.

    2. Re:Just recordings by FleaPlus · · Score: 4, Informative

      This reminded me of the research by Quian Quiroga et al in which they performed single-neuron recordings from MTL (upstream of IT, if I recall correctly) in humans. In that study they found neurons which would respond selectively to particular objects, such as Jennifer Aniston, Halle Berry, and the Sydney Opera House. Here's the abstract:

      R. Quian Quiroga, L. Reddy, G. Kreiman, C. Koch & I. Fried Invariant visual representation by single neurons in the humanbrain. Nature (2005) 435, 1102-1107

      It takes a fraction of a second to recognize a person or an object even when seen under strikingly different conditions. How such a robust, high-level representation is achieved by neurons in the human brain is still unclear. In monkeys, neurons in the upper stages of the ventral visual pathway respond to complex images such as faces and objects and show some degree of invariance to metric properties such as the stimulus size, position and viewing angle. We have previously shown that neurons in the human medial temporal lobe (MTL) fire selectively to images of faces, animals, objects or scenes. Here we report on a remarkable subset of MTL neurons that are selectively activated by strikingly different pictures of given individuals, landmarks or objects and in some cases even by letter strings with their names. These results suggest an invariant, sparse and explicit code, which might be important in the transformation of complex visual percepts into long-term and more abstract memories.

  10. sorry to dash your hopes, but... by Xochi77 · · Score: 5, Informative

    i am not at MIT, but I can tell you this aint about to happen any time soon.

    i am working on optical neuron-computer interfaces, and this is probably the most efficient and direct route for reading neurons. I know of researchers who can also stimulate neurons to fires via light, so in principle, we could build a complete neuroptical computers tomorrow... if neurons were not complete bastards to work with.

    you see, they just dont like to stay place. where i research, they often build tiny fences to keep them in place, but even then, they go shooting theyre axons anywhere they feel, with no concern for the feelings of the researcher.

    we also grow neurons on microchip surfaces, which allows for high speed and high resolution stimulation and reading of single neuron activity, but in two dimensions, which is excellent for retina etc.

    but the neuron-chip or old fashioned neuron-electrode are hard to place, and optical reading of neurons still has bugs to sort out (id guess from 4-10 years more basic research). whenever you see these cool brainscan pics with MRI etc, remember theyre resolution is on the order of millimeters, and thats a lot of complexity lost.

    http://www.biochem.mpg.de/mnphys/ has a nice review of the problems involved, if you like hardcore solidstate chemistry, silicon physics, and neurobiology

    1. Re:sorry to dash your hopes, but... by FleaPlus · · Score: 3, Informative

      What are your thoughts on using autonomously adjusting electrodes to deal with the problem of neurons shifting about? Granted, the current systems are rather bulky, but much more compact ones are under development.

    2. Re:sorry to dash your hopes, but... by Xochi77 · · Score: 5, Informative

      cute, but check out- http://scholar.google.com/url?sa=U&q=http://www.bi oon.com/biology/UploadFiles/200502/200502160347225 62.pdf "Functional imaging with cellular resolution reveals precise micro-architecture in visual cortex" also, i forsee the development of light-gated ionchannels, such as the one i mentioned before, that can be opens at various wavelengths, much like gfp has been mutated from green to the whole blue-red spectrum. geneticaly specified reading of neurons has furthers to go, but it will happen, and soon i think. in the end, why go with hacking into the brain to insert electrodes and chips etc, when two-photon microscopy can see though tissue?

  11. OMFG by russint · · Score: 3, Funny

    this could make LSD obsolete!

    --
    ^^
  12. What if... by nsasch · · Score: 2, Interesting

    What if the IT cortex was bypassed: The computers would get or simulate the input, and recognize and categorize the object and the computer would send that data directly to the other parts of the brain. Now the human doesn't see the ball, but knows there's a ball in front of them, and it's red, and about the size of their head, etc (all the details), but doesn't see it, just has a "feeling" that a ball is there.

    --
    Make your computer faster: rm -rf /mnt/windows/
    1. Re:What if... by kko · · Score: 2, Insightful

      Yeah, cool, well, actually getting any sort of input to the brain seems to be a big part of, if not THE actual problem.

      Jeez.

      --
      No, seriously, I just come here for the articles.
  13. Re:Cyborg possibilities by Kjella · · Score: 2, Interesting

    The most interesting one to me is the ability to supplement your own faulty memory with a hard drive and your own thinking power with a processor. You'd take a little snapshot of every person you met and file it away with their name, never to be forgotten. Think about what school would do for you! If I could remember all the science, history and literature I've been taught and forgotten, I'd be a much more educated guy than I am now.

    Not to go all Trinity on you, but why limit it to your own experiences? Basicly, say you wanted to recall the text of something you've never read, the HDD could supply it. That is simply on a request-response variety. You could do searches in information bases you've never read. You could do a two-way communication to make drill-downs. Let's say you were looking at a bird, and you could supply information to the base, the base might ask "questions" like color, size, beak, feathers, legs, sound to your brain to pull the information you want. The whole of wikipedia could easily fit on such a HDD. Sure there'd be a lot of trial and error here but this data could be gathered from everyone carrying it to improve the interface. It's more a matter if the human mind could keep up or if it'd go wacko from all this information at its fingertips. Then you could really talk about information overload.

    --
    Live today, because you never know what tomorrow brings
  14. Boon for Camouflage by schwit1 · · Score: 2, Interesting

    Could you test potential camouflage patterns with this and find which cause the most difficulty in visually deciphering? Or one day have computers generate camouflage on the fly based upon the surroundings.

  15. Code Talkers by Doc+Ruby · · Score: 2, Interesting

    It's not a "code". There's no objective reality that the brain is decoding for mere "referential integrity". The brain is organizing its responses to incoming sensory info, in a feedback loop with itself, including resonating "memory" response signals. Sure, object representations are recognized as repeats of previous object representations, and dispatched to brain areas sensitized to those representations. But it's not like objects outside the body have standard codes, the same from person to person, like say insulin has in our DNA. That would be way to static for us to survive in this changeable world. We're making it up as we go along, and living in the reality we generate. The closer our mind's model matches the world we encounter, the smarter we are.

    --

    --
    make install -not war

  16. it's not DNI at all by darkeye · · Score: 2, Informative

    as they don't actually connect to the neurons, but read the neuron acticity patterns, probably through fast MRI scanners. and there's no feedback either - they don't send any data to the neurons (other then through the natural eye of the monkey in the tests)

  17. Re:feedback loop? by CptNerd · · Score: 2, Funny

    But what if we're already in the matrix.. that would be a direct neural interface inside a direct neural interface... Talk about a mind bender.. or should I say 'spoon bender'?

    Or more likely just Bender: "We're boned!"

    --
    By the taping of my glasses, something geeky this way passes
  18. We have non-invasive signal injection technology by lightyear4 · · Score: 2, Informative


    We already have something called transcranial magnetic stimulation. See:

    http://ieeexplore.ieee.org/xpls/abs_all.jsp?arnumb er=1300793
    http://groups.csail.mit.edu/vision/medical-vision/ surgery/tms.html -- most relevant to discussion, has section on visual signal injection
    http://www.biomag.hus.fi/tms/
    http://www.mp.uni-tuebingen.de/mp/index.php?id=94
    http://en.wikipedia.org/wiki/Transcranial_magnetic _stimulation
    http://pni.unibe.ch/TMS.htm

  19. I'm currently doing similar work... by mr.+squishie · · Score: 2, Interesting
    ...but on human subjects using fMRI. This research really isn't related to the matrix or DNI's directly, it's about seeing whether or not electrical signals from the brain contain enough information for a classifier (ironically, in our case, artificial neural networks) to distinguish between some subjective cognitive state.

    Considering the progress we've made in distinguishing cognitive states (is this person looking at a face, a house, a squirrel, etc?) in human subjects using fMRI (an extremely noisy dataset), I'm not surprised that they found that there's enough information in a few neurons to perform classification.

    Really, the best pop-sci term to describe this would be "mind reading" -- the high level goal is to have a function that transforms physical space to some sort of cognitive space. I guess you could say it's the "I" of the I/O DNI in the matrix.