Slashdot Mirror


Robot Controlled By Human Brain Cells

destinyland writes "There's a new experiment from the British researchers who created a robot controlled by cultured rat neurons. They're now using a line of human brain neurons to control robots. The neurons are placed onto a multi-electrode dish that registers the neurons' electric signals. 'Every time the robot nears an object, the electrodes generate signals to stimulate the brain. In response, the brain's output is used to drive the wheels of the robot left and right so that it avoids hitting objects. The robot has no additional control from a human or a computer — its sole means of control is from its own brain.'"

20 of 86 comments (clear)

  1. I wasn't worried...at first. by Adustust · · Score: 2, Funny

    I always figured we'd be OK. Rats only attack if they're scared right? But people...oh God! Just what we need, a robot that thinks it'd be funny to be a zombie.

  2. bad summary? by Thornburg · · Score: 4, Insightful

    Is it just me, or do the video and article both CLEARLY state that it's rat brain cells, not human brain cells?

    1. Re:bad summary? by Hitman_Frost · · Score: 4, Insightful

      Blame Kevin Warwick.

      He's always exaggerating his claims, including his "I'm a cyborg" nonsense.

    2. Re:bad summary? by Rogerborg · · Score: 5, Insightful
      Aaaah... "Captain Cyborg"! I can't believe this chap (or robochap as he'd describe himself) is still getting funding for his joke research. I guess Reading University don't mind being laughed at, as long as they're being talked about.

      "It's difficult to describe how frustrating it is in the field seeing this man being our spokesman," says Richard Reeve, of the AI department at Edinburgh.

      Well, quite.

      --
      If you were blocking sigs, you wouldn't have to read this.
    3. Re:bad summary? by julesh · · Score: 2, Informative

      They did the experiment with the rat cells some time ago, now they are starting to work with human cells - the article states this clearly

      Yes, but the summary doesn't. The summary says they have _done_ it. "They're now using a line of human brain neurons to control robots." No, they're working on plans and beginning experiments by which they hope, at some point in the future, to use human neurons to control robots.

      (Q: Is it a "brain neuron" if it's cultured in vitro?)

    4. Re:bad summary? by mcmire · · Score: 4, Informative

      There are two experiments involved here, one using rat neurons and one using human neurons. The article is badly written -- it first introduces the two experiments, talks about the rat-neuron experiment for a bit (that's what the video refers to) and then abruptly segues to the human-neuron experiment (which there is no video for). Only the last three paragraphs are really about the human one. Looks like it's the same setup as the rat one though.

  3. Alas its' first words were... by gijoel · · Score: 4, Funny

    Ex-TeR-MiN-AtE! Ex-TeR-MiN-AtE!

  4. Re:AI? by timeOday · · Score: 3, Informative

    we have real intelligence controlling robots. The though of someone hooking this ip to a Predator drone are scary.

    The use of human brain cells doesn't imply "real" intelligence. Computationally, this thing is vastly weaker than "traditional" (silicon) computers from 30 years ago, much less whatever is on Predators (or your iPhone) today. There isn't any pixie dust in neurons.

  5. Not too smart by gr8_phk · · Score: 3, Interesting

    Since the cells are not connected to the motors directly, there must be some other electronics involved. Since there is no mention of learning, and the behavior seems consistent, we should raise the suspicion that the neurons are acting like nothing more than wires. Or is this a case of interesting work being dumbed down for a YouTube clip?

    1. Re:Not too smart by sonnejw0 · · Score: 2, Informative

      It's acting more like a diode, really, than just a wire, but yeah, you're basically on target.

      It is also plastic: i.e. the neurons are free to associate with whatever electrodes they are most attracted to, just like in a biological brain. That means that the electrodes that are most active will receive the most connections, and thus the "wires" are self constructing according to "need" (or frequency of stimulation in a molecular biology sense).

    2. Re:Not too smart by Anonymous Coward · · Score: 2, Informative

      The fact is there is no signal in this system that could be used in something such as reinforcement learning. The robot has absolutely no way of knowing if hitting a wall is good or bad. So what this system actually does is show that neurons can auto-organize and produce a coherent output based on inputs such as sensors values.

      What Warwick, forgot to say is that for every robot that converges toward an object avoidance behavior there are probably 30 other robots having completely difference behaviors. Since any of these behaviors is as good as the others for the robot (remember: no reward or pain of any sort), we are just shown the behaviors that make sens to us...

  6. Re:Does it count? by sonnejw0 · · Score: 3, Insightful

    Ahh, I see you're a materialist at heart. This is true of a fully human brain as well: an action potential is just a response to a molecular-mechanical stimulus that opens ion channels to change the polarity of the neuron. What makes a human "brain" controlled?

    Truth is, there's nothing special about this robot. It basically uses rat neuron cells to propogate an electrical signal instead of full-length wires. But if you believe that, then you also believe there's nothing special about the human brain. It just responds to environmental stimuli in a predictable, yet seemingly complex way. Big deal.

  7. 790 by morgauxo · · Score: 2, Funny

    790 is that you?

  8. Warwick has his own reality TV show by RandCraw · · Score: 5, Interesting

    In 1998, Kevin Warwick implants a trivial RF resonator in his arm (the sticky plastic strip that warns Walmart that someone is stealing a pair of socks). He contacts the press, calls himself a cyborg, and gets tenure at Reading U.

    In 2001, he replaces his implant with an RFID tag, and calls the press again, and says, "Look at me, look at me! Now I'm an ACTIVE cyborg!" And he becomes a full professor at Reading U.

    And now Warwick gets 300,000 neurons to produce a simple binary response (go straight vs turn). He calls the press again, and says, "Look at what I can do just by waving my arm".

    Jackass. Worse, the media can't tell the difference between poseurs like him and real scientists like Sebastian Thrun or Rod Brooks.

  9. Just A Few Thoughts by DynaSoar · · Score: 3, Interesting

    (pun unintended, and frankly not very funny, so forget it; same for that pun)

    1. The magazine h+ senior editor is RU Sirius. It is MONDO 2000, +20. Sirius is still Sirius, and seems to have foregone the +20 himself. He's actually no slouch, so when he senior-editorizes a magazine full of pseudoscience crap when he could have done better or at least different, I feel he's earned the right to the criticism rather than the fiction writers working as science article journalists. They start stupid, work stupid and produce stupid. He approves it for publication.

    2. Other articles in the magazine are equally absurd. Some make claims about specific phenomena or theories which are anywhere from fraudulent to simply goofy. I took one such article, claiming that depression is lack of "fun" to task over at The Daily Grail. The article is just bullshit in its best parts. There's worse.

    3. Using neurons in this design is enormously overly complex. There was an article in SciAm in which little battery powered cars were given photo-cells or photo-resistive cells as "eyes", those driving the back wheels on the same side, or on the other side, making 4 different designs that react to light. They approach fast and slow down, or slowly at first then rush in, or they zoom away and orbit the edges slowly or else creep away and at the outer area zoom around. The anthromorphization of their actions is multiplied when many of each 'species' are placed on th same floor with protruding light bulbs at several locations. It's not that the neurons don;t do the job, it's just that they're a computational device based far beyond the need, and that non-computational, analog increasing/decreasing voltage circuits do much more for much less.

    4. Given the falsification of a different h+ article of equally strong claims, I serious(!)ly doubt the existence of the items in TFA. In fact I wouldn't believe anything I read in this magazine, if for no other reason than that the New Agey products in the ads are, while absurd themselves, more believable than the articles.

    --
    "I may be synthetic, but I'm not stupid." -- Bishop 341-B
  10. Not a robot by AP31R0N · · Score: 2, Interesting

    If it's being controlled, it's not a robot. It's a car for rat brain cells. It's not following a program, it's being steered. While we're at it, the battle bots/robot wars are not robots either. They are remote controlled cars with weapons and armor. My car is not a robot just because it is a machine. If i attach some kind of Myth Busters control system... it's still not a robot. Until it's driving itself, it ain't a robot.

    Also: a robot is an android IFF is is human shaped. T1000, C3P0 - Yes. R2D2 - No.

    i think a show about ACTUAL ROBOTS hunting each other down would be way cooler. i'd have a category for 'Onboard Brain' and another for 'Remote Brain'. Maybe categories for environment: Water, Air, Land. Then maybe size class. It would be nerdgasmic.

    --
    Utilizing the synergization of benchmark e-solutions to pre-workaround action items!
  11. Finally by Locke2005 · · Score: 3, Funny

    Couple this with the "Robotic hand that can feel", and you got all the ingredients necessary for the first robotic sexual harassment lawsuit!

    --
    I've abandoned my search for truth; now I'm just looking for some useful delusions.
  12. ...unless it's for a sexbot. by Impy+the+Impiuos+Imp · · Score: 2, Funny

    This can't turn out well.

    Unless it's for a sexbot, of course.

    Oh, wait. Even just a few human brain cells are enough to know to steer clear of the oddness of nerds.

    --
    (-1: Post disagrees with my already-settled worldview) is not a valid mod option.
  13. What does the brain do? by im_thatoneguy · · Score: 4, Insightful

    What does the brain do in this robot? It sounds like all data processing and decision making is done on silicon with the brain along for the ride.

    Headline should be "Rat nerve cells get ride around lab in little cart."

  14. Re:Exterminate!, exterminate! by Cassini2 · · Score: 2, Informative

    Doctor Who references to Robots controlled by organic brain's include:
    - Mr. Sin, aka: the Peking Homunculus, a robot controlled by a pig's brain, from the Talon's of Weng Chiang. Mr. Sin probably matches the experiment in the article the most closely, because he was a robot controlled by a portion of an animal's brain.
    - Morbius, from the Brain of Morbius. Does everyone remember the talking brain in a jar?
    - The Genesis of the Daleks shows that Dalek's are fully formed aliens and can kill people without being inside the "Travel" machines. As can be also be seen in the early episodes that "made" Doctor Who, the Daleks.
    - The Attack of the Cyberman shows clearly how the Cyberman evolved by replacing body parts with metal parts, until they were just human brains controlling robots. Of course, the new series updates the brains in a robot with more nasty implications.
    - Master. In the Planet of Fire, a miniaturized Master remotely controlled the robot Chameleon.
    - Rani. In Time and The Rani, the Rani made a organic brain into a giant Beowulf cluster of the world's brightest minds. Her goal was to destroy the universe and a variety of other evil. Trouble ensued when the Rani realized that the Doctor was slightly crazy ...

    Other movie references are:
    - Saturn 3 - a movie where someone grows a brain in a jar and puts it in an 8 foot killer robot. Oh, did I mention that the personality imprint for the brain comes from a killer?
    - Robocop proof that good things could possibly happen if someone attaches a human brain to a robot, and the sequel, Robocop 2, that shows bad things are what normally occur.