Human Tests of Mind-Controlled Artificial Arm To Begin

kkleiner writes "The world's first human testing of a mind-controlled artificial limb is ready to begin. A joint project between the Pentagon and Johns Hopkins Applied Physics Laboratory, the Modular Prosthetic Limb will be fully controlled by sensors implanted in the brain, and will even restore the sense of touch by sending electrical impulses from the limb back to the sensory cortex. Last week APL announced it had been awarded a $34.5M contract with DARPA, which will allow researchers to test the neural prosthetic in five individuals over the next two years."

Comments

AWESOME!

[the_macman]

Excellent! First article I see after watching this. 2027 is only 17 years away!!! :D

Re:AWESOME!

[Kenoli]

The technology in ghost in the shell is quite a bit more advanced.
Full body prosthesis > arms that turn into guns.

"Joint Project"

[adamdoyle]

A joint project between the Pentagon and Johns Hopkins...

haha a joint project.

Re:"Joint Project"

[v1]

had to add that tag to it the moment I read the headline

Re:Invented by a star wars fan?

[scdeimos]

Actually it's the Modular Prosthetic Limb (MPL) - the Luke arm comes from a competitor Deka, which is owned and run by Segway inventor Dean Kamen.

Legal aspects?

[Dwonis]

I wonder if this is going to run afoul of... arms-control regulations?

/me ducks

Re:Legal aspects?

[fractoid]

the army

How much danger is there....

[mark-t]

...connecting something directly to the human brain?

What would happen if there was a malfunction and the current levels going into the brain for sensor feedback were unregulated?

Re:How much danger is there....

[HBoar]

That would lead to a zombie apocalypse, obviously.

Re:How much danger is there....

[stripathy]

Worst case scenario the neurons in that particular region would die. If the current stimulus was even higher then that local region of the brain would die and the person would no longer be able to sense his arm, or finger; but those sensations would likely regenerate because the brain is very adaptable.

Re:How much danger is there....

[cgenman]

We used to bridge current directly through people's brains for therapy, at pretty high power levels. That sucker seems reasonably resilient.

Some Issues with the Tech

[stripathy]

This technology is clearly very cool, but there's two major hurdles to overcome before everyone's running around with one of these.

1. Controlling the device. Currently scientists/doctors control these brain computer interfaces (BCIs) by implanting electrodes into the patient's brain and finding neurons which code for particular movements (arm up or ring finger down). As the output device gets more complicated, like the arm here, doctors need to find more and more neurons to represent each degree of freedom of the output device.

2. Quality of neural recordings degrade with time. The current shelf life of the electrode arrays used in these experiments is ~1-2 years because after implantation, the brain's immune system rejects the device and neurons which code useful information die or move elsewhere. http://en.wikipedia.org/wiki/Chronic_electrode_implants

Re:Some Issues with the Tech

[dgatwood]

I am not a neurosurgeon, but it seems to me that implanting probes in the motor cortex is probably not the best solution anyway. You already presumably have peripheral motor nerves coming off the spine and across your back to where the arm used to be (ignoring people who can't use limbs due to spinal damage for the moment). And peripheral motor nerves, unlike spinal nerves, don't suddenly stop controlling your arm and start controlling your leg or start controlling a different muscle in your arm, generally speaking.

To use a computer analogy, controlling implants with probes in the brain is like staying in a hotel in NYC and controlling the lights in your house in California by installing a box that introspect the packets that flow through a core router on a major Internet backbone in Cleveland. Ten hours later, you drive to D.C., and the packets that went to your house get routed through Detroit instead, so your house isn't controllable. As you get closer to your house, you encounter fewer possible paths that actually go to your house. Thus, the closer you are to the endpoint, the greater the likelihood that the packets are going to pass through your tap. By the time the packets get to your house, you can be pretty sure that they're intended for your house, or at worst, for somebody in your general neighborhood.

So by tapping the peripheral motor nerves, you'd reduce the number of problems you have to deal with down to one: the nerves near the implant site dying. And even in that case, repairs would be less dangerous; you would shorten the nerve in the person's shoulder area instead cutting into his/her skull. Also, the mechanisms that attack implants in the brain aren't present in the rest of the body, AFAIK, so you might have a whole different set of problems or you might have fewer problems, but it seems unlikely that you'd have the same problems.

Re:Some Issues with the Tech

[dgatwood]

I would think that this is where the DSP comes in.... You have a series of chemical signals coming in from specific nerves, and you turn each of these signals into a separate data stream. You ask the person, "Okay, try lifting your arm straight up." You then record what happens. Repeat for other actions to build up a rough map of what neuron does what. Then, you have the person try to use the limb, starting from various positions, and tell the person to do specific things, progressively tuning the amount of response for particular motions iteratively until the motion of the prosthetic limb feels natural. Easy? No. Quick? Also no. Possible? I don't see why not.

Re:Some Issues with the Tech

[plastbox]

Why does so many here seem to think this would be needed? I'm pretty sure you could hook up the prosthetic limb to more or less any signal picked up from the implants and send the patient to physical(/mechanical) therapy. The brain is easily the worlds biggest neural net, and incredibly flexible and adaptive. The way I figure, there is no effing way the sensomotoric correlations wouldn't emerge on their own with use and exploration.

Awesome stuff, with strange possibilities.

[CrazyJim1]

I love this stuff because people who lost limbs or are paralyzed can become fully functional if it comes to pass! This sort of thing inspires great hope. Still I think about strange mad scientist applications...

If an electrical connection can control an arm, how much longer until you can control a whole body?

Since its an electrical connection, it could also be a wireless connection so you could control things at a distance.

If you had a computer, it could control the body too.

If someone goes brain dead or a coma, a computer could use that body like a robot with the right wiring and WIFI.

Or what might happen is that it doesn't use people... The setup may use an animal instead.

Who wants a monkey butler with the brain of a computer? How about a spy cat?

I don't expect those things to actually happen because people have morality, but they could be possibilities. I think its more likely that robot bodies will be built by people, but this technology makes you wonder what strange things are possible.

Re:Awesome stuff, with strange possibilities.

[R3d M3rcury]

I don't expect those things to actually happen because people have morality [...]

You must be new here...

Re:Awesome stuff, with strange possibilities.

[kainosnous]

I don't expect those things to actually happen because people have morality,

People are not by nature moral. What most people call "morals" is really just a bunch of self serving excuses for doing bad things. The only influences, secularly speaking, that will keep this technology from being used in the most hideous ways are laziness, lack of sufficient resources, and greedy bickering between scientists, corporations, and bureaucrats.

Re:Awesome stuff, with strange possibilities.

[plastbox]

Why should input to this immense, self-organizing computer be any different than output? Read up on Sensory Substitution (or augmentation, or perceptual augmentation, or whatever you feel like calling it).

Just as you say, the brain would figure out how the arm worked if allowed to explore and test. The same thing is true about sensory information presented to the brain through the skin, as long as there is a correlation between the signals going out and the signals coming in. What's the reasoning behind thinking that dropping the wire from the skin to the brain and just "plugging it right in there" would make a difference? Be it the correlation between telling your body to turn, feedback from the inner ear and proprioception and feedback from the feelSpace belt, or the correlation between sending random signals to a prosthetic arm and observing what happens.. I'm pretty sure the brain would figure it out on it's own.

Re:Awesome stuff, with strange possibilities.

[tehcyder]

People are not by nature moral. What most people call "morals" is really just a bunch of self serving excuses for doing bad things. The only influences, secularly speaking, that will keep this technology from being used in the most hideous ways are laziness, lack of sufficient resources, and greedy bickering between scientists, corporations, and bureaucrats.

Yes, because we all know that it is only religious belief that enables ethical behaviour, and that all religious believers are truly moral people.
You are a fucking cunt on a tricycle.

Ghost in the Shell Tech

[jameskojiro]

This is just the first step, the next step will be interfacing a person's brain into a device for processing data, ie. A cyber brain. The first once will be about the size of a iPhone, but will be external and wirelessly connected to the brain implants, eventually the size will shrink where it will make sense to mount the thing inside of the head.

Re:Ghost in the Shell Tech

[Psaakyrn]

With plugs at the back of the neck so we can tap into our dreams...

No need for implants

[Clueless Nick]

Everybody must have seen this video on TED:

http://www.ted.com/talks/tan_le_a_headset_that_reads_your_brainwaves.html

If you can read the electrical impulses non-intrusively and with a lightweight headgear, and then use an adaptive algorithm to learn an individual's 'fingerprint' brainwave patterns, you can easily use the technology to control everything from powered wheelchairs to those cool animatronic prosthetics developed by the Japanese. Of course, you will also need some corrective algorithms so that empathically generated signals do not start to control the hardware ;)

Sense of touch

[Clueless Nick]

I was wondering about that ever since watching the robotic prosthetics on NHK and especially the said TED video. Would it be possible to tap into nerves on a patch of skin (e.g. where the missing appendage should have been) and 'train' the brain to read impulses there, rather than directly meddle with it surgically?

Sci-fi time.

Cigarette lighter inside the thumb

[Dr Max]

I’m one step closer to getting a super strong left arm; with a Cigarette lighter inside the thumb; fold away bottle opener; phone on the palm of my hand; and an electric screwdriver inside the ring finger.

Re:Invented by a star wars fan?

[mobby_6kl]

Yeah, they should have called it "The Stranger".

Re:Why the Pentagon?

[tsotha]

You're being overly cynical. Nearly all advances in reconstructive surgery and prosthetics have been driven by militaries over the years. These programs are explicitly for reconstruction and rehabilitation.

The super-soldier thing just isn't practical. For one thing, you could buy a couple tanks for the price you'd pay to wire someone up like that. For another, what would you do with your super-soldiers when they didn't reenlist or became unfit for duty? Then there are basic power/size/weight considerations. Far more practical would be an exoskeleton like HULC.

Re:Invented by a star wars fan?

[necro81]

the Luke arm comes from a competitor Deka, which is owned and run by Segway inventor Dean Kamen

Not competitor, exactly. When DARPA started the Revolutionizing Prosthetics project some five years ago, they created two independent development paths. DEKA was tasked with making the most advanced prosthetic arm available with current technology, while APL was tasked (primarily) with developing a neural interface for a prosthetic. APL also developed an arm, which they'll be using in their trails, but you don't hear as much about that. The division was primarily between applied engineering, leading to an actual product, and research translation that is a longer-term effort.

central nervous system vs. Peripheral nerves

[DrYak]

your proposition does not only make sense, but is even used in other experiments or products. Earlier prosthetic arms read signals from nerves and remaining fragments of muscles (mentionned in TFA). Also the HAL exoskeletton predicts which motion to assist by reading nerves and muscle.

BUT all this requires functionning nerves.
according to TFA, this artificial arm is intended for quadrplegic patients (with whom no useful brain impulse controls anything below the neck, except the main respiratory muscle)
for the intended patient, brain-computer-interfaces are the only way to go.

Also, a nerve requires a connexion to a muscle to function properly. You can't just put an electrode on it to read the signal. If the limb is missing, the nerve is un connected and dies of or degrades. That's why another artificial limb is mentionned to require renervation of muscles.
The golden target for non-quadraplegic patients would probably be to design which, to the body, exactly look like what the nerves grow onto, so the body will naturally make synapses to link the artificial limb.