Brain Interface Lets Monkeys Control Prosthetic Limbs

himicos was one of many readers to point out one recent success of scientists working to develop working brain-machine interfaces, writing "A team at the university of Pittsburgh has finally advanced a 2002 technology enough for use in prosthetic limbs, the targeted application all along. Training computer models to the firing patterns of the neurons in the parts of the brain that control motion, they are able to project the intentions of a monkey to a robotic arm, which follows the will of the animal. The sad thing about the articles is that the beauty of the mathematics used to create and train the models is totally ignored." Reader phpmysqldev adds a link to coverage at the BBC, and writes "This of course brings significant hope to amputees and other other people with physical disabilities." (Note that this research has been going on for quite some time.)

Comments

Forget Replacement Limbs...

[crymeph0]

How about custom appendages? If the brain can be trained to independently control a new arm, why couldn't it learn to control a genuine Doctor Octopus suit?

Re:Forget Replacement Limbs...

[PachmanP]

How about custom appendages? If the brain can be trained to independently control a new arm, why couldn't it learn to control a genuine Doctor Octopus suit?
Speaking from experience, it is because the grant money is better. If you say you need money to research brain/machine interfaces for prothetic limbs to help disabled people, you are more likely to get it than when you say you need the research to give yourself/your_cyborg_army superhuman appendages to be used for world domination.

Re:Forget Replacement Limbs...

[khayman80]

I would imagine the mental map we have of our bodies has four limbs. This would mean that, for purposes of sensation, motor control and proprioception, we can't operate more than four limbs at once. Why would we evolve the ability to control limbs that we don't even have? I mean, brains are flexible, but I would guess that trying to push the "body control/sensation/proprioception" map past four limbs may have some unintended (and possibly bad) consequences.

An alternative might be the use muscles in the face to control extra limbs. Frowning would perform one action with the prosthetics, smiling another, etc. But this would be considerably more clumsy than the intended use- replacing a limb that doesn't exist on the physical body, but does have a designated place in the brain that controls it.

Re:Forget Replacement Limbs...

[street struttin']

Speaking from experience, it is because the grant money is better. If you say you need money to research brain/machine interfaces for prothetic limbs to help disabled people, you are more likely to get it than when you say you need the research to give yourself/your_cyborg_army superhuman appendages to be used for world domination. You have GOT to be kidding! Getting government grants to find military applications for otherwise harmless things is a staple of the defense program. If you have an answer to "How many Commies/Terrorists can it kill", you've got grant money.

So go ahead and build your cybernetic superhumans to do your bidding, but you might have to sign a contract that says you'll do the bidding of the US government, too.

Re:Forget Replacement Limbs...

[Bones3D_mac]

I would imagine the mental map we have of our bodies has four limbs. This would mean that, for purposes of sensation, motor control and proprioception, we can't operate more than four limbs at once. Why would we evolve the ability to control limbs that we don't even have? I mean, brains are flexible, but I would guess that trying to push the "body control/sensation/proprioception" map past four limbs may have some unintended (and possibly bad) consequences.

Personally, I see this problem as more the case that we've only been conditioned to handle that many limbs over years of experience versus any sort of hard limit being imposed. (Not to mention it kind of runs across the grain of that whole "evolution" thing being needlessly debated...)

There have been numerous examples demonstrating that our brains are not only highly adaptive to new situations (such as the brain redistributing certain functions to different areas to overcome damaged areas), but are also highly receptive to new forms of input from external sources (such as invasive probing of the brain to create crude brain-to-computer interfaces to control simple devices, such as an on-screen cursor.)

The larger issue is really more of a case of creating a proper and convenient interface for cyborg-like add-ons. For example, do we necessarily have to invade the brain directly, or can we simply use existing connections by connecting jumper cables to the nerves running down the spine. And if that isn't an option, can we create or add extra, custom nerve sets to the spine and create connections to the brain that way?

Considering all that, a "third arm", or similar contraption is probably within the realm of possibility, but it may take time to adapt to and fine tune the system before it becomes effortless (or closer to that) to use. It's actually not all that dissimilar to the steps you have to go through for setting up a decent voice recognition system.

Re:Forget Replacement Limbs...

[TheLink]

"This would mean that, for purposes of sensation, motor control and proprioception, we can't operate more than four limbs at once."

I'm sure that's wrong.

We extend our mental maps to include vehicles, devices and tools that we operate on a regular basis. Believe me, some of us even feel pain when we ding our car on something. Some even feel pain if they get shot in a video game.

The fact that many people can be trained to see with their _tongue_ means the brain is very adaptable.

The Seeing Tongue:
http://findarticles.com/p/articles/mi_m1200/is_9_160/ai_78681631

Just because you start seeing with your tongue does not mean you lose sight in your eyes. So I do not believe that we are limited to controlling 4 limbs. When people use a tool they are skilled in, that tool becomes an extension of their body - and it does not even have to be physically connected to their body - ask people who do stunts with RC helicopters, or play FPS/RTS games.

Once you practice enough, it becomes learnt and integrated into your brain, you no longer think "Ah I must press this to do X", you just think "I need to go here" and you do whatever it takes to get it done.

A skilled typist does not think of each key stroke independently, the typist just thinks of the phrase (or sees stuff to type) and all the 8 fingers and 2 thumbs get it done. So controlling more than 4 limbs shouldn't be a huge problem.

However, just like when you concentrate on something a lot, say drawing an intricate design, you may lose awareness of what's going on with your little toe (until something significant happens to it, or even is about to happen to it - incoming object via peripheral vision - in which case the rest of your brain brings it to your attention).

Re:Explain the beauty?

[kalirion]

How would you explain the beauty of a sunset to the blind?

Monkey's opinion

[nategoose]

The monkey in the pictures had his own arms restrained within tubes so that he/she would be forced to use the mechanical arm in order to get the marshmallow, and the mechanical arm isn't oriented so that the monkey could possibly mistake it for his/her own arm. I can't help but wonder what the monkey's opinion of all this is. It's got to be more than a little confusing.

Re:sci-fi pondering

[somersault]

Do you merit all scientific advances purely by how much closer they're going to get us to the Star Wars universe? ;) I don't think that lightsaber mishaps are the only type that require the victim to use a prosthetic.

Re:Explain the beauty?

[hansraj]

I am amazed at the number of responses being just smug and claiming how you need to do math to appreciate the beauty. Reminds me of a guy doing PhD is Chemistry about effects of certain chiral isomer of nicotine on cancer. His first response when I asked what he worked on was "You won't get it". I am a PhD student in computational geometry and I frequently have to explain my work to relatives who have no idea about geometry. When I pestered the guy that whether or not he can explain his work to a layman reflects his understanding about his work, he agreed to try. Of course I could understand the central part once he replaced the technical name of the molecule with "a chiral isomer of nicotine". I am sure it could have been further simplified as "mirror image molecule of the stuff in tobacco" in case I didn't remember what "chiral" and "nicotine" are.

On the topic, I am not entirely sure about the exact math used in the said experiment but based on the fact that the link points to the notion of "information content", here is my guess how it should work (at least in principle). I will try just because no one else seems to. Feel free to correct me.

The state of the neurons of the relevant area of the brain (relevant for the goal in the experiment - say pick marshmallows or open the door) could be modeled as a random variable. The first problem when trying to figure out what a certain electrical activity in brain represents would be to figure out whether you are looking at a random electrical activity (brain doing lots of background work maybe) or some order (brain trying to focus and activate the subroutine for "move hand and open door"). This difference between order and chaos is captured in a neat formula describing the entropy or the information content of the random variable. Naturally, the less the entropy the more the order. I have no idea what possibly goes on after this step.

In any case, now coming to the "beauty" part. Of course you need an eye to appreciate beauty for the notion is quite subjective. The remarkable thing is that a simple formula captures the vague notion of "order" that we all have. The formula might not be the most beautiful thing because as I understood from the article, the log term is somewhat forced to make sure different things add up nicely. But then, one could think of this very fact (the extra log term) as a neat mathematical representation of the notion that disorder should be able to be combined with another disorder to create something bigger.

I hope my response is better than "drop whatever you are doing and go do a PhD in math before you can understand the beauty of math".

Re:And just like that...

[mpeskett]

You'd prefer they started out by sticking electrodes into humans with no idea what they were doing?

Of course some experimentation will be needed when they move to human subjects, but a monkey's brain is similar enough to ours that they can get a starting point to experiment around, rather than working blind on a human subject.

One other thing to note, there are no touch/pain receptors within the brain itself - people have brain surgery done while awake so the doctors can keep them talking and know they aren't accidentally removing something important. Once you've got an opening into the skull (which would be done under anaesthetic) you can poke and prod at the brain all you want without the subject feeling a thing.

Oh, and its on the news because its interesting and something of a step forward scientifically. Quit it with the conspiracy theories please.