Paralyzed woman uses thoughts to move robotic arm

[Kholdstare]

http://www.reuters.com/article/2012/05/16/us-paralysis-robotics-idUSBRE84F1CM20120516

This is a remarkable achievement by the scientists. Now, we don't need limbs any more? :P

To manage the feat, researchers used a tiny sensor about the size of a baby aspirin that is loaded with 96 electrodes designed to pick up nerve activity.

Scientists implant that into a part of the brain called the motor cortex that is involved in voluntary movement.

When a paralyzed patient implanted with the device thinks about moving an object, the electrodes pick up those nerve signals and send them to a nearby computer, which then translates them into commands to operate assistive devices, such as the robotic arms used in the study.

So far, the experimental device has been used to allow patients to operate a computer cursor and control simple robotic devices.

I wonder how did the scientists find out where to connect each electrodes? How did they know which portion sends what? Also, how did they interpret the brain signals and know which command the user wanted to perform?

 

[Drakkith]

I would guess from experiments on people prior to this. We can already see what the brain is doing using non-invasive techniques such as MRI's and PET scans.

 

[I like Serena]

Yep. Remarkable!

I've seen before that when a number of nerves to a limb have been cut, surgeons don't even try to reconnect the nerves properly.
They just reconnect them randomly.

At first the patient can't properly move his limb any more, but after training the brain rewires itself and the patient regains control over his limb.

I expect the same principle applies here.
As long as the electrodes pick up something, and they are picking up different signals, all is well.

 

[georgir]

The story on physorg's partner site:
http://medicalxpress.com/news/2012-05-paralyzed-individuals-robotic-arms-grasp.html

It is old news in many ways. Research in the field has been going on for decades now, with the first successful human trial in 2004. Google up Matt Nagle.
Even one of the subjects mentioned in this specific article, "S3", has been using the implant for so long now that they are probably bored out of it ;) Last year they had the 1000 days milestone, which means they had it implanted around 4 years ago:
http://phys.org/news/2011-03-braingate-neural-interface-day-milestone.html

What is new in this case is the fact that the subject is controlling an actual robotic arm in 3d space and not just a computer cursor. Not a bad development, but not as exciting considering the implant itself is the same old tech with very little improvements.

As to your curiosity about how this works, yea, it is a pretty blurry topic. They don't really need to hit a precise location in the brain, just somewhere in the motor cortex in general. Then they train the computer by having the subject imagine moving their hand and recording the patterns. I seem to recall that re-training and re-calibrating it was needed quite often in Matt Nagle's case, for example, but I don't know if that's still the case with the newer trials.

I often wondered if it wouldn't be easier and better to do the reverse, make the computer react to a fixed set of signals and leave the brain to adapt and learn. As "I like Serena" said, brain plasticity is quite amazing.

 

[I like Serena]

What is new in this case is the fact that the subject is controlling an actual robotic arm in 3d space and not just a computer cursor. Not a bad development, but not as exciting considering the implant itself is the same old tech with very little improvements.

Not as exciting?
It's very exciting as far as I'm concerned!

It's one thing to pick up the signal and move a mouse cursor with a tremendous amount of effort and many mistakes in the process.

Or to just be able to move a robotic arm as it pleases you.

The next step to obtain a working exoskeleton that you can manage with your brain seems to become feasible.
With something like that it won't matter anymore whether you are physically handicapped or not.
In some aspects you'd even be better off with such an exoskeleton.

I often wondered if it wouldn't be easier and better to do the reverse, make the computer react to a fixed set of signals and leave the brain to adapt and learn. As "I like Serena" said, brain plasticity is quite amazing.

I image that at first you might try to let the computer adapt to what the brain is already used to.
Then you might "freeze" the computer logic and let the brain adapt to the system as it is.

I think that usually you won't want a computer system that is "too" smart, because that is counterproductive.
Keep it as simple as possible and let the human brain adapt, since the human brain is much better suited to that than any computer.

 

[atyy]

As to your curiosity about how this works, yea, it is a pretty blurry topic. They don't really need to hit a precise location in the brain, just somewhere in the motor cortex in general. Then they train the computer by having the subject imagine moving their hand and recording the patterns. I seem to recall that re-training and re-calibrating it was needed quite often in Matt Nagle's case, for example, but I don't know if that's still the case with the newer trials.

I often wondered if it wouldn't be easier and better to do the reverse, make the computer react to a fixed set of signals and leave the brain to adapt and learn. As "I like Serena" said, brain plasticity is quite amazing.

I think that usually you won't want a computer system that is "too" smart, because that is counterproductive.
Keep it as simple as possible and let the human brain adapt, since the human brain is much better suited to that than any computer.

Possibly relevant leads:
http://www.ncbi.nlm.nih.gov/pubmed/18846203
http://www.ncbi.nlm.nih.gov/pubmed/20573887
http://www.ncbi.nlm.nih.gov/pubmed/18923392

 

[georgir]

Not as exciting?
It's very exciting as far as I'm concerned!

It's one thing to pick up the signal and move a mouse cursor with a tremendous amount of effort and many mistakes in the process.

Or to just be able to move a robotic arm as it pleases you.

Sorry, but your comments regarding the difficulty and success of the control are not making sense. There is no reason it should be different in the two cases, learning to use the system is not related to whether you are controlling a physical or a virtual object.

I can see how the physical variant can be perceived as "more real" and much more impressive when one first hears about the topic, and I definitely understand the much stronger appeal of it for handicapped people hoping to one day have restored movement thanks to this technology.

On the other hand, monkeys have controlled robotic arms since 2008.
http://www.nytimes.com/2008/05/29/science/29brain.html

You also saw how long ago experiments with controlling a computer have been carried out from my previous post. You have to admit that for someone keeping track on this field, this news might not be that big of a deal.

I don't want to make the wrong impression here - I am quite excited by this kind of technology, and I am glad to hear about the continued work in the field and the progress they are making.

I just don't think you, or a lot of other people for that matter, are completely grasping the potential of the technology and what the really exciting aspects of it are.

For the geek that I am, the virtual applications of this technology have much more potential. Contrary to what one might think as a first impression, they are not at all limited to controlling a 2d cursor. Simultaneous control of multiple complex moving parts, objects, viewpoints in virtual space is much easier, cheaper and safer to experiment with than controlling real hardware.

And its applications can be far greater too. I know most people's first association is just playing games and such, but it can be very useful in computer modelling, design, art. A virtual application doesn't even have to be related to movement as such, but can be related to a change of color, sound tone or whatever, and it's still equivalently easy to control. Its impact on the existing arts would be very interesting to see. It may even lead to some interesting new form of communication/interaction bewteen people. Frankly, I think the potential is so huge that no one can even completely predict it yet.

 

[moonman239]

Tomorrow's headlines: "Lady with robotic arm terrorizes town"

Seriously, though, it's really cool that scientists have developed a robotic arm that can be controlled by the brain.