6 ms·

Amputee controls prosthetic arms by thinking about them [video]

74 points by taf2 12y ago

nchelluri 12y ago
Ok, that was fucking cool.
fsloth 12y ago
Holy crap, after 40 years without hands he can pick objects and manipulate them with relative dexterity. I wonder could the hands be scaled up to a point where he could drive a "man-dozer" and be a one man construction company :)
fsloth 12y ago
Summary of the technique: 1. Surgically reconfigure amputees nerves that once controlled the arm and the hand so they are compatible with their pattern recognition scheme 2. Teach the pattern recognition system to respond to patients neural stimuli to said nerves : “We use pattern recognition algorithms to identify individual muscles that are contracting, how well they communicate with each other, and their amplitude and frequency,” Chi explained. “We take that information and translate that into actual movements within a prosthetic.” 3. Attach the robotic hands to the patient.
netcan 12y ago
Here is what I keep thinking when I see these prosthetics. After you train to control the arm while it's attached, can you more it around regardless of it being actually connected to your shoulder (assuming of course the sensors stay in place) and move it around like thing.
- mparlane 12y ago
  Must not have watched the video: https://www.youtube.com/watch?v=9NOncx2jU0Q#t=250 https://www.youtube.com/watch?v=9NOncx2jU0Q#t=250 This part of the video shows the arms are not attached. I'm sure he was trained to use the arms when they were not on him first.
opendomain 12y ago
This is amazing. There is no direct connection to his brain, so he may not b controlling the arms by just 'thinking'. It looks like almost using smaller muscles to control the arms, but using the nerve connections themselves. Also, he controls one pivot per action - so he moves the shoulder, then the forearm, then the hand, then the fingers with rests between each step.
- HackinOut 12y ago
  This is the one thing I didn't get: if it's mapped to nerves why can't he do two pivots at a time? Is it a limitation of the robotics?
  - bpicolo 12y ago
    It's mapped to the base of nerves. He doesn't have the entire nervous network that an arm usually has.
HackinOut 12y ago
Despite the cool robotics and brain control (nerves to be exact), seeing that guy do the one practical thing, picking up a glass and trying to bring it to his lisp after 40 years, was actually the coolest thing.
robg 12y ago
Amazing that 15 years ago I was watching primates do these things and more with neural implants. Perfect instance where the idea is amazing and revolutionary but the reality is so hard it takes decades if not longer to accomplish.
- innguest 12y ago
  The reality is if we dumped as much tax money on it as we dump on spaceships landing on comets, we'd be all done with this in 1 year and headlines like this would be old news.
  - randyrand 12y ago
    What would be the point of dumping tax dollars into amputee research? We invest in space because we think it is worthwhile, and still mostly unprofitable (that's changing). That is not the case with prosthetics. It makes sense to let the market decide how much to invest in this research acoording to the size of the amputee prosthetic market. It frees up the government to work I. Other things and gets rid of the HUGE inefficiencies when it comes to the government trying to accomplish anything.
    - innguest 12y ago
      > What would be the point of dumping tax dollars into amputee research? Helping real people that need help now. > We invest in space because we think it is worthwhile, and still mostly unprofitable (that's changing). That doesn't help real people that need help now. Remote-controlled robots on comets, versus more affordable and better prosthetics for all. If the whole country could vote I doubt they'd choose the former over the latter. So who is "we" in "We invest in space because we think it is worthwhile"? I'll turn your next line around: It makes sense to let the market decide how much to invest in this research acoording to the size of the space exploration and satellite launching market. It frees up the government to work on other things and gets rid of the HUGE inefficiencies when it comes to the government trying to accomplish anything. I like how you say "and still mostly unprofitable (that's changing)." You're deluding yourself if you think new industries go through a "not yet profitable" stage where they need help from the government and then afterwards they can be set free.
milesf 12y ago
My lifelong dream of becoming Dr. Octopus is one step closer :)
ciokan 12y ago
One (close) day such things will be in fashion magazines
fantanfantan 12y ago
I am excited for these kinds of technology for non-amputee use too: https://www.thalmic.com/en/myo/ https://www.thalmic.com/en/myo/
- koevet 12y ago
  Received a myo yesterday. Pretty awesome hardware execution - aka it works consistently - but the "things" you can do out-of-the-box are fairly limited. I will dig into the SDK over the weekend.
orthecreedence 12y ago
It must be incredibly hard to remember how to even use your arms 40 years after not having them. I'm not a neuroscientist, but I would expect that sooner or later the brain would start to rewrite the paths devoted to arm movement to something else. Looks like this guy made a ton of progress though. Very encouraging to see.

74 points by taf2 12y ago

Ok, that was fucking cool.

Holy crap, after 40 years without hands he can pick objects and manipulate them with relative dexterity. I wonder could the hands be scaled up to a point where he could drive a "man-dozer" and be a one man construction company :)

fsloth 12y ago

Summary of the technique: 1. Surgically reconfigure amputees nerves that once controlled the arm and the hand so they are compatible with their pattern recognition scheme 2. Teach the pattern recognition system to respond to patients neural stimuli to said nerves : “We use pattern recognition algorithms to identify individual muscles that are contracting, how well they communicate with each other, and their amplitude and frequency,” Chi explained. “We take that information and translate that into actual movements within a prosthetic.” 3. Attach the robotic hands to the patient.

netcan 12y ago

Here is what I keep thinking when I see these prosthetics. After you train to control the arm while it's attached, can you more it around regardless of it being actually connected to your shoulder (assuming of course the sensors stay in place) and move it around like thing.

mparlane 12y ago
Must not have watched the video: https://www.youtube.com/watch?v=9NOncx2jU0Q#t=250 https://www.youtube.com/watch?v=9NOncx2jU0Q#t=250 This part of the video shows the arms are not attached. I'm sure he was trained to use the arms when they were not on him first.

opendomain 12y ago

This is amazing. There is no direct connection to his brain, so he may not b controlling the arms by just 'thinking'. It looks like almost using smaller muscles to control the arms, but using the nerve connections themselves. Also, he controls one pivot per action - so he moves the shoulder, then the forearm, then the hand, then the fingers with rests between each step.

HackinOut 12y ago
This is the one thing I didn't get: if it's mapped to nerves why can't he do two pivots at a time? Is it a limitation of the robotics?
- bpicolo 12y ago
  It's mapped to the base of nerves. He doesn't have the entire nervous network that an arm usually has.

HackinOut 12y ago

Despite the cool robotics and brain control (nerves to be exact), seeing that guy do the one practical thing, picking up a glass and trying to bring it to his lisp after 40 years, was actually the coolest thing.

robg 12y ago

Amazing that 15 years ago I was watching primates do these things and more with neural implants. Perfect instance where the idea is amazing and revolutionary but the reality is so hard it takes decades if not longer to accomplish.

innguest 12y ago
The reality is if we dumped as much tax money on it as we dump on spaceships landing on comets, we'd be all done with this in 1 year and headlines like this would be old news.
- randyrand 12y ago
  What would be the point of dumping tax dollars into amputee research? We invest in space because we think it is worthwhile, and still mostly unprofitable (that's changing). That is not the case with prosthetics. It makes sense to let the market decide how much to invest in this research acoording to the size of the amputee prosthetic market. It frees up the government to work I. Other things and gets rid of the HUGE inefficiencies when it comes to the government trying to accomplish anything.
  - innguest 12y ago
    > What would be the point of dumping tax dollars into amputee research? Helping real people that need help now. > We invest in space because we think it is worthwhile, and still mostly unprofitable (that's changing). That doesn't help real people that need help now. Remote-controlled robots on comets, versus more affordable and better prosthetics for all. If the whole country could vote I doubt they'd choose the former over the latter. So who is "we" in "We invest in space because we think it is worthwhile"? I'll turn your next line around: It makes sense to let the market decide how much to invest in this research acoording to the size of the space exploration and satellite launching market. It frees up the government to work on other things and gets rid of the HUGE inefficiencies when it comes to the government trying to accomplish anything. I like how you say "and still mostly unprofitable (that's changing)." You're deluding yourself if you think new industries go through a "not yet profitable" stage where they need help from the government and then afterwards they can be set free.

milesf 12y ago

My lifelong dream of becoming Dr. Octopus is one step closer :)

ciokan 12y ago

One (close) day such things will be in fashion magazines

fantanfantan 12y ago

I am excited for these kinds of technology for non-amputee use too: https://www.thalmic.com/en/myo/ https://www.thalmic.com/en/myo/

koevet 12y ago
Received a myo yesterday. Pretty awesome hardware execution - aka it works consistently - but the "things" you can do out-of-the-box are fairly limited. I will dig into the SDK over the weekend.

orthecreedence 12y ago

It must be incredibly hard to remember how to even use your arms 40 years after not having them. I'm not a neuroscientist, but I would expect that sooner or later the brain would start to rewrite the paths devoted to arm movement to something else. Looks like this guy made a ton of progress though. Very encouraging to see.