Wednesday, August 8, 2007


I'll post more on this later (post prelim), but Wired has a nice overview of the DARPA arm project. They make mention of the Deam Kamen arm as well, and that the current control mechanism for both is myoelectric (EMG pads recording rerouted, off site, or residual muscle movement).

I just thought I'd mention that the reasons many of my posts are relate to robotic prostheses are that:
1) They make for good media, so they get plenty of coverage
2) They represent something of a parallel but separate line from what I do (decoding = neuroscience, prosthesis = engineering)
3) Aside from implantable stimulation systems, artificial prosthetics are the Mecca of neuroprosthesis. Cursor movement is great, but it doesn't tie your shoe. Additionally, they provide the possibility of extending the human body. In Sci-Fi Fantasy Future Land, there is always the possibility of having more than two arms (you laugh, but there's no evidence that this would not be possible!)


IConrad said...


And evidence that it is possible, as I'm sure you know: the neuroplasticity of tool incorporation as demonstrated by a three (2 meat, 1 robotic) armed monkey ring any bells? :)


Drew said...

Nicolelis has been able to demonstrate that a primate can "retune" neurons in its motor/premotor and dorsal parietal cortices to control a third arm independently of its two biological arms.

Plasticity in the nervous system is the rule, not the exception.

Brandon King said...

Yup and yup. Remind me to name my band "Two Meat, One Robot". I like the ring of that.

Nicolelis, and others, have shown that an arm-like effector can be controlled while the biological arms remain motionless. But, there has not been a series of consistent metrics to say, "Yes, this monkey is controlling that arm with the same accuracy as his biological arm." Researchers generally take the 'good trials' and present them as the best possible case.

Also, whether these systems are performing at the same level of accuracy and speed as their arms, without impacting accuracy of other parts of the body has yet to be seen. Maybe your general accuracy is reduced because you have less cortical mass available to process each arm.

And lastly, these experiments are done with arms without hands. Reconstructing finger movements is far more complex.

Post a Comment