Dissing our educational system seemed to be oh so fashionable these days. But, when the University of Granada says they are developing a 'robotic cerebellum' to help people with Alzheimer's, I think we're doing just fine. (Okay, so they site Parkinson's, which still doesn't make sense, but is a little closer to the mark.)
The idea is interesting, though. The cerebellar cortex is the cytoarchitecturally the most well mapped out part of the brain, due to a very stereotyped pattern. Simulations of the cerebellum have been worked on for years (I remember making a crude model on a Purkinje cell in my Comp Neuro class back in 2000), and this seems like the next logical step. While ataxia is the most notable symptom of cerebellar damage, there are distinct cognitive deficits as well. Whether these are due to the influence of cerebellum on other parts of the brain, or appears as an epiphenomenon (I can't coordinate my reaching for objects, which makes me depressed and irritable) has yet to be completely worked out.
Current self balancing robots, like the Segway and less well-known smaller projects like nBot, implement a Kalman filter to remain balanced and avoid tipping over. While the benefit of an artificial cerebellum in humans is self-evident (the cerebellum is one of the most likely places for a stroke to occur), the application in pure robotics is dubious. Might make for a nice smart-prosthetics add-on, though.
From Medgadget.
No comments:
Post a Comment