Slashdot Mirror
Scientists Discover What You Are Thinking
neurospace writes "Caltech scientists have successfully decoded movement plans from the brains of awake humans. This work has direct application to the development of a neural prosthesis, a brain-machine interface that will give paralyzed people the ability to move and communicate simply using their thoughts. The lead scientist on this project will be interviewed on Sunday, March 20, on the SETI Institute's weekly radio show, 'Are We Alone?'"
Good as this is, what I'm really waiting for is a way to tap into the language center. Imagine: an interface which can work, regardless of the language spoken by the person. No more need for translation, everything could be held in a form identified directly by the brain.
But I doubt this will happen in my lifetime.
If we could find the mechanical outcome of what we think when we listen to music.
... and I shall strike upon thee with great vegeance, furious anger and a slightly positive karma.
Since the brain seems to adapt its structure to suit its environment (such as giving someone partial "vision" by stimulating their back with an array of little elements which correspond to the pixels on a camera), won't it mean that different people will have slightly different "wiring" for this to totally work on everyone? On the other hand, since the brain is somewhat adaptive, maybe you can get the brain to adapt or to learn to communicate with the target electrode areas...
I have had an idea for years, Im not sure if it is possible technically or phsyically because I dont have a fantastic understanding of the human brain. I also have no idea whether anyone has thought of this before but here goes.
One of the nice things about neural networks is that you dont neccesarily have to understand processes that occur during translation. I have often hypothesised that it might be possible to use the traits of a neural network to create an interface with the brain. Suppose there was a patient who had a degenerative eye condition that meant in 10yrs he or she would be completely blind. Forgetting the implications of connecting wetware to hardware for a moment- imagine if we could use a neural network to interface with the visual cortex of a patient , to learn to understand the electrical impulses on the patients visual cortex by way of matching them up with a camera mounted on the side of the head. Might it be possible for the patient to look at a tree using his real eye - the nueral network sees the tree with its camera and this way "Learns" what the patterns in the cortex represent.
Something like this (if it is possible) would have some quite phenomenal implications - especially if it were possible to "playback" the patterns into the cortex from the camera.
Would anyone who knows a bit more about these subjects care to discuss the possibilities of something like this?
Electronic Music Made Using Linux http://soundcloud.com/polyp