no blue brains here

the brain is always compared to the most sophisticated technology. right now, it's computers. the new project blue brain is an attempt to model the mammalian cerebral cortex's neocortical columns using one of the fastest computers on the world (it's the size of four refrigerators).

this project is ultracool and i want to like it. however, computational analysis, no matter how fast, does not accurately simulate the brain's processing capabilities. even if technology could mimic over one million parallel neural columns, the computational structure is entirely different. it will not give is the "correct answers," just as the fastest and most sophisticated telephone switchboard, another old brain analogy, would not have.

coincidentally, yesterday i came across not only an article on blue brain (thanks dr. greider), but also an article stating that "human brain is no computer":

Upsetting a long-held theory, Cornell University scientists say the mind works in a continuous, dynamic process, not in a series of distinct stages like a computer.

the point of blue brain is to, within three years, be able to use it as an animal-free medical model to test psychopharm drugs. peta would be proud, but i worry about basing any judgments on an inherently flawed model. blue brain may be able to display some of the outputs of a brain, like deep blue did to beat kasparov, but the inner workings are completely different.

may a next-generation synthetic processor someday be able to do this? maybe, but i won't hold my breath. we should instead spend our time and money working with the best processor there is - the actual brain.

Labels: neuroscience