***an initiative started by the EFPL (École Polytechnic Fédérale de Lausanne) in 2005.
** Goal: to use a computer to model the human brain. Ideally, the result would be a virtual brain that would behave just as an organic brain would.
**It could be a boon to neuroscience research.
**The Blue Brain Project uses one of IBM's top supercomputers -- the Blue Gene/L, which is among the top supercomputers in the world. It can do multiple trillions of operations per second. And for what do they need all of this computing superpower? It's needed to do such things as model a single structure in a rat's brain. The structure is called a neocortical column (NCC), a clump of about 10,000 neurons in the cerebral cortex.
**Our own human brains are composed of millions of such columns. The project team's idea was to be able to model just that single NCC, and then if it could work artificially in the same manner as it would in a real human brain, they'd have a terrific proof of concept with which to move forward to the task of using a similar simulation on a human brain simulation.
**The Blue Brain Project, it turned out, was indeed able to replicate the NCC, and while that's exciting, the simulation doesn't run nearly as fast as a true, biological NCC. Even the IBM supercomputer isn't nearly enough computational muscle to achieve that -- and that's just for a single rat's NCC model! Replicating millions of them, at real-time speed, to make an artificial brain will take raw computing power that isn't yet available. What's needed, the team says, is more computing pop and more funding.
***Understanding information processing in the brain and its higher emerging properties is arguably one of the major challenges in the life sciences. Research at the BMI focuses on three main areas: i) Molecular neurobiology and mechanisms of neurodegeneration; ii) Molecular and cellular mechanisms of synapse and microcircuit function up to the behavioural level and including metabolic aspects; iii) Sensory perception and cognition in humans. In all areas,