Hi,
I am looking with doubt at some of assumptions made here:
'Neuron follows only very simple integrate-and-fire rules.'
The complexity theory tells us that small changes in agents behaviors can shows itself by an global vast change. The dynamics in neuron is much more complex. There are lots of details that maybe important. Integrate and fire is a level of abstraction. Your abstraction model is completely dependent to your aim. If you want to model a real brain i think you should not use the simplest level of abstraction.
'had on average 10^4 connections resulting in a 10^15 computational 'size' for the brain.'
I think we does merely not confronted with a set of 10^4 numbers. There are papers suggesting the spacial structure of brain have some computational significance. The connections also have their own dynamics. There are papers emphasizing these dynamics rule in whole system output.
I think the problem of modeling the brain is not a matter of computational power. It is a matter of lack of knowledge about the brain. We can not simulate brain simply because we do not know how exactly it is working. we can make assumptions about Brain and simulate our model and compare its output with real model, but we can not model real brain, at least now.
Best regards,
Habib Talavaty.
============================================================ FRIAM Applied Complexity Group listserv Meets Fridays 9a-11:30 at cafe at St. John's College lectures, archives, unsubscribe, maps at http://www.friam.org
