Discussion of CA weaknesses
- What is a good CA model
- Stochastic Models: This is an abstraction away from neurons; the models
state that CAs behave such and such way, but it is not clear that biological
CAs do behave that way.
- Synfire chains (bursting neurons): this is better. There is some work
to show they are the same as neurons but we don't really know what neurons
do at this level.
- Loosely Simulated Neurons: this is what I do. I hope I'm catching the
essential properties
- Simulated Neurons: computationally very expensive. Do they do too much?
- Weaknesses of my model
- Am I catching the essential properties of neurons
- Too few neurons and connections
- Simplified Topology
- Poor (absent) interface with the environment
- How big is a CA
- a part of a cortical column
- large areas of the brain
- both
