Cell Assemblies
- If enough neurons in the CA are fired by an external source
(e.g. senses), there is a cascade of activation causing a large
percentage of the neurons to fire.
- These neurons, in turn, cause other neurons to fire and
the circuit will remain active.
- This is a Short term memory item and will remain active after
external input ceases.
- Long term memory is formed via learning, and in general you
can not activate a short term memory that does not have a
long term one.
- This means that a range of input patterns will ignite a CA, and
that a CA is a categoriser.
- We use Pearson's product to measure reliable activation, uniqueness and
persistence
- Show the
simple system
- Benefits of CAs
- Short Term Memory
- Long Term Memory
- Figure Ground Separation
- Variable Binding
- There are a range of complex behaviours that humans must do that
you can theoretically describe using the CA model
- Existing solutions (mammals) to a
range of unsolved problems are closely related.
- There has been a fair amount of closely related research (e.g. attractor nets, neurology, fMRI, psychology, and machine learning).
