Neurons
- CAs are based on mammalian neurons.
- My CANT model is trying to be a computationally efficient version
of these neurons.
- Neurons fire if their activation passes a threshold. This sends
activity across its synapses giving activation to other neurons.
- They're leaky integrators. If they fire they lose all activation,
if they don't some leaks away.
- They fatigue. The more they fire, the higher the threshold. Rest
allows recovery.
- Some neurons are excitatory, some are inhibitory.
- Synapses tend to connect to nearby neurons.
- There are different types of neurons. (I'm currently working with
pyramidal, and chandelier.)
