David A. Robinson proposed that networks of neurons could integrate in a
variety of ways. The first way is that a collateral forms a strong excitatory (+) input back onto
itself. If an action potential arrived as input it would be sent as an output as
well as used to re-activate the neuron itself after a small delay. This would
produce a second action potential and a third etc.
A second way is the 'figure eight' network through two inhibitory synapses.
Two inhibitory connections in series can act like a net excitation.
These neural networks are the simplest examples of short term memory.
|