ConceptSynapseWeight

Related: ConceptSynapse, ConceptSynapseWeightAdaption

The Amount of the voltage that is given over to the target neuron.

Whether it is a 'fixed' or 'variable' weight depends on the ConceptSynapseModel.

"As currently implemented, a 0.0-weight synapse can be brought to a weight of 1.0 with 11 spikes but returning it to zero requires 33. Using a 4 ms neuron cycle time, this means that it takes up to 44 ms to set a 0.0-weight synapse to an arbitrary value. While this timing is plausible, note that it restricts the synapse weight to only 11 discrete values."

"Synapse weights do offer higher precision the smaller they are, but, once again, this leads to a corresponding reduction in speed or increase in complexity as it takes progressively more incoming spikes to stimulate a neuron to fire."

"Note also that even when attempting to set a synapse weight to some precise value, there is no practical way, within the network, to learn what the precise value is. A synapse weight of 0.5 will cause a target neuron to spike on every other cycle—but so will a weight of 0.75 or 0.9. So you may observe that a neuron is firing every other cycle but this only gives a general indication of the aggregate incoming synapse weight, never precise values."

"Most networks rely on synapse weights which are significant relative to the threshold, so a small number of incoming spikes can cause the target neuron to spike. While we can assume that individual biological synapses have a much smaller maximum weight, a single high-weight synapse is functionally equivalent to a number of smaller-weight synapses in parallel."