Skip to content

ConceptSynapseModel

DavidFreely edited this page Nov 10, 2025 · 9 revisions

Related: ConceptLearning, ConceptNeuronModel, ConceptSynapse

List

Fixed Synapse Model

'Fixed' ConceptSynapseWeight ("no plasticity") of a synapse.

The default model for synapses (in BS).

Form the basic architecture of a circuit. Their task can be described as to "control" the signal flow in a predefined, stable manner and to define the structure of the network. They could be regarded as the fixed, unchangeable cabling of a system. If the “control” synapses were to change as well like the Hebbian synapses, the entire architecture of the network could become too complex.

Hebbian Synapse Model

(More generally: "spike-timing-dependent plasticity")

'Variable' ConceptSynapseWeight ("plasticity") of a synapse.

They are responsible for storing the actual information—the “content”— and forming/loosening associations (in a circuit).

Mechanism

  • If a neuron A targets another neuron B and 'does' fire immediately prior to the firing neuron B, we could assume that neuron A caused the firing of neuron B (or should-have if it didn’t). Therefore, the 'synapse-weight' should/will be 'increased'.
  • If a neuron A targets another neuron B and 'doesn't' fire immediately prior to the firing neuron B, then neuron A did not cause the firing of neuron B. Therefore, the 'synapse-weight' should/will be 'decreased'.

Result

  • Increase: "remembering" an apparently useful/probable causality of neuron A to B more.
  • Decrease: "remembering" the probable no-causality of neuron A to B more.

WeightAdaptionAmount

ConceptSynapseWeightAdaption.

Clone this wiki locally