How Gibbs distributions may naturally arise from synaptic adaptation mechanisms. A model-based argumentation (Q841142)

In order to study the statistics of spike trains in a neural network and the effects of synaptic plasticity on the network dynamics, thermodynamics technology and formalism is applied. The neural state is introduced as solution for a deterministic dynamical model and one presents two models for the evolution of the membrane potential. Some basics in Spikes Dynamics and Statistics as well as Gibbs measures, Gibbs states are largely presented. One shows how Gibbs distributions are natural probability measures to describe the statistics of spike trains. The effect of the synaptic adaption in case of specific synaptic plasticity rules is described. The rules are formulated still by means of the thermodynamic formalism. A few practical examples and numerical results are shown.