Limits and dynamics of stochastic neuronal networks with random heterogeneous delays (Q1938840): Difference between revisions

This is a study on the dynamics and limits of large-scale neural networks in the presence of noise. The considered neural network is modeled by a system of stochastic differential equations on the space of continuous functions. In case the initial condition is a square integrable stochastic process and the coefficients of the model fulfill some special conditions (Lipschitz continuity, growth conditions), the network system accepts a unique strong solution. For translation invariant as well as for non translation invariant systems the mean field equations are derived. The limit equations when the number of neurons \(N\) goes to infinity, in both cases, build a system of delayed McKean-Vlasov equations with distributed delays. Under same assumptions on the model coefficients (Lipschitz continuity, growth conditions) and for square integrable stochastic initial condition, the existence and uniqueness of a strong solution for the delayed McKean-Vlasov equations is proved. Further, convergence in law of the solutions of the network equations towards the solution of the delayed McKean-Vlasov equations is proved too. In the last part of the article the author concentrates on the study of the dynamics of the firing-rate model. Mean-field equations are derived and the impact of the distribution of delays on the dynamics of the network is largely discussed.