Noise-induced extreme events in Hodgkin–Huxley neural networks

Abstract

Extreme events are rare, large-scale deviations from typical system behavior that can occur in nonlinear dynamical systems. In this study, we explore the emergence of extreme events within a network of identical stochastic Hodgkin–Huxley neurons with mean-field coupling. The neurons are exposed to uncorrelated noise, which introduces stochastic electrical fluctuations that influence their spiking activity. Analyzing the variations in the amplitude of the mean field, we observe a smooth transition from small-amplitude, out-of-sync activity to synchronized spiking activity as the coupling parameter increases, while an abrupt transition occurs with increasing noise intensity. However, beyond a certain threshold, the coupling abruptly suppresses the spiking activity of the network. Our analysis reveals that the influence of noise combined with neuronal coupling near the abrupt transitions can trigger cascades of synchronized spiking activity, identified as extreme events. The analysis of the entropy of the mean field allows us to detect the parameter region where these events occur. We characterize the statistics of these events and find that, as the network size increases, the parameter range where they occur decreases significantly. Our findings shed light on the mechanisms driving extreme events in neural networks and how noise and neural coupling shape collective behavior.