Extended analysis of stochastic resonance in a modular neuronal network at different scales

Based on the firing dynamics at three different levels of microscale, mesoscale and macroscale, this study presents an extended analysis of stochastic resonance in a modular neuronal network in the spatially correlated white noise environment. Two well-defined modules of small-world subnetwork and scale-free subnetwork constitute the modular neuronal network in a hierarchical way. When a subthreshold periodic input is incorporated into this network, numerical results illustrate that a collective pattern of stochastic resonance emerges at macroscopic scale when the intensity of the correlated noise is appropriately tuned. Through extended analysis, one can detect that the firing rhythms of individual neurons gradually follow those of the periodic input at microscale. In addition, the occurrence of stochastic resonance at mesoscale in the small-world subnetwork is earlier than that in the scale-free subnetwork, and the peak height of resonance curve in the former subnetwork is remarkably higher than that in the latter one. These combined results indicate that the small-world subnetwork is more favorable than the scale-free subnetwork to induce stochastic resonance in this constructed modular network. The robustness of the extended analysis of stochastic resonance against variations in noise correlation coefficient and intra-module probability is also unveiled. This study provides a new perspective and tool to understand the collective phenomenon of stochastic resonance in realistic neuronal systems.