Synaptic delay induced macroscopic dynamics of the large-scale network of Izhikevich neurons

Abstract: We consider a large network of Izhikevich neurons. Each neuron has a quadratic integrate-and-fire type model with a recovery variable modelling spike frequency adaptation (SFA). We introduce a biologically motivated synaptic current expression and a delay in the synaptic transmission. Following the Ott-Antonsen theory, we reduce the network model to a mean-field system of delayed differential equations. Numerical bifurcation analysis allows us to locate higher-codimension bifurcations and to identify the regions in the parameter space where the network exhibits changes in the macroscopic dynamics, including transitions between states where the individual neurons exhibit asynchronous tonic firing and different types of synchronous bursting. We investigate the impact of the heterogeneity of the quenched input current, the SFA mechanism and the synaptic delay on macroscopic dynamics. In the limit that the heterogeneity goes to zero, our perturbation and bifurcation analysis shows that the behaviour of the mean-field model remains consistent, although this limit breaks an assumption of the model reduction. For a single population of neurons with SFA, the synaptic delay has little effect on the generation of COs for weak coupling, but favours their emergence beyond that, and even induces new macroscopic dynamics. In particular, Torus bifurcations may occur, and these are a crucial mechanism for the emergence of population bursting with two nested frequencies. We discuss how these solutions may relate to cross-frequency coupling which is potentially relevant for understanding healthy and pathological brain functions.