Synchronized bursts following instability of synchronous spiking in chaotic neuronal networks

We report on the origin of synchronized bursting dynamics in various networks of neural spiking oscillators, when a certain threshold in coupling strength is exceeded. These ensembles synchronize at relatively low coupling strength and lose synchronization at stronger coupling via spatio-temporal intermittency. The latter transition triggers multiple-timescale dynamics, which results in synchronized bursting with a fractal-like spatio-temporal pattern of spiking. Implementation of an appropriate technique of separating oscillations on different time-scales allows for quantitative analysis of this phenomenon. We show, that this phenomenon is generic for various network topologies from regular to small-world and scale-free ones and for different types of coupling.