We have examined the effects of nearest-neighbour gap-junction coupling in computer simulations of a neuronal network of noisy subthreshold oscillators (10*10 neurons) at tonic-bursting transitions and in the chaotic regime in between. When the neurons are operating in a regularly tonic or burst firing mode they are going into in-phase synchronization already at very low coupling strength without significant changes of the impulse patterns. The situation changes close to the bifurcation point where an otherwise synchronized network transiently exhibits chaotic activity at intermediate coupling strengths. Also a network of chaotic neurons goes into a synchronized state but needs higher coupling strengths. Moreover, the network can shift between in-phase and out-of-phase synchronization in irregular, unpredictable intervals. Remarkably, the interval distributions and return-maps of individual neurons look rather noisy even in completely deterministic simulations. On the other hand, the effects of noise seem to be clearly less significant in coupled neurons compared to often dramatic noise effects on single neurons dynamics. Further simulations shall elucidate the impact of synaptic coupling and the variability of individual neurons.

Supported by the European Union through the Network of Excellence BioSim, Contract No. LSHB-CT-2004-005137.