Synchronization in networks of noisy interneurons

It is generally believed that rhythm synchronization of different neuronal regions of the brain are of paramount importance to understand the perception problem.1 A particular type of synchronous neuronal oscillations, which may be linked to the binding of different brain regions, occurs in the 20–80 Hz frequency range, known as gamma oscillations.2 Recent advances in the understanding of this problem have indicated that random networks of GABA-ergic interneurons are capable of producing the gamma oscillations under physiological conditions.3 This synchronization by mutual inhibition was reproduced in computer simulations using physiologically realistic models.3 In the experimental preparation there is always some heterogeneity in neuronal properties, as well as synaptic noise. An important issue is therefore whether the mechanism suggested by these simulations is robust. In a recent paper Wang and Buzsaki (WB),4 and also White et al,5 have carried out computer simulations of a network of interconnected GABA-ergic interneurons. The synchronization was found to be only moderately robust against heterogeneity in the distribution of driving currents. In this work we consider the stability of this type of gamma oscillations against synaptic noise. In addition, we consider the spatio-temporal coherence of a set of interneurons with specific spatial connectivity properties. To analyze the results quantitatively we calculate measures that characterize the network’s temporal coherence as well as its spatial synchronization.