Theoretic models, as well as physiological data, suggest that neuronal circuits are able to maintain similar functionality with variable architectures. The organization of our clustered networks into connected circuits was self-executed by the Dinaciclib neurons and the glia cells. Consequently, the exact architecture of each neuronal cluster was different. In addition, our cortical cultures contained many cell types, each having distinct morphological and function features. For small clusters, this implies that the distribution of cell types was different for every cluster. It is also likely that the exact connectivity scheme of the cells within each cluster was different. Despite the above variability, all the clusters showed spontaneous persistent collective activity in the form of NBs with markedly similar features. This hints that almost every network, independently of its architecture and size, self regulates its activity to sustain persistent activity patterns. This assumption is supported by the well known existence of both redundant cellular mechanism that support synchronization, and homeostatic mechanisms that support activity regulation. We have also demonstrated that our small clustered networks exhibit persistent network-level oscillation in the range of 25�C 100 Hz. These frequencies are of particular interest as they are manifested in brain activity and are typically associated with functional properties such as temporal encoding, MK-1775 abmole sensory binding, and storage and recall of information. Oscillations were observed in most of the analyzed clusters, suggesting that they are a generic property of small neuronal populations rather than the outcome of specific network architecture. In addition, the oscillations were more prominent at the decaying phase of the NBs. Such delayed activation may suggest that the oscillatory state is the outcome of a collective dynamics process that has to evolve until oscillations appear. Alternatively, the time delay may be related to a delayed activation of a synchronizing mechanism. It was previously shown, both in experimental and in theoretical studies, that oscillations in the cortex are generated by a combination of network interactions and cellular mechanisms. More specifically, the combined action of recurrent excitation and modulating inhibition are required to produce the oscillations. In addition, gap junctions were shown to play an important role in synchronizing neurons during oscillations.