A direct approach to neuronal connectivity

In this paper, we present a direct approach to characterize the potential of neuronal cells for connectivity by investigating the distribution of synaptic connections and cluster formation in simulated neuronal networks. The influence of the cell morphology onto the overall connectivity of the network is estimated through a set of novel measurements immediately related to the observed number of connections and the distribution of the size of the obtained clusters of connected cells. Such measurements provide interesting indication about the potential of the cell for connectivity and about critical phase transitions of the network formation dynamics, characterizing three distinct regimes. Because connectivity is closely related to neuronal function, the proposed functionals become particularly relevant for characterizing the morphology and respective dynamics of different classes of neuronal cells. Such a potential is corroborated through the application of the proposed methodology over images of real neuronal cells, namely cat retinal ganglion neurons, allowing good separation between the different types.