The analysis of spontaneous functional connectivity (sFC) reveals the statistical connections between regions of the brain consistent with underlying functional communication networks within the brain. In this work, we describe the implementation of a complete all-to-all network analysis of resting state neuronal activity from magnetoencephalography (MEG). Using graph theory to define networks at the dipole level, we established functionally defined regions by k-means clustering cortical surface locations using Eigenvector centrality (EVC) scores from the all-to-all adjacency model. Permutation testing was used to estimate regions with statistically significant connections compared to empty room data, which adjusts for spatial dependencies introduced by the MEG inverse problem. In order to test this model, we performed a series of numerical simulations investigating the effects of the MEG reconstruction on connectivity estimates. We subsequently applied the approach to subject data to investigate the effectiveness of our method in obtaining whole brain networks. Our findings indicated that our model provides statistically robust estimates of functional region networks. Application of our phase locking network methodology to real data produced networks with similar connectivity to previously published findings, specifically, we found connections between contralateral areas of the arcuate fasciculus that have been previously investigated. The use of data-driven methods for neuroscientific investigations provides a new tool for researchers in identifying and characterizing whole brain functional connectivity networks.