We used magneto-encephalography to study the temporal dynamics of band-limited power correlation at rest within and across six brain networks previously defined by prior functional magnetic resonance imaging (fMRI) studies. Epochs of transiently high within-network band limited power (BLP) correlation were identified and correlation of BLP time-series across networks was assessed in these epochs. These analyses demonstrate that functional networks are not equivalent with respect to cross-network interactions. The default-mode network and the posterior cingulate cortex, in particular, exhibit the highest degree of transient BLP correlation with other networks especially in the 14–25 Hz (β band) frequency range. Our results indicate that the previously demonstrated neuroanatomical centrality of the PCC and DMN has a physiological counterpart in the temporal dynamics of network interaction at behaviorally relevant timescales. This interaction involved subsets of nodes from other networks during periods in which their internal correlation was low. ► MEG resting state networks (RSNs) can be identified by BLP temporal correlation ► DMN is a core of BLP interaction with other RSNs especially in the β band ► RSNs with stronger internal coupling tend to interact less with other RSNs de Pasquale et al. investigate the dynamic segregation and interaction among resting state networks showing that the default mode network and the posterior cingulate cortex are highly connected with other resting state networks via the β band.