Neural, behavioral and social phenomena are intrinsically multi-scale in space. But their dynamics also span different scales in time. A major challenge for social neuroscience is to understand how cognition and behavior are linked to the coordination of these highly complex spatial-temporal patterns that exist within and across different levels of observation. Hyper-scanning studies have already shown how the coordination of multiple humans engaged in a social interaction is accompanied by the coordination of their brain activities: (Dumas et al. 2010) demonstrated that states of interactional synchrony at the behavioral level correlate with the emergence of inter-individual synchronization at the brain level across multiple frequency bands. These results concerned average measurements across either subjects or time, and provided crucial descriptive data on which we now attempt to build an explanatory, integrative account of social behavior. In the current study and with the same data, we focused on how these interbrain couplings vary in time with respect to one another and in relation to behavior. In particular, we sought to determine how intra- and inter-individual functional connectivities (e.g. in alpha, beta and gamma frequency bands) are modulated by slower timescale activity. An on-line dynamical account uncovered similar temporal structures in brain patterns within and between subjects such as modulation by a common rhythm at slower frequency, suggesting a potential relationship with ongoing behavior. Such interleaving of neural, behavioral and social levels was also supported by a detailed investigation of the different behavioral transitions that occurred during real-time social interactions. Though challenges remain, taken together our results demonstrate the power of an integrative approach to understanding the neurobehavioral mechanisms involved in social interactions.