Brain signal data are inherently "big" and challenging: massive in amount, complex in structure, high in dimensions, and low in signal to noise ratio. As a result, the integration of multiple data modalities to efficiently analyze multi-modal brain data is of prime importance. In this talk we present a joint analysis of brain connectivity with other data modalities, where the main focus is brain connectivity estimated by electroencephalogram (EEG) coherence at two conditions: rest and a reinforcement learning experiment. Our initial assessment of brain connectivity suggests that coherence (EEG) and function connectivity (fMRI) produce correlated results, which provides a solid justification of data integration. Integrating with genetic, behavioral, and fMRI data, we then find that different frequency bands of coherence show varying degrees of heritability. Finally, we conduct a systematic investigation on how genetic and brain signal jointly affect human behavior.