Concurrent EEG- and fMRI-derived functional connectomes exhibit linked dynamics

Large-scale functional connectivity of the human brain, commonly observed using functional Magnetic Resonance Imaging (fMRI), exhibits a whole-brain spatial organization termed the functional connectome. The fMRI-derived connectome shows dynamic reconfigurations that are behaviorally relevant. Due to the indirect nature of fMRI, it is unclear whether such topographic changes reliably reflect modulation in neuronal connectivity patterns. Here, we directly compared concurrent fMRI-derived and electrophysiological connectivity dynamics on a connection-wise basis across the whole connectome.

Dynamic whole-brain functional connectivity (dFC) was assessed during resting-state in two independent concurrent fMRI-electroencephalography (EEG) datasets (42 subjects total) using a sliding window approach. FMRI- and EEG-derived dFC shared significant mutual information in all canonical EEG frequency bands. Notably, this was true for virtually all connections. Across all EEG frequency bands, connections with the strongest link between EEG and fMRI dynamics tied the default mode network (DMN) to the rest of the brain. Beyond this frequency-independent multimodal dFC, fMRI connectivity covaried with EEG connectivity in a frequency-specific manner in two distributed sets of connections for delta and gamma bands, respectively. These results generalized across the two datasets.

Our findings promote the DMN as a universal hub of dynamics across frequencies, but also show that spatial distribution of fMRI and EEG dFC differ across the canonical EEG-frequency bands. This study reveals a close relationship between time-varying changes in whole-brain connectivity patterns of electrophysiological and hemodynamic signals. The results support the value of EEG for studying the whole-brain connectome and provide evidence for a neuronal basis of fMRI-derived dFC.