Functional connectivity has been shown to change over short time scales of seconds to minutes, giving rise to the so-called dynamic functional connectivity (dFC). However, the electrophysiological underpinnings of dFC states remain unclear. We investigate EEG spectral correlates of dFC states using simultaneous EEG-fMRI data, by using a high temporal resolution fMRI acquisition combined with a phase coherence approach for dFC estimation and by computing k-means clustering with a varying number of dFC states. We found an association between high alpha power topographies and specific dFC states, which included regions of the frontoparietal network and the default mode network.