We computed the spectral time series of the centroid voxel of each ROI and considered it representative of the spectral activity of the whole ROI. The resulting 378 pairs of intra-cerebral electrical sources were used to estimate FC in the brain (Schoffelen and Gross, 2009). We adopted the eLORETA connectivity algorithm, described in two methodological reports (Pascual-Marqui, 2007; Pascual-Marqui et al., 2011) and applied in recent EEG (Canuet et al., 2011, 2012; Lehmann et al., 2012, 2014; Pagani et al., 2012; Olbrich et al., 2014) and ERP (Mulert et al., 2011) studies. This FC method represents connectivity indices in instantaneous and lagged components. While the instantaneous (zero-lag connectivity) component in a given frequency band is sensitive to intrinsic limitations such as the effect of volume conduction and to low spatial resolution, the lagged (non-instantaneous connectivity) component has a physiological origin. The physiological lagged connectivity index, namely Lagged Phase Synchronization (LPS), measures the functional connection (a corrected lagged phase synchrony value after the instantaneous zero-lag contribution has been excluded) between two intracerebral electrical source signals in the frequency domain based on
