The effects of source leakage on these different connectivity measures is shown in Figure 4. Synthetic source time series were simulated as described in section Source space MEG. Functional connectivity between reconstructed source time courses was estimated by using either coherence (C; top left), or phase locking value (PLV; bottom left), or imaginary part of coherency (ImCoh; top center), or imaginary part of phase locking value (iPLV; bottom center), or lagged coherence (ρ2; top right), or weighted phase lag index (wPLI; bottom right). For each connectivity measure, the ratio between the estimated functional connectivity values and the true value (i.e., the one obtained directly from the generated source time sources) is plotted as a function of the distance between sources 1 and 2. Artificially inflated connectivity estimates (ratio values larger than 1) can be observed for coherence and phase locking value as an effect of source leakage. Of note, this effect is not limited to close by sources but is present also when the sources are far apart (e.g., with distance of about 6 cm). Conversely, measures which were specifically
