used to detect the linear and nonlinear inter-dependences between two dynamical systems (Stam and van Dijk, 2002). However, some statistical measures for signal interdependencies can be spuriously biased by the volume conduction effect of source activities within the brain (Guevara et al., 2005; Nunez et al., 1997). EEG is sensitive to volume conduction because EEG potentials measured at the scalp are highly dependent on tissue conductivity of activity from one or more sources. In the present study, 29 EEG electrodes were used, which is a relatively small number compared to previous network analyses of MEG data (Buldu et al., 2011; Kitzbichler et al., 2011; Stam et al., 2009). Previous reports suggest that voltage correlations due to volume conduction would be insubstantial for electrodes separated by 4 cm or more (Doesburg et al., 2005; Nunez et al., 1999), and the effect of volume conduction may be reduced with a small number of sparsely-located channels. Further studies are needed to determine an unbiased measure to detect the true neuronal activity between brain nodes or regions. Alternative measures sensitive only to the functional coupling between channels excluding volume conduction have been suggested — e.g., phase lag index (PLI — Stam et al., 2007b;
