, 2000, Fransson, 2006, Greicius et al., 2008, Greicius and Menon, 2004, Larson-Prior et al., 2009, Morgan and Price, 2004 and Vincent et al., 2007). It is also supported by the observation that intrinsic BOLD fMRI fluctuations account for variability in task-evoked activity ( Fox et al., 2006) and

associated behavioral performance ( Fox et al., 2007). In the context of MEG BLP-correlation, this second hypothesis predicts maintenance of within/between see more RSN topography during natural vision, and an increase of interaction between RSN as they go from a state of relative segregation at rest to a state of greater integration during task. We focus on networks (visual, dorsal attention, auditory, language) that have been modulated in fMRI during the

HKI-272 in vivo observation of natural scenes (Golland et al., 2007, Hasson et al., 2004 and Nir et al., 2006), and the default mode network (Raichle et al., 2001 and Shulman et al., 1997) that is active at rest but suppressed during task performance. We recorded MEG signals in a group of twelve participants, each performing three different experimental blocks (runs) both during visual fixation (fixation) and the observation of three movie segments (about 5 min each) taken from the Italian version of The Good, the Bad and the Ugly (movie). In a separate recording session, each participant underwent fMRI during fixation and movie conditions ( Figure 1). The MEG data analysis pipeline is the same as in de Pasquale et al., 2010 and de Pasquale

et al., 2012) and Mantini et al. (2011) and returns estimates of band limited power (BLP) from the source-space signals for delta (δ), theta (θ), alpha (α), beta (β), gamma (γ) band (see Supplemental PAK6 Information and Figure S1 available online). To evaluate the modulation produced by movie watching on BLP interaction with respect to fixation, in sensory and attention networks, we computed the total interdependence function from BLP, a global measure of interaction at different frequencies, obtained from all the possible pairs of the principal nodes of each RSN (visual, auditory, and dorsal attention networks) (Experimental Procedures and Supplemental Information). The nodes of each RSN were defined a priori from an independent set based on meta-analyses of task fMRI studies (Baldassarre et al., 2012, He et al., 2007 and Lewis et al., 2009; Table S1). In both α and β bands, the within-network inter-nodal BLP interaction was stronger at lower (<2 Hz) than higher frequencies, with a moderate peak at about 0.1 Hz during fixation in agreement with previous MEG studies (Brookes et al., 2011b, de Pasquale et al., 2010, de Pasquale et al., 2012 and Hipp et al., 2012; Figure 2A). Movie watching decreased the total interdependence compared to fixation, at frequencies below 0.3 Hz in each RSN (Figure 2A, dotted lines; Supplemental Information).