assumption that the synaptic currents giving rise to the summated MEG signal observed on the scalp are confined to the cortical ribbon (Dale et al., 2000). It applies a real head model obtained by reconstructing each person’s cortical surface from high-resolution MRI to constrain a distributed minimum norm inverse solution (Dale and Sereno, 1993) and modeling studies indicate a reasonable localization accuracy given the applied assumptions (Liu et al., 2002). The TF analysis of the oscillatory signal developed in the context of the aMEG approach allowed us to obtain spatio-temporal estimates of the event-related theta changes associated with semantic retrieval and alcohol effects. In the present study, SW elicited stronger theta power compared to PW in the left LT and prefrontal cortices in the ∼370–520 ms time window. Both the temporal window and the spatial distribution of these effects correspond to the left-lateralized fronto-temporal network subserving the generation of the N400m analyzed in time-domain (Helenius et al., 1998; Dhond et al., 2001, 2005; Halgren et al., 2002; Marinkovic, 2004). These spatial estimates are also consistent with a large body of research based on the BOLD–fMRI method showing that the main foci of language-related activity are the left-lateralized temporal and prefrontal
