MEG and fMRI dynamics during movie viewing

Abstract

Movies, mimicking dynamic visual scenes experienced in our everyday life, provide a useful tool to study brain processes related to natural viewing. Despite the apparent complexity of the movie stimuli, functional magnetic resonance imaging (fMRI), magnetoencephalography (MEG) and electroencephalography (EEG) studies have shown that movies can generate consistent brain activity across viewers. However, little is known about the similarities and differences of hemodynamic and electromagnetic brain dynamics during such natural viewing. We compared MEG and fMRI dynamics in eight subjects who viewed a 15-min black-and-white movie ("At Land" by Maya Deren, 1944) twice during 3-T fMRI and twice during 306-channel MEG recordings. We calculated voxel-wise intra- and intersubject correlations within each imaging modality as well as the correlation between MEG envelopes and fMRI signals. The fMRI signals showed intra- and intersubject correlations up to r = 0.66 and 0.37, respectively, whereas correlations were weaker for the envelopes of band-pass filtered (7 frequency bands within 0.03–100 Hz) MEG signals (intrasubject correlation r < 0.14 and intersubject correlation r < 0.05). Both for fMRI and MEG, the strongest intra- and intersubject correlations took place in occipital areas. Although direct MEG–fMRI voxel-wise correlations were unreliable, MCCA-based spatial-filtering of the MEG data uncovered signal components with intersubject correlations up to r = 0.25. Furthermore, the envelopes of these MEG components below 11 Hz showed statistically significant association with fMRI signals in a general linear model (GLM). Similarities between envelopes of MEG signals and fMRI voxel time-courses were seen mostly in occipital, but also in temporal and frontal brain regions. The results show that the most consistent brain activity across movie viewers occurred in early visual areas, but the association between hemodynamic and electromagnetic activity extended also to higher-order brain areas.

Meeting abstract presented at VSS 2018