An fMRI dataset in response to "The Grand Budapest Hotel", a socially-rich, naturalistic movie

Abstract

Naturalistic stimuli evoke strong, consistent, and information-rich patterns of brain activity, and engage large extents of the human brain. They allow researchers to compare highly similar brain responses across subjects, and to study how complex representations are encoded in brain activity. Here, we describe and share a dataset where 25 subjects watched part of the feature film "The Grand Budapest Hotel" by Wes Anderson. The movie has a large cast with many famous actors. Throughout the story, the camera shots highlight faces and expressions, which are fundamental to understand the complex narrative of the movie. This movie was chosen to sample brain activity specifically related to social interactions and face processing. This dataset provides researchers with fMRI data that can be used to explore social cognitive processes and face processing, adding to the existing neuroimaging datasets that sample brain activity with naturalistic movies.

Fig. 1

Framewise displacement for each subject across all runs. Subject motion was low in this dataset, as indicated by a median framewise displacement well below 0.5 mm for all subjects (the median value across subjects was 0.09 mm, minimum median across subjects of 0.06 mm, max 0.19 mm). Twenty out of 25 subjects had less than 5% volumes marked as motion outliers (fMRIprep defines an outlier as a volume in which framewise displacement is greater than 0.5 mm or standardized DVARS is greater than 1.5; see Methods).

Fig. 2

Temporal SNR across subjects. (a) Violin plots showing tSNR values across the brain. For each subject, a tSNR map was first generated by computing the median tSNR value across runs within each voxel. This plot shows the distribution of values in the tSNR map within a brainmask, computed in each subject’s volumetric anatomical space. Subjects are ordered in increasing median tSNR. Across subjects, the mean tSNR was 74.42 ± 3.91. (b) Median tSNR across subjects computed on data that was projected to the template surface fsaverage. As expected, areas closer to air-tissue boundaries such as the anterior temporal lobe and orbito-frontal cortex show signal dropout, while tSNR is high across the whole cortex.

Fig. 3

Inter-subject functional correlation. As expected from an audio-visual movie, visual and auditory areas showed the largest correlation in brain responses across subjects. Additional areas belonging to the theory-of-mind network, such as precuneus, temporo-parietal junction (TPJ) and medial prefrontal cortex (MPFC) also showed high correlation across subjects, as well as prefrontal areas, possibly highlighting the richness in social information available in the movie used for this dataset.

Fig. 4

Between-subject time-segment classification on hyperaligned data. The left panel (split 1) shows results obtained from hyperaligning on the first half of the data (runs 1–3), and classifying on the second half (runs 4, 5). The right panel shows the complementary analysis, that is, hyperaligning on the second half of the data, and classifying on the first half. Despite differences in absolute classification values due to differences in amount of data, the results are qualitatively similar. The highest classification values could be found in visual and auditory areas, as well as theory-of-mind areas such as precuneus, TPJ, and MPFC, and also prefrontal areas.