Fig. 8: Experimental design and stimulus structure.

Subjects viewed a 33-min movie composed of alternating baseline and movie blocks. Baseline periods (totaling 5:50 min) consisted of a fixation point on a gray background with no audio. Movie segments included clips from two nature documentaries, featuring a variety of species and environments. The film contained naturalistic visual and auditory stimuli, including conspecific and heterospecific vocalizations, human speech, music, and environmental sounds. At the top, the 7 Tesla and 9.4 Tesla scanners used for fMRI sessions in humans (n = 19) and common marmosets (n = 8, with 5 functional runs per marmoset) are shown. On the left and right sides of the figure is represented the statistical approach used for both marmosets (left) and human (right). After preprocessing, individual fMRI maps were analyzed using tICA, extracting 20 independent components per species. The timecourses of non-noise components (n = 12 for marmosets and n = 14 for humans) were then correlated within and across species, generating correlation matrices (example shown at the bottom). Additionally, these functional components were clustered using hierarchical clustering to further explore functional network relationships in both species.