Appendix 1—figure 11. Proportion of vertices of the 30 most predictive regions in 12 cortical networks defined in Ji et al., 2019.

(A) The 30 regions that are most predictive of general intelligence based on hyperaligned fine-grained task functional magnetic resonance imaging (fMRI) connectivity are in the default (34.0% of vertices in the 30 regions), frontoparietal (33.0%), cingulo-opercular (20.2%), dorsal attention (5.7%), and visual 2 (6.7%) networks. (B) The 30 regions that are most predictive of general intelligence based on hyperaligned fine-grained resting fMRI connectivity are in the default (39.3%), frontoparietal (22.5%), cingulo-opercular (14.4%), dorsal attention (13.9%), visual 2 (4.9%), language (2.8%), and visual 1 (1.7%) networks. The 12 cortical networks are defined based on Ji et al., 2019. Note that different cortical network parcellations are in agreement with each other in general, and the proportion of cortical vertices in these 12 networks is similar to the proportion in the seven cortical systems based on Yeo et al., 2011. Most of the vertices are in default, frontoparietal, and dorsal attention networks based on both Yeo et al., 2011 and Ji et al., 2019; some regions in the ventral attention network based on Yeo et al., 2011 are labeled as cingulo-opercular in Ji et al., 2019.