Table 1.
Sources of open access datasets used for analysis of 1011 scans.
Figure 1.
Significant lateralization of gray matter density.
Colored regions included ROIs that showed significantly greater left- or right-lateralization of gray matter density across 1011 subjects, correcting for multiple comparisons using a false discovery rate correction of q<0.05 across 7266 ROIs. Color bars show t-statistics for the left and right hemispheres, respectively. Images are in radiologic format with subject left on image right.
Figure 2.
Degree maps for significantly left- and right-lateralized connections after regression of structural laterality index from all connections.
Significantly lateralized connections (after correcting for multiple comparisons using a false discovery rate of q <0.05, across all 14.1 million intrahemispheric connections) were used to construct a graph of significantly left-lateralized connections among left hemisphere ROIs and a separate graph of significantly right-lateralized connections among right hemisphere ROIs. Color scale shows graph-theoretical degree (i.e., sum of all significantly lateralized connections in which a given node is represented) for each ROI. Images are in radiologic format with subject left on image right.
Table 2.
MNI coordinates of 20 lateralization hubs.
Figure 3.
Significantly lateralized connections to each hub.
The hemispheric lateralization maps for the nine hubs of the left-lateralized network and 11 hubs of the right-lateralized network are shown in lateral and medial projections. Color scale (t-statistic) shows significantly left-lateralized (warm colors) or right-lateralized (cool colors) to the seed (i.e., hub). A black circle marks the position for each seed.
Figure 4.
Significantly lateralized connections between each of the 20 hubs.
Warm colors show significant left lateralization and cool colors show significant right lateralization. Color bar shows t-statistic for each connection. All colored squares were significant after correcting for multiple comparisons using a false discovery rate of q<0.05 among all possible connections between the hubs. See Table 2 or Figure 3 for the hubs’ two-letter abbreviations.
Figure 5.
Significant correlation of lateralized connections across subjects.
Yellow nodes represent connections between left hubs and green nodes represent connections between right hubs. An edge is present if lateralization was found to significantly correlate across subjects between the two connections, with red edges showing positive correlation and blue edges negative correlation, after correcting for multiple comparisons using a false discovery rate of q<0.05 across all possible connection-to-connection pairs. Virtually all edges are between nodes with a hub in common. A Kamada-Kawai algorithm was implemented in Social Network Image Animator software (http://www.stanford.edu/group/sonia/). The software was also used to visualize the relationship between connections. See Table 2 or Figure 3 for the hubs’ two-letter abbreviations.
Figure 6.
Change in mean functional lateralization with age.
Mean functional lateralization index for all connections between left (A) and right (B) hubs, respectively, is shown for each subject, plotted against subject age. Pearson correlation coefficients and p-values are shown above both plots.
Table 3.
Connections between right-lateralized hubs that change in lateralization across development between the ages of 7 and 29.
Figure 7.
Reproducibility of lateralization.
A, Mean functional lateralization index for the 91 intrahemispheric connections (blue, connections involving right-lateralized hubs; red, connections involving left-lateralized hubs) is compared when averaging across all subjects except those from the Beijing site and when averaging across only subjects from the Beijing site. Pearson correlation coefficients and p-values are shown in both plots. B, Mean functional lateralization index for the 91 intrahemispheric connections (blue, connections involving right-lateralized hubs; red, connections involving left-lateralized hubs) is compared when averaging across all subjects and when averaging across all subjects except those that come from a site with less than 10 subjects.