Total Brain Volumes
As expected, there were significant TBV differences between all men (n=48) and all women (n=48), with larger volumes in men (p<0.001). shows the means and standard deviations for male and female subsamples. As expected, there were no significant differences with respect to TBV between the matched men and the matched women (there were also no significant differences with respect to GM, WM, and CSF ratios).
There were significant differences between the extreme men and the extreme women (p<0.001) with larger TBVs in men. There were also significant differences between the extreme men and the matched women (p<0.001), as well as between the matched men and the extreme women (p<0.001), with larger TBVs in men. The mean differences between men and women in the latter two comparisons were similar (i.e., 217 ml and 185 ml, respectively) and considerably smaller than when comparing male and female extremes (402 ml).
Regional Gray Matter Volumes
As illustrated in , for the comparisons of regional GM volumes between all men and all women, we revealed significant main effects of sex at p=0.05 (FDR-corrected) and restricting outcomes to clusters exceeding k=1000 voxels (i.e., the calculated spatial extent threshold according to the theory of Gaussian random fields). There were no significant effects of brain size and no significant interactions between sex and brain size.
Figure 1 Sex differences in regional GM (main effect and subsamples). Displayed are maximum intensity projections superimposed onto the SPM standard glass brain template (sagittal and coronal view). Panel 1a illustrates the main effect of sex (bidirectional). (more ...)
When conducting post hoc tests, there were no regions of larger GM volumes in men compared to women at p=0.001 (uncorrected), regardless of which subgroups were compared with each other. In contrast, we revealed a number of regions where women had larger GM volumes compared to men at p=0.001 (uncorrected). These sex effects (women > men) decreased slightly when brain size differences between men and women declined (). Importantly, we also revealed clusters of significantly larger GM volumes in women compared to men in all four comparisons when applying FDR corrections at p=0.05 and restricting outcomes to clusters exceeding k=1000 voxels.
provides the detailed T statistics associated with the specific comparison within the matched sample at p=0.05 (FDR-corrected) and k=1000. As illustrated, we revealed significantly larger GM volumes in women compared to men in the following three main clusters: Cluster 1 (k=6505) constituted of the left and right caudate (extending into other regions of the basal ganglia) as well as into the left orbito-frontal region. Cluster 2 (k=2224) comprised regions of the left superior temporal gyrus, and Cluster 3 (k=1982) encompassed regions of the left superior frontal gyrus (for MNI coordinates pertaining to cluster-specific local maxima refer to ).
Figure 2 Sex differences in regional GM (matched sample). Displayed are section views of the single subject SPM standard brain. The clusters indicate brain regions where the matched women had significantly larger GM volumes than the matched men. The color intensity (more ...)
Although there was a lack of clusters indicating a significant main effect of brain size (or interactions between brain size and sex), we conducted further exploratory analyses in order to investigate possible trends when comparing small and large brains within males only (i.e., matched vs. extreme men; mean difference: 217 ml) and also separately within females only (i.e., extreme vs. matched women; mean difference: 185 ml). For this purpose we abstained from applying corrections for multiple comparisons. However, none of these comparisons revealed any significant clusters at p=0.001, uncorrected.