Atypical Regional Brain Volume in Women but Not Men with Autism Overlaps with Sexually Dimorphic Regions: Neuroanatomical Evidence for Females As a Sub-Group on the Autism Spectrum

Background: Females with autism spectrum conditions (ASC) have been relatively ignored in research, or implicitly assumed as similar to their male counterparts. This leaves two questions unanswered: In what ways are females and males with ASC similar or different? What mechanisms account for such similarities and differences? Recent studies suggest females with ASC are distinct from males with ASC in terms of cognitive, serum biomarker and early brain overgrowth profiles. This suggests that sex-linked factors may contribute substantially to the observed heterogeneity and potentially obscure the scope of inferences that can be drawn from etiological studies of ASC.

Objectives: We aimed (1) to characterize brain morphometric features for women with autism; (2) to investigate how women and men with ASC are similar or different in brain morphometry; (3) to test if typical sexual dimorphism is linked to autism; and (4) to investigate the extent to which a proxy marker of early androgen exposure (the 2D:4D ratio) is associated with the link between brain sexual dimorphism and autism.

Methods: High-resolution structural MRI (3T) images were obtained and preprocessed with a standard voxel-based morphometry (VBM) pipeline (DARTEL) in SPM8. Thirty women with ASC (aged 18-49 years) were compared with 30 age and IQ-matched typical control women to characterize the brain morphometry in women with autism. They were then compared with age and IQ-matched men with and without ASC (N = 30 per group) in a 4-group design, using two statistical strategies: (1) a factorial ANOVA (to test if women and men with autism have different brain morphometry) and (2) spatial overlap (conjunction) analyses on planned pair-wise comparisons (to test if brain feature of autism is linked to typical sexual dimorphism, in males and females, respectively).

Results: Compared to typical women, women with ASC have smaller relative gray matter (GM) volume in the anterior cingulate cortex, smaller relative white matter (WM) volume of ponto-cerebellar fibers and larger WM volume bilaterally in temporo-parieto-occipital regions. In particular, women with ASC showed substantial ‘masculinization’, evidenced by the finding that up to 25.14% of GM voxels and 55.34% of WM voxels showing a diagnostic effect were also sexually dimorphic voxels. In contrast, virtually no overlap was observed between diagnostic effect and sexual dimorphism in men. Finally in women, for GM regions involved in both features of autism and sexual dimorphism, a proxy measure of prenatal androgen stimulation effect (2D:4D ratio) correlated with regional volume in typical women but not women with ASC.

Conclusions: Brain structural characteristics are strikingly distinct between men and women with ASC, and the effect of autism overlaps substantially with the effect of sexual dimorphism in women with ASC. Future research should thus avoid combining males and females with ASC in studies as this increases heterogeneity. The underlying mechanisms that contribute to variation in brain structure are likely not the same in males and females with ASC. Predictions from the ‘extreme male brain’ theory may be more evident in females with ASC.