Autism spectrum disorder (ASD) and schizophrenia are neurodevelopmental disorders often characterized by contrasting cognitive styles. ASD is associated with over-systemizing and under-empathizing, while schizophrenia exhibits hyper-mentalizing, potentially mirroring average sex differences. Previous research suggested ASD might represent an extreme form of male brain characteristics, while schizophrenia the opposite. Both disorders, however, share commonalities such as impaired executive functions and social cognition deficits. While studies indicate shared genetic contributions, the genetic correlation between ASD and schizophrenia is relatively low. Left-right brain asymmetry is a fundamental organizational principle of the brain and shows sex differences and alterations in both ASD and schizophrenia. The study aimed to investigate the association between polygenic risk scores for ASD and schizophrenia and brain structural asymmetry, specifically exploring whether these associations reflect masculinization or feminization.

Existing literature presents contrasting views on the relationship between ASD and schizophrenia, with some framing them as opposing disorders reflecting exaggerated male versus female brain characteristics, respectively. Others highlight shared deficits in executive functions and social cognition. Studies have shown shared genetic components, with common single nucleotide polymorphisms (SNPs) accounting for a moderate genetic correlation. Brain asymmetry, a key organizational principle, displays sex differences and alterations in both conditions, suggesting a potential link between genetics, brain structure, and diagnosis. This research builds upon prior work demonstrating heritability in brain asymmetry measures, suggesting a genetic basis for variations in this trait, which is partially influenced by prenatal testosterone and likely established prenatally through genetic regulation. Previous research also identified a genetic overlap between brain asymmetries and these disorders.

The study used data from 32,256 participants (15,288 males, 16,968 females) in the UK Biobank. Imputed genotype data was used to calculate polygenic risk scores (PRS) for ASD and schizophrenia using the PRS-CS software and summary statistics from previous genome-wide association studies (GWAS). Brain imaging data from T1-weighted MRI scans provided measures of regional cortical surface area, cortical thickness, and subcortical volumes. Asymmetry indexes were calculated as (Left - Right)/((Left + Right)/2) for each region. Canonical correlation analysis was employed to investigate the multivariate associations of ASD and schizophrenia PRS with the 42 regional asymmetry indexes that showed significant SNP-based heritabilities. The correlations between loadings across these asymmetry indexes were calculated and the significance was assessed using permutations to account for inter-correlations between brain regions. Functional annotation of brain regions with the strongest associations was performed using Neurosynth, a database for large-scale synthesis of functional MRI data. Associations of polygenic risks with sex were tested using t-tests and linear regression. Logistic regression examined associations with handedness. Phenome-wide association analysis (PHESANT) explored relationships with other phenotypes in the UK Biobank. The influence of brain size on the associations between polygenic risks and brain asymmetries was also examined.

A significant, albeit weak, positive correlation (r = 0.08, p = 7.13 × 10⁻⁵⁰) was found between ASD and schizophrenia PRS. Both ASD (r = 0.03, p = 2.17 × 10⁻⁹) and schizophrenia (r = 0.04, p = 2.61 × 10⁻¹¹) PRS showed significant multivariate associations with brain asymmetry. However, the multivariate patterns of association were largely distinct for the two disorders. No significant correlation existed between the loading patterns for ASD and schizophrenia PRS. Neurosynth analysis of regions with the strongest loadings revealed associations with language and executive functions for both disorders. Females showed higher polygenic risks for both ASD and schizophrenia compared to males. There was no evidence that either disorder's polygenic risk was associated with a more male-like or female-like pattern of brain asymmetry. Non-right-handedness was positively associated with ASD PRS, particularly mixed-handedness. Phenome-wide association analysis revealed several significant associations: for ASD, mainly related to hearing difficulties, socioeconomic status and long-standing illness; for schizophrenia, mainly related to poorer cognitive performance across various domains. Schizophrenia PRS showed a negative association with brain size, but this didn't affect the primary findings.

The findings demonstrate that while ASD and schizophrenia PRS are weakly correlated, their associations with brain asymmetry are largely distinct, not supporting the concept of opposing disorders along a single neurobiological dimension. Instead, the results suggest largely distinct, but overlapping, etiological mechanisms. The absence of an association with a more male-like or female-like pattern of asymmetry challenges the "extreme male brain" theory for ASD. The consistent involvement of language and executive function regions highlights potentially shared etiological mechanisms in these domains, particularly implicating the pars opercularis. The differing phenome-wide associations between ASD and schizophrenia PRS further emphasizes the distinct nature of these disorders despite some shared features. The small effect sizes necessitate caution in interpreting findings.

This study reveals distinct patterns of brain asymmetry associated with ASD and schizophrenia polygenic risks. These findings implicate language and executive functions, challenging the "extreme male brain" theory. The relatively low genetic correlation and distinct neurobiological profiles underscore the largely independent etiologies, with notable overlap in language-related regions. Future longitudinal studies are needed to clarify causal relationships.

The study used cross-sectional data, limiting inferences about causality. The relatively small effect sizes, while statistically significant due to the large sample size, may limit generalizability. The reliance on macro-structural brain measures may have missed more subtle asymmetries. The under-representation of males with high polygenic risk could have influenced the results. Finally, the lack of direct validation of polygenic risk scores against clinical diagnoses limits the definitive interpretation of the findings.