Sex Related Brain Structure of Social Cognition: An Autism Twin Study

Background:

A normative female advantage in social cognition skills has been proposed to contribute to the reduced risk of females to develop autism spectrum disorder (ASD) (Baron-Cohen et al., 2005; Christov-Moore et al., 2014). However, it is unclear if differences in brain structure related to social cognition exist between the sexes, and if such potential differences are associated with autistic trait severity. Further, studies on social cognition in ASD have often used tasks with low ecological validity, that are not challenging enough for higher functioning subjects (Dziobek et al., 2006).

Objectives:

The present study therefore aimed to investigate the relationship between brain structure and social cognition skills, assessed with a more complex, ecologically valid, social cognition task, as well as autistic traits, in males and females along the continuum of ASD. By using a co-twin control design, we were able to control for familial confounding factors (e.g. age, socio-economic status, etc.).

Methods: From the Roots of Autism and ADHD Twin Study Sweden (RATSS) 57 twin pairs (24 female and 33 male pairs), mean age 17.57 (range 12.52 - 23.69), completed the movie for the assessment of social cognition (MASC). In addition, autistic traits related to social cognition difficulties were assessed with a subscale from the parent-report Social Responsiveness Scale-2 (SRS-2). Twin pairs were either discordant or concordant for ASD, discordant or concordant for other neurodevelopmental disorders, or concordant for neurotypical development. Surface based morphological estimates including cortical surface, thickness and volume were computed from T1 anatomical images. Across and within-pair associations were calculated between social cognition and anatomical structure in the cerebral social network.

Results:

Females displayed better social cognition skills (MASC) compared to males, but there was no sex difference in the amount of autistic traits associated with social cognition difficulties (SRS-2). After controlling for shared factors within twin pairs, overall autistic traits predicted reduced social cognition skills only in males. In line with this, only in males were decreased social cognition skills (MASC) associated with increased thickness of the left inferior frontal, insula, mid temporal and right fusiform gyri. These associations between social cognition skills and brain structure were largely replicated in a subsample consisting of ASD discordant pairs (6 female and 8 male pairs). In contrast, autistic traits (related to social cognition difficulties (SRS-2)) were associated with reduced volume and surface area of most regions of interest only in females.

Conclusions: When controlling for familial factors, variation in social cognition skills was associated with brain structure mostly in males, while autistic traits related to social cognition difficulties were associated with brain structure only in females. Our findings therefore point at potential differences between males and females in autism phenotype and associated underlying neural structure. Finally, the results highlight the power of using a co-twin control design to detect subtle associations between brain structure and behavior, and suggest the importance of non-shared environmental factors in the relationship between neuroanatomy and social cognition skills.