White Matter Abnormalities Support the Female Protective Effect in Autism: A Group-Matched, Quality-Controlled, Diffusion Tensor Imaging Study

Poster Presentation
Friday, May 11, 2018: 11:30 AM-1:30 PM
Hall Grote Zaal (de Doelen ICC Rotterdam)
E. K. Lecarie1, J. Lei2, S. C. Boland1, D. Yang3, P. E. Ventola4, D. G. Sukhodolsky4, K. Pelphrey5 and R. J. Jou6, (1)Yale Child Study Center, New Haven, CT, (2)Centre for Applied Autism Research, University of Bath, Bath, United Kingdom, (3)The George Washington University, Ashburn, VA, (4)Yale Child Study Center, Yale University School of Medicine, New Haven, CT, (5)The George Washington University, Washington, DC, (6)Yale University, New Haven, CT
Background: Many earlier diffusion tensor imaging (DTI) studies reported widespread and heterogeneous white matter (WM) abnormalities in autism spectrum disorder (ASD). However, these findings were not replicated in more recent studies which stringently controlled for MRI data quality. Moreover, there is growing awareness that ASD manifests differently in girls, specifically the Female Protective Effect, which posits that girls require a greater etiologic load to manifest significant symptoms of autism. Therefore, neuroimaging studies should also account for gender differences, as girls with ASD may be more severely affected and exhibit a neurophenotype unlike that of boys with ASD.

Objectives: To evaluate gender differences in WM between individuals with ASD and neurotypical (NT) peers, while stringently controlling for data quality and matching for confounding variables known to produce artefactual findings.

Methods: A total of 130 children were divided into four groups: 25 ASD girls (age = 9.30 years ±4.25), 65 ASD boys (age = 9.79 years ±4.16), 16 NT girls (age = 11.30 ±4.14) and 24 NT boys (age = 11.16 years ±4.52). Participants underwent clinical assessment and DTI on a 3T MRI scanner. All scans were pre-processed using FSL, including eddy current correction and estimation of the diffusion tensor, which enables calculation of fractional anisotropy (FA), an accepted measure of WM integrity. A multi-step procedure was used to ensure stringent quality control, which consisted of visual inspection by two independent raters for removing compromised data (intraclass correlation coefficient >0.90). Voxel-wise group comparisons were performed using Tract-Based Spatial Statistics. Areas of significant difference were identified (p < 0.05, corrected) using Threshold-free Cluster Enhancement. A tractography atlas was used to characterize fiber tract abnormalities.

Results: No main effects of group or gender were observed for age or full-scale IQ. For the Social Responsiveness Scale, there was no main effect of gender, though a main effect of group was noted. No main effect of group was observed for any motion parameters (displacement, rotation, and translation). There were widespread regions in both hemispheres where FA was significantly reduced in ASD girls. All fiber types were represented, including association, commissure, and projection fibers. Notably, there were no significant differences in FA between ASD and NT boys.

Conclusions: The presence of neither localized nor widespread WM abnormalities in ASD boys was supported in the current study. Combined with comparable SRS scores between genders, the widespread WM abnormalities detected only in ASD girls suggests that comparable symptoms of autism require substantial structural aberrations. These results support the Female Protective Effect and the continued importance of gender differences in autism.