Decreased Cortical Thickness in Autism Spectrum Disorder: A Large-Sample Investigation Using Surface-Based Morphometry

Background: Cortical thickness (CT) abnormalities in autism spectrum disorder (ASD) are reported in the current research literature. However, several major limitations exist, including (a) highly mixed results with some studies reporting increased CT in ASD while others reporting decreased CT, (b) heterogeneous methods with some studies using voxel-based morphometry (VBM) leading to less accurate results (because the cerebral cortex is highly folded and curved) relative to surface-based morphometry (SBM), and (c) small sample sizes which may lead to erroneous conclusions. 

Objectives: This study addresses the aforementioned limitations by comparing CT using SBM in large samples of individuals with ASD and typically developing (TD) participants.

Methods: There were 213 total participants: 105 children and adolescents with ASD, including 88 males (mean age = 9.7 ±3.7 years) and 17 females (mean age = 9.3 ±3.8 years), and 108 TD participants (mean age = 10.37 ± 3.4 years) matched on age, sex, and cognitive functioning. Cognitive functioning was measured by Differential Ability Scales-II (DAS-II). The diagnosis of autism was confirmed with Autism Diagnostic Observation Schedule (ADOS) and Autism Diagnostic Inventory-Revised (ADI-R). T1-weighted, whole-brain structural MRI scans were acquired using a 3-Tesla scanner. The structural MRI scans were processed and analyzed using the FreeSurfer image analysis suite version 5.1.0, which consists of automated tools for reconstruction of the brain from high-resolution MRI data. With some manual correction, the program facilitated accurate and precise CT measurements of 66 different brain regions based on the Desikan-Killiany atlas. These measurements were entered into the Statistical Package for the Social Sciences (SPSS) version 19, and between-group comparisons of CT were conducted using multivariate analysis of covariance (MANCOVA).

Results: MANCOVA controlling for age, cognitive functioning, and intracranial volume revealed that children and adolescents with ASD had consistently lower CT than TD across several regions known to be involved in processing social information.  These regions included the bilateral banks of the superior temporal sulcus (STS) as well as the left superior temporal, middle temporal, entorhinal, and fusiform gyri, and finally the left pars orbitalis and supramarginal gyrus.

Conclusions: By using SBM and a large sample of participants, this study documents specific areas of cortical thinning in ASD confined to regions involved in social information processing. These abnormalities likely have important implications for the social-communication deficits in ASD.