Neuroanatomical Signatures of Autism

Friday, May 16, 2014
Atrium Ballroom (Marriott Marquis Atlanta)
D. Yang, R. J. Jou and K. A. Pelphrey, Child Study Center, Yale University, New Haven, CT
Background: In an fMRI study of children with autism spectrum disorder (ASD), unaffected siblings of children with ASD (US), and typically developing children (TD), Kaiser and colleagues (2010) identified three different kinds of biomarkers: (a) state regions: functionally weaker in ASD (relative to TD and US), reflecting the state of having ASD; (b) trait regions: functionally weaker in both ASD and US (relative to TD), reflecting the underlying genetic vulnerability for ASD; and (c) compensatory regions: functionally stronger in US (relative to ASD and TD), suggesting a pathway by which US might avoid developing ASD. It remains unclear how these neurofunctional biomarkers might map onto neuroanatomical characteristics in ASD.

Objectives: We aimed at elucidating the neuroanatomical biomarkers of autism by comparing and contrasting the cortical thickness among ASD, TD, and US in a large sample of children with groups rigorously matched on sex, age, and IQ.

Methods: Study participants included 209 children (age: 4-12 years old), including 100 children with ASD (25 females and 75 males), 70 TD children (28 females and 42 males), and 39 US children (17 females and 22 males). Expert clinical diagnosis of ASD was confirmed via the ADOS and ADI-R. For every participant, a high-resolution, T1-weighted MPRAGE structural scan was acquired using a 3-Tesla Siemens Tim Trio scanner. Cortical reconstruction was performed using FreeSurfer 5.1.0. We sought to identify (a) state regions, where the cortical thickness in ASD was thinner than both TD and US; (b) trait regions, where cortical thickness in both ASD and US children is thinner than TD children; and (c) compensatory regions, where cortical thickness in US is thicker than both children with ASD and TD children. To correct for multiple comparisons, we employed a Monte Carlo simulation (cluster analysis) with an a priori threshold of p< .05 for whole-brain correction.

Results: Trait regions were not apparent following whole-brain correction. State and compensatory regions were localized mainly to the temporal and frontal lobes as well as the insular cortex. Among the state regions, the cortical thickness of the right posterior superior temporal sulcus (pSTS) and that of the right lateral orbitofrontal cortex were highly negatively correlated with ADOS module 3 social affect scores (ps< .01, corrected), suggesting that cortical thickness in these regions is a sensitive biomarker of social-communication dysfunction in ASD.

Conclusions: We generally replicated the prior functional MRI findings using a large structural MRI dataset while identifying additional state and compensatory regions, providing convergent evidence for the neural signatures of ASD originally identified by Kaiser et al. (2010). Future studies are warranted to further characterize these neurobiological signatures using other imaging modalities, including diffusion MRI. Full characterization of the neural signatures of ASD is essential to guide future work aimed at elucidating the neural-systems-level pathophysiology of ASD and developing guidance for treatment.