17846
Cerebral Morphometry in the Abide Data Set

Friday, May 16, 2014
Atrium Ballroom (Marriott Marquis Atlanta)
M. Schaer1,2, C. J. Lynch3 and V. Menon4, (1)Stanford University, Stanford, CA, (2)Office Medico-Pedagogique, University of Geneva Medical School, Geneva, Switzerland, (3)University of Georgetown, Washington, DC, (4)Psychiatry and Behavioral Sciences, Stanford University, Stanford, CA
Background:  

Numerous published studies have delineated morphometric differences in the brain of children, adolescents and adults with autism spectrum disorders (ASD). However, most of these studies analyzed limited sample size, or specific populations in terms of age range, symptoms’ severity or cognitive abilities, so that large-scale studies are needed to better understand how demographics and clinical profile influence the cerebral phenotype in ASD.

Objectives:  

This study aims at exploiting the largest sample of subjects available to date to examine the structural brain differences in patients with ASD as compared to controls. Further, we sought to explore potential structural correlates of the varying symptoms’ severity and cognitive level in autism.

Methods:  

The ABIDE data set (http://fcon_1000.projects.nitrc.org/indi/abide/) comprises 1112 structural MRI, collected from 539 patients with ASD and 573 controls aged between 6.5 and 64 years old. Volumetric estimations and 3D cortical reconstructions were obtained using FreeSurfer (http://surfer.nmr.mgh.harvard.edu). As the ABIDE dataset is distributed without any quality control, intensive inspection was achieved for each subject and manual edits were used as needed. Quality control was conducted in 852 MRIs from 13 sites to date, among which 128 MRIs (15%) were excluded because of motion, artifact or poor cortical reconstruction. The resulting 724 scans were used to compare cerebral and regional cortical and white matter volumes between controls and ASD. Further analyses were conducted within the ASD group to correlate cerebral morphometry with symptoms’ severity as measured with the ADOS (Autism Diagnostic Observation Scale, Lord et al.) and with IQ.

Results:  

In the entire sample, no difference in global brain volume was observed in ASD as compared to controls. Trends for increased cortical volume in ASD was observed in the bilateral superior temporal gyri, right precuneus and left isthmus of the cingulate (p<0.05, uncorrected for multiple comparisons). Within the ASD group, patients that had higher severity of symptoms had larger cerebral volumes (cortical, white and subcortical) than patients with lower severity of symptoms (all p<0.002). At the regional level, this increased volume in the most severely affected patients was mostly lateralized in the left hemisphere, affecting prefrontal medial and lateral regions, inferior and medial temporal areas, as well as the parieto-temporo-occipital junction. We also observed that the patients with ASD with the lower IQ had smaller cerebral and white matter volume as compared with those with higher IQ.

Conclusions:  

In this large sample of patients with ASD, we observed that ASD diagnosis alone was not a significant parameter related to different brain morphometry, suggesting that the clinical heterogeneity is also related to heterogeneous cerebral phenotype. Disentangling the different direction of the effect of higher symptom severity and lower cognitive abilities may help reconciliate previously divergent results and provide a framework to better understand the spectrum of neurodevelopmental pathways that can lead to autism.