Previous studies with small samples have reported that individuals with autism spectrum disorders (ASD) exhibit numerous anatomical brain abnormalities when compared with controls. The recent release of the Autism Brain Imaging Data Exchange (ABIDE, >1000 participants, ages 6-65) offers an important opportunity to perform more definitive large-scale comparisons between ASD and control individuals who span a wide range of IQ scores, ages, and symptom severities.Anatomical Magnetic Resonance Imaging (MRI) scans of 348 male ASD participants and an equal number of age, gender, and IQ matched controls were selected for the current study. We performed comprehensive univariate and multivariate analyses using multiple volumetric, cortical thickness, and cortical surface area measures derived for each of 150+ brain areas. Significant findings were almost entirely confined to comparisons of ASD individuals with relatively severe symptoms (Autism Diagnostic Observation Schedule, ADOS > 14) and controls. Even significant differences between groups, however, were on the order of 2-4%, while between-subject differences within each group reached 80-100% for most measures. Multivariate classification analyses yielded marginal decoding accuracies of ASD and control individuals, which significantly exceeded chance levels only when classifying ASD individuals with ADOS >14 and controls. We suggest that previous reports of anatomical differences across ASD and control groups and high decoding accuracies of ASD and control individuals may have resulted from the substantial anatomical heterogeneity across individuals and the recruitment of small samples. These sobering findings suggest that measures of brain volumes, cortical thickness, and cortical surface areas are likely to be of low clinical and scientific significance for understanding the pathophysiology of children, adolescents, and adults with ASD.
|Original language||English GB|
|Title of host publication||JOURNAL OF MOLECULAR NEUROSCIENCE|
|State||Published - 2014|