An MRI Investigation of Neuroanatomical Differences in High Functioning Adults with Autism Spectrum Disorder Using Non-Parametric Cluster Based Statistics

Friday, May 16, 2014
Atrium Ballroom (Marriott Marquis Atlanta)
D. S. Andrews1, E. Daly1, J. Horder1, M. A. Mendez2, V. Giampietro3, M. Brammer3, C. E. Wilson1, N. Gillan1, C. Ecker1 and D. G. Murphy1,4, (1)Department of Forensic and Neurodevelopmental Sciences, Institute of Psychiatry, King's College London, London, United Kingdom, (2)Forensic and Neurodevelopmental Sciences, Institute of Psychiatry, King's College London, London, United Kingdom, (3)Centre for Neuroimaging Sciences, King's College London, London, United Kingdom, (4)The Sackler Institute for Translational Neurodevelopment, Institute of Psychiatry, King’s College London, London, United Kingdom
Background: Voxel-based morphometry (VBM) is a statistical method for identifying regional differences in grey (GM) and white (WM) volumes. VBM investigations of autism spectrum disorder (ASD) have often been confounded by grouping individuals across different bands of the spectrum, periods of development, and/or comparison to inappropriately matched controls with regards to age and intellectual ability. Further more these studies commonly rely on parametric statistical methods and potentially overly stringent corrections for multiple comparisons. These methodological issues all potentially serve to conceal elements of the larger pattern of neuroanatomical differences observed in ASD.

Objectives: This study sought to identify regional differences in GM and WM volumes between a homogenous group of ASD participants compared to matched typically developing controls using non-parametric cluster based VBM. Correlations between GM volumes and number of observed ASD symptoms were also investigated.  

Methods: Group comparisons of GM and WM volumes were conducted between 21 adult males with ASD (Age: 32 ± 8.87, IQ: 116 ± 13.88) and 21 typically developing controls (Age: 32 ± 9.76, IQ: 119 ± 8.42) matched for age and general intelligence. All ASD participants were diagnosed according to the International Statistical Classification of Diseases, 10th Revision (ICD-10) and were categorized as high functioning (IQ>80). Diagnosis was confirmed using the Autism Diagnostic Interview-Revised (ADI-R). Current symptoms were assessed using the Autism Diagnostic Observation Scale (ADOS). High-resolution T1 anatomical magnetic resonance images (MRI) were collected and segmented using VBM methods with Diffeomorphic Anatomical Registration using Exponentiated Lie algebra (DARTEL). Non-parametric cluster level statistical analysis using X-Brain Activation Mapping (XBAM v4.1) was performed to identify regional differences in GM and WM volumes, as well as correlations between GM volumes and ADI-R and ADOS symptom scores.  

Results: Regional differences in the neuroanatomy of ASD participants compared to controls were revealed in the form of spatially distributed significant clusters (permutation test P=.001) of GM and WM volumetric increases. GM increases encompassed several regions including the orbitofrontal and dorsolateral prefrontal cortices, superior temporal gryus, posterior parietal cortex, primary visual areas, fusiform gyrus and cerebellum. Spatially distributed WM abnormalities were identified across the brain including tracts that connect associative areas, namely the longitudinal facsiculus and corpus callosum. GM regions of the prefrontal cortex were also found to significantly correlate (permutation test P=.002) with ADI-R social scores.  

Conclusions: Previous VBM studies in ASD have reported several contradictory findings. Cluster based non-parametric VBM methods allow for more sensitive identification of GM and WM volume differences in ASD. Furthermore, the recruitment of a homogenous group of ASD participants and appropriately matched controls serves to limit potential confounds. The current study identified significant distributed GM and WM volume increases across several brain regions known to mediate behaviors associated with ASD. By measuring tissue volume alone, VBM methods are inherently limited. Techniques such as surface-based morphometry (SBM) and diffusion tensor imaging (DTI) can aid in better defining the neuro-abnormalities present in ASD and provide valuable clues as to the underlying developmental processes potentially responsible for observed deficits in the disorder.