Describing the brain in autism in five dimensions--magnetic resonance imaging-assisted diagnosis of autism spectrum disorder using a multiparameter classification approach

Abstract

Autism spectrum disorder (ASD) is a neurodevelopmental condition with multiple causes, comorbid conditions, and a wide range in the type and severity of symptoms expressed by different individuals. This makes the neuroanatomy of autism inherently difficult to describe. Here, we demonstrate how a multiparameter classification approach can be used to characterize the complex and subtle structural pattern of gray matter anatomy implicated in adults with ASD, and to reveal spatially distributed patterns of discriminating regions for a variety of parameters describing brain anatomy. A set of five morphological parameters including volumetric and geometric features at each spatial location on the cortical surface was used to discriminate between people with ASD and controls using a support vector machine (SVM) analytic approach, and to find a spatially distributed pattern of regions with maximal classification weights. On the basis of these patterns, SVM was able to identify individuals with ASD at a sensitivity and specificity of up to 90% and 80%, respectively. However, the ability of individual cortical features to discriminate between groups was highly variable, and the discriminating patterns of regions varied across parameters. The classification was specific to ASD rather than neurodevelopmental conditions in general (e.g., attention deficit hyperactivity disorder). Our results confirm the hypothesis that the neuroanatomy of autism is truly multidimensional, and affects multiple and most likely independent cortical features. The spatial patterns detected using SVM may help further exploration of the specific genetic and neuropathological underpinnings of ASD, and provide new insights into the most likely multifactorial etiology of the condition.

Figure 1.

Summary of the five morphometric parameters measured at each cerebral vertex. These included average convexity (A), cortical thickness (B), pial area (C), metric distortion (Jacobian) (D), and mean (radial) curvature (E).

Figure 2.

A, B, ROC graphs for the six discrete classifiers in the left hemisphere (A) and in the right hemisphere (B). Individual points on the graph depict classifiers on the basis of all parameters (A), cortical thickness (B), metric distortion/Jacobian (C), average convexity (D), pial area (E), and mean (radial) curvature (F). C, D, The classification plots for the left (C) and right (D) hemispheres.

Figure 3.

Classification plots showing group allocation of individuals with ADHD (red squares) in the left (A) and right (B) hemispheres using the ASD classifier.

Figure 4.

Discrimination maps for the five different morphometric features in the left (L) and right (R) hemispheres. Color maps represent the weight vector on the basis of the five modality classification for cortical thickness (A), average convexity (B), metric distortion (C), and pial area (D). Weights for the mean (radial) curvature did not exceed the set threshold. Positive weights (i.e., overall excess patters in ASD relative to controls) are displayed in red, and negative weights (i.e., overall deficit patterns in ASD relative to controls) are displayed in blue.

Figure 5.

Visualization of the morphometric abnormalities in the right intraparietal sulcus. Color maps represent the weight vector score (A). B, Outlines of the cortical surface for ASD (red) and control (blue) group. This main discriminating factor in this group was an increase in sulcal depth in ASDs relative to controls. Differences in sulcal depth for this ROI are shown for both groups in C. D, Morphometric profile for this region. Profiles were derived by averaging the weight vector scores across vertices within this region of interest, and for the different morphometric parameters. Weights were identified on the basis of the concatenated SVM model, thus showing the relative contribution of parameters in this ROI.

Figure 6.

A, Visualization of the morphometric abnormalities in the left inferior parietal lobe (BA39). B, Outlines of the cortical surface for ASD and control group. Differences in metric distortion for this ROI are shown for both groups in C. D, Morphometric profile (see Fig. 4 legend).

Figure 7.

A, Morphometric abnormalities in the middle temporal sulcus. B, Visualization of cortical thickness measures for ASD (red straight line) and control (blue straight line) group. In this region cortical thickness exclusively discriminated between groups with individuals with ASD displaying increased thickness relative to controls (C, D).

Figure 8.

A, Morphometric abnormalities in the posterior cingulate gyrus. B, C, Here, a combination of cortical thickness and folding pattern led to a high contribution to the classification in that region.