Skip to main page content
Access keys NCBI Homepage MyNCBI Homepage Main Content Main Navigation
. 2015 May;66:46-59.
doi: 10.1016/j.cortex.2015.02.008. Epub 2015 Mar 3.

Multimodal Neuroimaging Based Classification of Autism Spectrum Disorder Using Anatomical, Neurochemical, and White Matter Correlates

Affiliations
Free PMC article

Multimodal Neuroimaging Based Classification of Autism Spectrum Disorder Using Anatomical, Neurochemical, and White Matter Correlates

Lauren E Libero et al. Cortex. .
Free PMC article

Abstract

Neuroimaging techniques, such as fMRI, structural MRI, diffusion tensor imaging (DTI), and proton magnetic resonance spectroscopy (1H-MRS) have uncovered evidence for widespread functional and anatomical brain abnormalities in autism spectrum disorder (ASD) suggesting it to be a system-wide neural systems disorder. Nevertheless, most previous studies have focused on examining one index of neuropathology through a single neuroimaging modality, and seldom using multiple modalities to examine the same cohort of individuals. The current study aims to bring together multiple brain imaging modalities (structural MRI, DTI, and 1H-MRS) to investigate the neural architecture in the same set of individuals (19 high-functioning adults with ASD and 18 typically developing (TD) peers). Morphometry analysis revealed increased cortical thickness in ASD participants, relative to typical controls, across the left cingulate, left pars opercularis of the inferior frontal gyrus, left inferior temporal cortex, and right precuneus, and reduced cortical thickness in right cuneus and right precentral gyrus. ASD adults also had reduced fractional anisotropy (FA) and increased radial diffusivity (RD) for two clusters on the forceps minor of the corpus callosum, revealed by DTI analyses. 1H-MRS results showed a reduction in the N-acetylaspartate/Creatine ratio in dorsal anterior cingulate cortex (dACC) in ASD participants. A decision tree classification analysis across the three modalities resulted in classification accuracy of 91.9% with FA, RD, and cortical thickness as key predictors. Examining the same cohort of adults with ASD and their TD peers, this study found alterations in cortical thickness, white matter (WM) connectivity, and neurochemical concentration in ASD. These findings underscore the potential for multimodal imaging to better inform on the neural characteristics most relevant to the disorder.

Keywords: Autism; Classification; DTI; MRI; Multimodal neuroimaging; Spectroscopy.

Figures

Fig. 1
Fig. 1
(A) The input feature space for our toy example showing a scatter plot of the three variables X1, X2 and X3 for two classes (shown in red and blue); (B) The decision tree obtained for the data in the toy example shown in A; (C) A projection of the data in A onto the X1–X3 plane. The green box corresponds to the feature space for the left arm of the decision tree in Fig. 1B, i.e., X3 < .732. It can be seen all that class-1 features (blue) are correctly classified; and (D) A projection of the data in A onto the X2-X3 plane. The green box corresponds to the feature space for the right arm of the decision tree in Fig. 1B, i.e., X3 < .732 & X2 < .483 and X3 < .732 & X2 > .483. It can be seen that a class-1 feature (blue) is correctly identified for X3 < .732 & X2 < .483 while another class-1 feature (blue) is mis-classified along with class-2 features for X3 < .732 & X2 > .483.
Fig. 2
Fig. 2
(A) Group means for fractional anisotropy (FA) for the nodes along the forceps minor of the corpus callosum for the TD (depicted in blue) and ASD (depicted in green) groups. The clusters with significant reduction in FA in ASD participants (p < .05, corrected) are indicated with a star; (B) A rendering of FA measurements for the forceps minor of the corpus callosum for one subject as a visualization of the tract properties. The tract segmentation was based on the previously defined Mori white matter atlas (Hua et al., 2008; Wakana et al., 2007).
Fig. 3
Fig. 3
Decision tree for classification of autism (ASD) and typically developing (TD) groups, including the following predictors: right forceps minor radial diffusivity (RD), left Inferior Frontal Gyrus pars opercularis cortical thickness (CT), and left forceps minor fractional anisotropy (FA). Classification accuracy reached 91.9% ± .42.
Fig. 4
Fig. 4
(A) A projection of the input feature space onto a 2D feature space containing the right forceps minor RD and left forceps minor FA as the two axes. The highlighted green box indicates the region in the feature space being utilized for classification which corresponds to the right side of the decision tree in Fig. 3. Blue indicates TD control participants and Red indicates ASD participants. Mis-classified subjects are shown using arrows; (B) A projection of the input feature space onto a 2D feature space containing the right forceps minor RD and left pars opercularis CT as the two axes. The highlighted box indicates the region in the feature space being utilized for classification which corresponds to the left side of the decision tree in Fig. 3. Blue indicates TD participants and Red indicates ASD participants. Mis-classified subject is shown using an arrow.
Fig. 5
Fig. 5. Decision tree for a regression model including symptom severity scores (measured by RAADS-R) and significant factors including left forceps minor fractional anisotropy (FA), cortical thickness (CT) for left isthmus cingulate, left posterior cingulate, and right cuneus, and radial diffusivity (RD) for right forceps minor
Fig. 6
Fig. 6. Original (blue) and predicted (red) symptom severity (measured by RAADS-R) using a regression model including typically developing (TD) and autism (ASD) participants

Similar articles

See all similar articles

Cited by 36 articles

See all "Cited by" articles

Publication types

MeSH terms

Feedback