White Matter Connectivity and Autism Clinical Symptomatology: a Diffusion MRI Study

Background:

Anatomo-functional abnormalities in the regions of the ‘social brain’ (superior temporal sulcus, orbito-frontal cortex, amygdala and fusiform gyrus) have previously been described in autism spectrum disorders (ASD) . Besides, several studies have reported reduced fronto-temporal functional connectivity among this network . More recently, diffusion neuroimaging techniques have also pointed out ASD-associated white matter abnormalities, suggesting anatomical connectivity disruptions in this disorder .

Objectives:

In our study we tested the hypothesis of a relationship between white matter integrity and two clinical characteristics of children with ASD: autism severity and verbal communication deficit.

Methods:

White matter integrity was voxel-wise assessed over the whole brain using high-angular resolution diffusion tensor imaging (DTI) in 24 children with ASD (age = 8.4 ± 3.5 years; IQ = 65.2 ± 24.6; mean ± SD). The ASD diagnosis was based on DSM IV-R and ADI-R criteria. Autism clinical severity was assessed with ADI-R total score. High angular resolution diffusion images (HARDI) were acquired on a GE-Signa 1.5 T using an echoplanar sequence (41 directions, TE=70 ms; TR=9000 ms; 2*1.8*1.8 mm3; b=1500 s/mm2). Fractional anisotropy (FA) & apparent diffusion coefficient (ADC) maps were estimated with Brainvisa software (http://brainvisa.info) and then non linearly spatially normalized on study-specific FA template  and spatially smoothed (8 mm) using SPM5 software (http://www.fil.ion.ucl.ac.uk/spm/).

Two statistical analyzes of FA and ADC maps were performed: 1) correlation analysis with autism severity (total ADI-R score) and 2) comparison between ASD children with language vs. ASD children without language. Analyses, controlled for age and IQ, were performed on the whole brain using a voxelwise threshold at p < 0.005.

Results:

We found a significant correlation between ASD severity and white matter microstructure in the main pathways of the social brain network (arcuate and uncinate fasciculi): the more severe the autistic syndrome, the more FA reduction and the more ADC increase in these regions. We also found white matter microstructure differences between ASD children with and without language in white matter regions of the language network: temporal pole, superior and middle temporal regions, inferior frontal region (temporal and frontal parts of the arcuate fasciculus). In these regions, the FA was higher and the ADC was lower in the ASD children with verbal communication.

Conclusions:

In this first high resolution whole-brain analysis of diffusion indices in children with ASD, we found a correlation between autism severity and white matter disruption in fronto-temporal pathways. We also found white matter differences in the arcuate fasciculus according to the presence or absence of language. Our results suggest white matter disorganization and/or reduced myelination in ASD, and support abnormal anatomical connectivity in ASD within the social and the language networks.