Neurite Orientation Dispersion and Density Imaging in Autism Spectrum Disorders

Background: Diffusion tensor imaging (DTI) is an integral neuroimaging technique to assess white matter microstructure and has been influential to the study of white matter alterations in ASD. However, one of the major limitations of DTI is that the quantitative measures estimated via this technique are inherently non-specific; that is, a variety of microstructural or biochemical processes could lead to similar DTI parameter estimates. As a result, the ability to identify specific microstructural features using DTI is limited. Biophysical modeling techniques, such as neurite orientation dispersion and density imaging (NODDI), may improve the level of microstructure specificity available and thus could provide new insights into the microstructural changes observed in ASD.

Objectives: We sought to investigate, for the first time, white matter microstructural differences between individuals with and without a ASD diagnosis using the NODDI imaging technique.

Methods: Participants for this study consisted of 108 individuals between 5 and 42 years of age, 41 of which were diagnosed with ASD. A three-shell diffusion weighted imaging (DWI) protocol was acquired with b-values of 350, 800, and 2000 s/mm² at 63 non-collinear diffusion encoding directions on a 3.0 Tesla GE MR750 scanner. Following acquisition, DWI images were corrected for eddy-current distortions and head motion and subsequently fit to the NODDI model, yielding parameter maps of intracellular and isotropic volume fractions (ν_IC, ν_ISO, respectively), and orientation dispersion index (ODI). These maps were nonlinearly aligned to a population specific template using DTI-TK, while permutation testing adjusted by age and a threshold free clustering approach was used to examine voxelwise differences between the ASD and typical development (TD) groups.

Results: Comparison of NODDI measures yielded widespread differences between the ASD and TD groups (Fig. 1). Specifically, TD controls were observed to have higher ν_IC within regions including the internal capsules, thalamus, brain stem and pons (p<0.05, TFCE-adjusted). Differences in ODI were mixed. For example, controls were observed to have higher ODI with deep white matter areas, such as the internal capsules and brain stem, while the ASD group was observed to have higher ODI in more peripheral white matter, including the genu of the corpus callosum, and regions of parietal white matter (p<0.05, TFCE-adjusted). Isotropic volume fraction was observed to be higher in the ASD group extensively across the brain (p<0.05, TFCE-adjusted).

Conclusions: Our preliminary findings suggest a reduction of ν_IC and ODI in regions of the internal capsules, thalamus and brain stem within the ASD group, while having increased ODI and ν_ISO in more widespread brain regions. These findings agree well with the current DTI literature that describe alterations of white matter microstructure associated with ASD, while also suggesting the possibility that these white matter alterations may stem from alterations to the neurite architecture, such as decreased neurite density and/or increased white matter angular dispersion. Future analyses will investigate the age-related relationships of these neurobiological changes and the behavioral processes that may be involved with these alterations.