Constrained spherical deconvolution-based tractography and tract-based spatial statistics show abnormal microstructural organization in Asperger syndrome

Abstract

BackgroundThe aim of this study was to investigate potential differences in neural structure in individuals with Asperger syndrome (AS), high-functioning individuals with autism spectrum disorder (ASD). The main symptoms of AS are severe impairments in social interactions and restricted or repetitive patterns of behaviors, interests or activities.MethodsDiffusion weighted magnetic resonance imaging data were acquired for 14 adult males with AS and 19 age, sex and IQ-matched controls. Voxelwise group differences in fractional anisotropy (FA) were studied with tract-based spatial statistics (TBSS). Based on the results of TBSS, a tract-level comparison was performed with constrained spherical deconvolution (CSD)-based tractography, which is able to detect complex (for example, crossing) fiber configurations. In addition, to investigate the relationship between the microstructural changes and the severity of symptoms, we looked for correlations between FA and the Autism Spectrum Quotient (AQ), Empathy Quotient and Systemizing Quotient.ResultsTBSS revealed widely distributed local increases in FA bilaterally in individuals with AS, most prominent in the temporal part of the superior longitudinal fasciculus, corticospinal tract, splenium of corpus callosum, anterior thalamic radiation, inferior fronto-occipital fasciculus (IFO), posterior thalamic radiation, uncinate fasciculus and inferior longitudinal fasciculus (ILF). CSD-based tractography also showed increases in the FA in multiple tracts. However, only the difference in the left ILF was significant after a Bonferroni correction. These results were not explained by the complexity of microstructural organization, measured using the planar diffusion coefficient. In addition, we found a correlation between AQ and FA in the right IFO in the whole group.ConclusionsOur results suggest that there are local and tract-level abnormalities in white matter (WM) microstructure in our homogenous and carefully characterized group of adults with AS, most prominent in the left ILF.

Informed constrained spherical deconvolution (iCSD)

“…However, their method produces non-zero fiber densities also in the CSF, unlike the proposed iCSD method. In CSD, which is a widely used method (Metzler-Baddeley et al, 2012a;Emsell et al, 2013;Forde et al, 2013;Kristo et al, 2013;McGrath et al, 2013a;McGrath et al, 2013b;Reijmer et al, 2013b;Reijmer et al, 2013a;Thompson et al, 2014;Roine et al, 2015), the isotropic PVEs were not taken into account earlier.…”

“…CSD is currently one of the most used, clinically feasible and readily available HARDI methods Metzler-Baddeley et al, 2012a;Emsell et al, 2013;Farquharson et al, 2013;Forde et al, 2013;Kristo et al, 2013;McGrath et al, 2013a;McGrath et al, 2013b;Reijmer et al, 2013b;Reijmer et al, 2013a;Thompson et al, 2014;Roine et al, 2015). In CSD, the fiber orientation distribution function (fODF) is estimated from clinically feasible DW data by deconvolving the data with a kernel constructed from a response function (RF) representing a single coherently oriented population of fibers (Tournier et al, 2004;Tournier et al, 2007).…”

Function–structure connectivity in patients with severe brain injury as measured by MRI‐DWI and FDG‐PET

“…Tractography based on constrained spherical deconvolution is optimal with b-values of 2,500-3,000 s/mm 2 [Tournier et al, 2013], but crossing fibres can still be more reliably modelled than with simple DTI-based models using our lower b-value of 1,000 s/mm 2 [e.g., see for effective application: Roine et al, 2015]. Future studies should strive to acquire diffusion-weighted images using isotropic voxels, as anisotropic voxel sizes produce datasets in which the fibre orientation estimates depend on the position of the subject in the scanner.…”

Tracking Dynamic Interactions Between Structural and Functional Connectivity: A TMS/EEG-dMRI Study

“…However, despite of a low b-value, with a sufficient amount of directions crossing fibers can be reliably modeled with CSD and the result is still significantly better than with a simple DTI-based model, e.g., see (Roine et al, 2015) for a successful application.…”

“…For both the response estimation and spherical deconvolution steps we chose a maximum harmonic order max l of 6, based on a successful application of constrained spherical deconvolution with similar b-value and number of directions (Roine et al, 2015).…”