Increased left hemisphere impairment in high-functioning autism: A tract based spatial statistics study

Highlights

• We used diffusion tensor imaging to compare participants with autism to controls.

• Participants with autism demonstrate greater white matter impairment.

• The autism group had greater impairment in left compared to right hemisphere.

• Left hemisphere impairment in autism was attributable to myelination anomalies.

Abstract

There is evidence emerging from Diffusion Tensor Imaging (DTI) research that autism spectrum disorders (ASD) are associated with greater impairment in the left hemisphere. Although this has been quantified with volumetric region of interest analyses, it has yet to be tested with white matter integrity analysis. In the present study, tract based spatial statistics was used to contrast white matter integrity of 12 participants with high-functioning autism or Aspergers syndrome (HFA/AS) with 12 typically developing individuals. Fractional Anisotropy (FA) was examined, in addition to axial, radial and mean diffusivity (AD, RD and MD). In the left hemisphere, participants with HFA/AS demonstrated significantly reduced FA in predominantly thalamic and fronto-parietal pathways and increased RD. Symmetry analyses confirmed that in the HFA/AS group, WM disturbance was significantly greater in the left compared to right hemisphere. These findings contribute to a growing body of literature suggestive of reduced FA in ASD, and provide preliminary evidence for RD impairments in the left hemisphere.

Introduction

Autism spectrum disorders (ASD) are pervasive, developmental, neurological conditions, which adversely impact behavior in two key domains: social communication, and repetitive and/or stereotyped patterns of behavior (DSM V, 2013). ASD is estimated to have a global prevalence of approximately 2.0% of the population (Blumberg et al., 2013). The symptoms of ASD are believed to be associated with atypical neurological development, where environmental factors alter the function of complex, multi-focal, neural networks across the lifespan (Amaral et al., 2008). Support for this view comes from fMRI evaluations of functional connectivity (FC), which examines inter-regional, co-activation across the brain. Among participants with ASD, research has demonstrated both increased and decreased FC compared to typically developing (TD) individuals (cf. reviews by Just et al., 2012, Muller et al., 2011).

An inherent limitation of FC is that it is unable to probe the underlying structure and organization of white matter (WM) that underlies cortical connectivity (Travers et al., 2012). Overcoming this limitation, Diffusion Tensor Imaging (DTI) is a non-invasive technique that can identify differences in microstructural and macroscopic organization of WM (Lange et al., 2010). In WM bundles, the membranes of axons and myelin cause the diffusivity of water perpendicular to WM tracts (Radial Diffusivity) to decrease relative to directions parallel to WM (axial) (Lee et al., 2007). This directional restriction is known as diffusion anisotropy, and is represented by three eigenvalues (λ₁, λ₂, and λ₃), which reflect the length of each eigenvector (Travers et al., 2012).

In DTI, several measures can be extracted. Fractional Anisotropy (FA) is a normalized value ranging from 0 to 1, which represents the fraction of the tensor that can be assigned to anisotropic diffusion, and is sensitive to structural differences in myelination, axonal density, axonal caliber and fiber coherence (Cheng et al., 2010). Mean diffusivity (MD) represents the average radius of the three eigenvalues, and is sensitive to the density of tissue barriers in all directions.

Separating the eigenvalues is argued to provide a more complete picture of WM structure (Song et al., 2005). Axial diffusivity (AD) looks at water diffusivity parallel to WM tracts (λ₁), whilst Radial Diffusivity (RD) expresses water diffusivity perpendicular to tracts ([λ₂,+λ₃]/2). RD is believed to be sensitive to dysmyelination and demyelination (Harsan et al., 2006), whilst AD is sensitive to axonal injury (Travers et al., 2012).

There are several approaches to analyzing DTI data. Region of interest (ROI) based analyses involve placing a seed, which follows pathways of maximal diffusion to reconstruct a specific WM tract. However, this method is limited in that only preselected WM tracts can be investigated (Gibbard et al., 2013). Voxel-based analyses such as statistical parametric mapping (SPM) have also been utilized, but are limited by issues related to registration error and smoothing techniques (Jones et al., 2005). Tract based spatial statistics (TBSS) addresses several of these shortcomings by implementing non-linear registration, producing an FA WM skeleton with alignment invariant tract representation, and by avoiding smoothing statistics which do not require normally distributed data (Tamm et al., 2012).

Compared to TD participants, the most common finding to arise from TBSS based analyses of ASD participants has been reductions in FA, and increases in MD distributed widely across the brain (Cheon et al., 2011, Kumar et al., 2010; Shukla et al., 2011). Using TBSS, WM tracts commonly associated with reduced FA in participants with ASD include the superior longitudinal fasciculus (SLF), cingulum bundle, uncinate fasciculus (UF) and the corpus callosum (Barnea-Goraly et al., 2010, Jou et al., 2011; Kumar et al., 2010; Pardini et al., 2009, Shukla et al., 2010, Thakkar et al., 2008). There are also reports of increased FA in ASD samples (Billeci et al., 2012, Cheng et al., 2010, Weinstein et al., 2011), which is usually attributable to younger cohorts of participants. Additionally, there are reports of increased RD . (Amies et al., 2011; Shukla et al., 2010), and decreased AD (Barnea-Goraly et al., 2010). Thus, the distribution of WM anomalies in participants with ASD remains heterogeneous, and there is a need to identify more specific WM disturbances linked to the disorder.

A past theory (Hier et al., 1979) that has received renewed interest since the advent of DTI is that participants with ASD demonstrate greater left hemisphere impairment. Among TD participants, there is a trend toward increased FA in the arcuate and uncinate fasciculi of the left hemisphere (Catani et al., 2007). In participants with an ASD, ROI based research has identified reduced lateralization in the left hemisphere (Fletcher et al., 2010, Lo et al., 2011, Nagae et al., 2012), whilst another study investigating tensor shape reports this asymmetry as being reversed in ASD, with reduced FA in the left compared to right hemisphere (Lange et al., 2010). Moreover, a recent meta-analysis of six previous DTI studies reported that participants with ASD demonstrated significantly reduced FA in the left, but not right SLF and UF (Aoki et al., 2013). There is also evidence from volumetric (Rojas et al., 2002) and functional neuroimaging (Eyler et al., 2012) research for greater left hemisphere impairment in participants with ASD.

No research to date has implemented voxel-based methods to explore hemispheric WM differences in ASD participants. Thus, the present study used TBSS to examine WM integrity of adolescent and adult participants with high functioning autism or Aspergers׳ syndrome (HFA/AS). In this study FA, MD, RD and AD were investigated as outcome measures. Based upon previous literature, two hypotheses were generated. Firstly, that the HFA/AS group would demonstrate reduced FA by comparison to TD participants. Secondly, based upon the limited literature to have investigated hemispheric differences, the HFA/AS group would demonstrate greater WM impairment in the left compared to right hemisphere.