Diffusion tensor imaging of the corpus callosum in Autism

Abstract

The corpus callosum is the largest commissural white matter pathway that connects the hemispheres of the human brain. In this study, diffusion tensor imaging (DTI) was performed on subject groups with high-functioning autism and controls matched for age, handedness, IQ, and head size. DTI and volumetric measurements of the total corpus callosum and subregions (genu, body and splenium) were made and compared between groups. The results showed that there were significant differences in volume, fractional anisotropy, mean diffusivity, and radial diffusivity between groups. These group differences appeared to be driven by a subgroup of the autism group that had small corpus callosum volumes, high mean diffusivity, low anisotropy, and increased radial diffusivity. This subgroup had significantly lower performance IQ measures than either the other individuals with autism or the control subjects. Measurements of radial diffusivity also appeared to be correlated with processing speed measured during the performance IQ tests. The subgroup of autism subjects with high mean diffusivity and low fractional anisotropy appeared to cluster with the highest radial diffusivities and slowest processing speeds. These results suggest that the microstructure of the corpus callosum is affected in autism, which may be related to nonverbal cognitive performance.

Introduction

The neuropathology of autism appears to be complex (Lainhart, 2006). Several lines of evidence suggest that abnormalities of the corpus callosum are involved. Complex information processing, which requires cortico-cortical interhemispheric as well as intrahemispheric transfer of information, has been found to be deficient across multiple domains in autism (Minshew et al., 1997). Abnormalities of corpus callosum mid-sagittal area and white matter density have been found (Chung et al., 2004, Egaas et al., 1995, Filipek, 1996, Hardan et al., 2000, Manes et al., 1999, Piven et al., 1997, Vidal et al., 2006, Waiter et al., 2005). Interhemispheric functional underconnectivity has been suggested by several recent fMRI studies of autism in language processing and working memory (Just et al., 2004, Koshino et al., 2005, respectively). Postmortem studies have found thinning of the corpus callosum in some cases (Bailey et al., 1998). A better understanding of corpus callosum white matter differences between autistic individuals and controls, in general and throughout development, may help identify potential neuroanatomical markers of autism and important neurobiological subtypes of the disorder.

The corpus callosum is responsible for conduction of signals between homologous and heterotopic cortical regions and is an essential component for brain lateralization and interhemispheric communication (Innocenti, 1986, Pandya and Seltzer, 1986, Zaidel and Iacoboni, 2003). The most rostral region of the corpus callosum, the genu and the rostrum (hereafter referred to as the genu), has connections between prefrontal brain regions (Witelson, 1989). The most caudal region, the splenium, contains connections between occipital, temporal and parietal regions (Witelson, 1989). The midsections between the genu and splenium are the body and isthmus (hereafter referred to as the body). Methods for subdividing the corpus callosum into subregions have been developed although there is no clear consensus regarding approaches for the divisions (de Lacoste et al., 1985, Witelson, 1989, Clarke and Zaidel, 1994). Recently Huang et al. (2005) used diffusion tensor imaging (DTI) and white matter tractography to parcellate the corpus callosum into regions that connect to specific cortical areas.

Neuroimaging techniques have provided insight into general corpus callosum development and more localized changes that occur with age (see reviews in Brambilla et al., 2003, Lainhart et al., 2005). In typically developing individuals, overall area of the corpus callosum increases during childhood and adolescence. The greatest increase occurs in the posterior regions of the corpus callosum and during the childhood years (Giedd et al., 1999, Giedd et al., 1996, Keshavan et al., 2002). Overall callosal area may continue to increase into the twenties (Pujol et al., 1993), but the anterior corpus callosum may reach adult size well before then (Giedd et al., 1996).

Studies investigating the development of the corpus callosum in autism have provided mixed results. Although a decrease in the area of the anterior corpus callosum has been noted (Hardan et al., 2000), a recent study (Rice et al., 2005) found no differences in area, shape, or contour of the corpus callosum between autistic participants with macrocephaly and normal participants with benign macrocephaly. These findings suggest that group differences in head size may have influenced previous results. In addition to area, another type of measure may be necessary to identify abnormalities in the corpus callosum in autism.

Diffusion tensor imaging (DTI) is a non-invasive method for mapping the diffusion properties of tissue water (Basser and Pierpaoli, 1996). DTI is extremely sensitive to subtle differences in the architecture of white matter at the microstructural level. The white matter tracts of the corpus callosum are highly coherent which makes them well suited for study with DTI. The diffusion tensor defines the magnitude, anisotropy (variation of the diffusion properties with direction) and orientation of anisotropic water diffusion in biological tissues. The diffusion tensor may be decomposed into three principal eigenvalues with corresponding eigenvectors. The major eigenvector (e₁), corresponding to the largest eigenvalue (λ₁), also referred to as the axial diffusivity D_a, is the direction of fastest diffusivity and is generally assumed to be parallel to the direction of axon bundles in white matter. The medium and smallest eigenvalues (λ₂ and λ₃, respectively) are assumed to be perpendicular to the white matter tracts. A measure of diffusivity in the perpendicular plane is the radial diffusivity, D_r = (λ₂ + λ₃)/2 (Song et al., 2002). The average of the three eigenvalues is referred to as the mean diffusivity (MD). A commonly used measure of diffusion anisotropy is the fractional anisotropy (FA), which is a normalized (ranges between 0 and 1) version of the eigenvalue standard deviation (Pierpaoli and Basser, 1996).

To date, the only published study examining DTI in autistic children showed lower FA values in the genu and rostral body in seven high-functioning autistic male children and adolescents compared to controls (Barnea-Goraly et al., 2004). Filippi et al. (2003) noted significantly lower FA and higher MD values in both the genu and splenium of developmentally delayed children compared to controls, although autistic participants were excluded from their study. Thus, the present study was designed to elucidate some of the developmental changes that occur in the corpus callosum in autism.

In this study, diffusion tensor measurements (MD, FA, D_a and D_r) in corpus callosum were investigated in a large group of high-functioning autistic male children, adolescents, and young adults compared to matched controls. Relationships between age, volume of the corpus callosum, FA, and MD measures were explored. Potential functional correlates of white matter organization, such as IQ and social functioning, were also examined.