Less white matter concentration in autism: 2D voxel-based morphometry

Abstract

Autism is a neurodevelopmental disorder affecting behavioral and social cognition, but there is little understanding about the link between the functional deficit and its underlying neuroanatomy. We applied a 2D version of voxel-based morphometry (VBM) in differentiating the white matter concentration of the corpus callosum for the group of 16 high functioning autistic and 12 normal subjects. Using the white matter density as an index for neural connectivity, autism is shown to exhibit less white matter concentration in the region of the genu, rostrum, and splenium removing the effect of age based on the general linear model (GLM) framework. Further, it is shown that the less white matter concentration in the corpus callosum in autism is due to hypoplasia rather than atrophy.

Introduction

Autism is a neurodevelopmental disorder of brain function that has begun to attract in vivo structural magnetic resonance imaging (MRI) studies in the region of the corpus callosum Egaas et al., 1995, Hardan et al., 2000, Manes et al., 1999, Piven et al., 1996, Piven et al., 1997. There is little understanding about the link between the functional deficit and the underlying abnormal anatomy in autism, which provides motivation for our study. These studies use the Witelson partition or a similar partition scheme of the corpus callosum (Witelson, 1989). Witelson partitioned the midsagittal cross-sectional images of the corpus callosum along the maximum anterior–posterior line (Talairach and Tournoux, 1988) and defined the region of the genu, rostrum, midbodies, isthmus, and splenium from the anterior to posterior direction. Based on the Witelson partition, there has been a consistent finding in abnormal reduction in anterior, midbody, and posterior of the corpus callosum Brambilla et al., 2003a, Brambilla et al., 2003b.

Piven et al. (1997) compared 35 autistic individuals with 36 normal control subjects controlling for total brain volume, gender, and IQ and detected a statistically significant smaller midbody and posterior regions of the corpus callosum in the autistic group. Manes et al. (1999) compared 27 low functioning autistic individuals with 17 normal controls adjusting for the total brain volume. They found a smaller corpus callosum compared to the control group in genu, rostrum, anterior midbody, posterior midbody, and isthmus but did not find statistically significant differences in the rostrum and the splenium, although the sample mean of the rostrum and splenium size are smaller than that of the control group. Hardan et al. (2000) compared 22 high functioning autistic to 22 individually matched control subjects and showed smaller genu and rostrum of the corpus callosum adjusting for the total brain volume based on the Witelson partition. The smaller corpus callosum size was considered as an indication of a decrease in interhemispheric connectivity. They did not detect other regions of significant size difference. For an extensive review of structural MRI studies for autism that have been published between 1966 and 2003, one may refer to Brambilla et al., 2003a, Brambilla et al., 2003b.

The shortcoming of the Witelson partition is the artificial partitioning. The Witelson partition may dilute the power of detection if the anatomical difference occurs near the partition boundary. Alternative voxel-wise approaches that avoid predefined regions of interests (ROI) have begun to be used in structural autism studies. Vidal et al. (2003) used the tensor-based morphometry (TBM) to show reduced callosal thickness in the genu, midbody, and splenium in autistic children. Hoffmann et al. (2004) used a similar TBM to show curvature difference in the midbody. Abell et al. (1999) used voxel-based morphometry (VBM) Ashburner and Friston, 2000, Ashburner and Friston, 2001, Good et al., 2001a, Good et al., 2001b, Wright et al., 1995 in high functioning autism to show decreased gray matter volume in the right paracingulate sulcus, the left occipito-temporal cortex, and increased amygdala and periamygdaloid cortex.

The advantage of the VBM framework over the Witelson partition approach is that it is completely automated and does not require artificial partitioning of the corpus callosum that introduces undesirable bias. Further, it is not restricted to a priori ROIs enabling us to perform the statistical analysis at each voxel level and to pinpoint the exact location of the anatomical differences within ROI even if there is no ROI size differences. Although VBM was originally developed for whole-brain 3D morphometry, our study concentrates on the midsagittal cross-sectional corpus callosum regions to be able to compare the result with the previous 2D Witelson partition studies such as Hardan et al. (2000), Manes et al. (1999), and Piven et al. (1997). Hence, we will prefer 2D VBM over 3D VBM in this study. Let us review the basis of VBM and its connection to ROI morphometry briefly.