Atlasing location, asymmetry and inter-subject variability of white matter tracts in the human brain with MR diffusion tractography

Abstract

The purpose of this study is to create a white matter atlas of the human brain using diffusion tensor imaging (DTI) tractography and to describe the constant and variable features of the major pathways. DTI was acquired from 40 healthy right-handed adults and reconstructed tracts mapped within a common reference space (MNI). Group effect maps of each tract defined constant anatomical features while overlap maps were generated to study inter-subject variability and to compare DTI derived anatomy with a histological atlas. Two patients were studied to assess the localizing validity of the atlas. The DTI-derived maps are overall consistent with a previously published histological atlas. A statistically significant leftward asymmetry was found for the volume and number of streamlines of the cortico-spinal tract and the direct connections between Broca's and Wernicke's territories (long segment). A statistically significant rightward asymmetry was found for the inferior fronto-occipital fasciculus and the fronto-parietal connections (anterior segment) of the arcuate fasciculus. Furthermore, males showed a left lateralization of the fronto-temporal segment of the arcuate fasciculus (long segment), while females had a more bilateral distribution. In two patients with brain lesions, DTI was acquired and tractography used to show that the tracts affected by the lesions were correctly identified by the atlas. This study suggests that DTI-derived maps can be used together with a previous histological atlas to establish the relationship of focal lesions with nearby tracts and improve clinico-anatomical correlation.

Introduction

Until the advent of diffusion tensor imaging (DTI), our knowledge of white matter anatomy was based on a small number of influential 19th and early 20th century post-mortem dissection atlases (Burdach, 1819, Déjerine, 1895). In common with their contemporary counterparts (Talairach and Tournoux, 1988), these atlases emphasize the constant or average anatomy of representative subjects at the expense of normal variability between subjects. Few post-mortem histological studies addressed the variability of the tracts between the two hemispheres and reported asymmetries for the cortico-spinal tract (CST) (Flechsig, 1876, Yakovlev & Rakic, 1966, Rademacher et al., 2001), the optic radiations (Bürgel et al., 1999), and the uncinate (Highley et al., 2002). Bürgel et al. (2006) showed that a significant inter-subject variability also exists for each tract within the single hemispheres. Little is known about whether such anatomical variability differs between genders and extends to other tracts underlying complex cognitive functions.

Diffusion tensor imaging has allowed the study of the in vivo anatomy of white matter tracts in the human brain (Catani et al., 2002, Mori et al., 2002, Mori et al., 2005, Wakana et al., 2004, Catani & Thiebaut de Schotten, 2008, Ciccarelli et al., 2008, Lawes et al., 2008) and addressed some of the questions that were difficult to answer with post-mortem dissections such as the location, asymmetry and inter-subject variability of white matter tracts (Ciccarelli et al., 2003, Barrick et al., 2007, Catani et al., 2007, Wakana et al., 2007, Verhoeven et al., 2010). In the recent years several groups have used DTI to produce group atlases of the major white matter tracts (Hua et al., 2008, Lawes et al., 2008, Verhoeven et al., 2010, Wassermann et al., 2010). These atlases contain maps of the major white matter tracts that have a good correspondence with post-mortem blunt dissections. However, none of the previous atlases have addressed the inter-subject variability between left and right hemisphere in relation to gender.

The aim of this study was to create a normative atlas of white matter human connections for clinical and research purposes on a large age-matched population of male and female participants. We combine for the first time complementary approaches to address the question whether the lateralization of the association and projection fibres differs between genders and hemispheres. A comparison between the DTI-derived atlas and a previously published post-mortem histological atlas (Bürgel et al., 2006) is also presented and limitations of both approaches discussed. An atlas of the anatomical variability in the normal population could help improve localization of white matter lesions in brain disorders and understand mechanisms of brain recovery and plasticity (Amunts and Willmes, 2006).