Elsevier

Cortex

Volume 44, Issue 8, September 2008, Pages 1105-1132
Cortex

Special issue: Research report
A diffusion tensor imaging tractography atlas for virtual in vivo dissections

https://doi.org/10.1016/j.cortex.2008.05.004Get rights and content

Abstract

Diffusion tensor imaging (DTI) tractography allows perform virtual dissections of white matter pathways in the living human brain. In 2002, Catani et al. published a method to reconstruct white matter pathways using a region of interest (ROI) approach. The method produced virtual representations of white matter tracts faithful to classical post-mortem descriptions but it required detailed a priori anatomical knowledge. Here, using the same approach, we provide a template to guide the delineation of ROIs for the reconstruction of the association, projection and commissural pathways of the living human brain. The template can be used for single case studies and case–control comparisons. An atlas of the 3D reconstructions of the single tracts is also provided as anatomical reference in the Montreal Neurological Institute (MNI) space.

Introduction

The possibility of performing virtual dissections of white matter tracts and visualizing pathways in the living human brain is one of the most promising applications of diffusion tensor imaging (DTI) tractography (Catani and Mesulam, 2008a, Catani, 2006). Current DTI tractography methods require the delineation of regions of interest (ROIs) as starting “seed points” for tracking (Jones, 2008, this issue). One approach for ROIs delineation is the automatic application of normalized cortical or subcortical masks to single brain data sets (see for example, Lawes et al., 2008). The use of cortical masks is of particular help when trying to reduce tractography analysis time and operator-dependent biases. But these methods perform poorly when applied to pathological brains (e.g., when the anatomy is distorted by the underlying pathological process) or when the experimenter aims at describing inter-individual variability in tract anatomy (e.g., studying differences in the cortical projections of an individual tract in the normal population). Also the use of cortical masks is prone to generate artefactual reconstructions of tracts due to high uncertainty of the fiber orientation in the cortical voxels or surrounding white matter (Jones, 2003, Jones, 2008). An alternative strategy is to define the ROIs manually. This approach may overcome some of the problems mentioned above and has been successfully used in several tractography studies (Conturo et al., 1999, Concha et al., 2005, Basser et al., 2000, Catani et al., 2002, Mori et al., 2000). One limiting step of this second approach is that the method requires a priori anatomical knowledge to identify the course of white matter pathways and delineate ROIs on DTI images. Here, we provide a tool to teach tractography-derived white matter anatomy and to perform virtual in vivo dissections of the major tracts of the human brain. First, we have created a 3D tractography atlas of the associative, commissural and projection pathways in a standardized system of coordinates (Montreal Neurological Institute, MNI space) (Fig. 1, Fig. 2, Fig. 3, Fig. 4, Fig. 5, Fig. 6, Fig. 7, Fig. 8, Fig. 9, Fig. 10, Fig. 11). The atlas, together with the description of each tract in the figure legends, can be used as anatomical reference. Second, we provide some guidelines for the identification of the pathways in the color maps and show how to delineate ROIs on axial fractional anisotropy (FA) images from an average data set (Fig. 11). We hope that the atlas and the template for ROIs will be used as a tool for teaching and guiding virtual brain dissections in single cases and case–control studies.

Section snippets

DTI data set acquisition, processing, and averaging

Twelve right-handed male subjects (34.3 ± 5.7 years old) gave written consent to participate in the study, which was approved by the local ethics committee at the Service Hospitalier Frédéric Joliot, Orsay. MRI data were acquired using echo-planar imaging at 1.5 T (General Electric Healthcare Signa) with a standard head coil for signal reception. High resolution T1-weighted anatomical images were acquired (gradient-echo sequence, repetition time 9.9 ms, echo time 2 ms, matrix 256 × 192, field of view

Acknowledgements

We would like to thank Cyril Poupon for the MRI diffusion database. The database is the property of CEA SHFJ/UNAF and can be provided on demand to [email protected]. Data were post-processed with AIMS/Anatomist/BrainVisa software, freely available at http://brainvisa.info.

We also would like to thank Flavio Dell'acqua and Luca Pugliese from the NATBRAINLAB (http://www.natbrainlab.com) for the helpful discussion. This project was generously supported by the Medical Research Council (UK)

References (32)

Cited by (0)

View full text