Elsevier

NeuroImage

Volume 19, Issue 4, August 2003, Pages 1545-1555
NeuroImage

Regular article
Diffusion tractography based group mapping of major white-matter pathways in the human brain

https://doi.org/10.1016/S1053-8119(03)00190-3Get rights and content

Abstract

Tractography uses diffusion tensor imaging data to trace white matter pathways in vivo within the brain. We have constructed group maps that represent three major white matter tracts—the anterior callosal fibers, optic radiations, and pyramidal tracts—in a group of 21 volunteers. For each individual tract the fast marching tractography (FMT) algorithm was used to generate a VSC (voxel scale connectivity) map in native space. Using SPM99 these maps were transformed into a standard reference frame and three group mapping techniques were investigated: the first averaged the individual VSC maps, the second produced maps that demonstrate intersubject tract variability and degree of overlap, and the third used an SPM analysis to construct a statistical image that represents the group effect. The group maps reconstructed for each tract under investigation conform well to known anatomy and are consistent with data derived from postmortem human brains. Greater intersubject variability is found around the terminal projections of the tracts adjacent to cerebral cortex, whereas the “core” of each tract is characterized by lower variability. No significant differences were found between the left and right side of the pyramidal tracts and optic radiations. The group mapping techniques utilize the VSC maps in different but complementary ways. In the future, group mapping could investigate in vivo white matter differences between normal subjects and patients affected by neurological and psychiatric diseases.

Introduction

Investigating structural connections within the brain is fundamental to our understanding of the effects pathology may have on neural plasticity, development, and function. Until recently, studies of white matter pathways have been limited to postmortem and animal brains, using histological tracer techniques (Jones, 1999). However, the advent of diffusion tensor imaging (DTI) (Basser et al., 1994) has allowed the in vivo investigation of brain microarchitecture and, more recently, interregional fiber tracking Basser et al 2000, Conturo et al 1999, Xue et al 1999. DTI is a magnetic resonance imaging technique that provides information about the random displacement, or passive diffusion, of water molecules Basser et al 1994, Basser and Pierpaoli 1996. In human tissues, the mobility of water molecules is not the same in all directions, as its motion is limited by the presence of tissue components, such as cell membranes and fibers. If these components are arranged in a way that restricts the water diffusion in some directions, the displacement of water becomes directionally preferential, or anisotropic. For example, in brain white matter, in which the axons are organized in parallel bundles, water diffusion preferentially occurs in the direction of the axons rather than across them Le Bihan et al 1993, Pierpaoli and Basser 1996.

The success of DTI is in its ability to provide indirect information about the geometric organization of brain tissue (Le Bihan, 1995). The magnitude of white matter anisotropy and the dominant direction of diffusion that are measured using DTI, at the macroscopic voxel level, reflect the degree of fiber coherence and the tract orientation within the voxel. However, DTI does not directly give any information about how fiber tracts are connected between neighboring voxels. Such information may be inferred using diffusion tractography techniques (also referred to as fiber tracking or axonography). Several tractography methods have been developed in order to infer continuity of fiber orientation from voxel to voxel and to reconstruct connections between brain regions Basser et al 2000, Conturo et al 1999, Jones et al 1999b, Mori et al 1999, Mori et al 2000, Poupon et al 2000.

Fast marching tractography (FMT) Parker et al 2002a, Parker et al 2002b is a tractography technique that generates voxel scale connectivity2 (VSC) maps in vivo within the brain using the information provided by DTI. These maps provide at each voxel in the brain a scalar index that estimates the degree of connection to a particular starting voxel. The FMT method has received some validation by application to animal data (Parker et al., 2002a) and by comparison with standard atlases (Parker et al., 2002b). A recent study has been performed using FMT to investigate the reproducibility of quantitative tract measurements (Ciccarelli et al., 2003), and the maps generated by FMT corresponded well with known anatomy Burgel et al 1999, Rademacher et al 2001.

This work describes the evaluation of estimated white matter tract variability between normal subjects, as determined using FMT. This step is essential in order to move on to the investigation of neurological and psychiatric diseases, in which the white matter pathways are known to be disrupted Foong et al 2000, Mori et al 2002, Pierpaoli et al 2001, Steel et al 2001. The methods of creating group maps of white matter tracts, and therefore, compensating for normal intersubject variability, may ultimately allow the construction of “brain atlases” that represent in vivo white matter tracts. These would be invaluable for investigating how white matter is affected by pathological conditions of the central nervous system.

As an initial step toward this goal, we have generated group maps of white matter pathways in a group of 21 healthy subjects using FMT. Three clinically relevant white matter pathways were chosen: the left and right pyramidal tracts, the left and right optic radiations, and the anterior callosal fibers. We created three different group mapping techniques for each pathway and investigated the contribution of each technique to understanding the behaviour of the estimated pathways between subjects. The first technique simply averaged the individual VSC maps generated by FMT, the second produced maps that demonstrated intersubject tract variability and degree of overlap, and the third employed SPM (Ashburner and Friston, 2000) to construct a statistical image that represented the group effect. With the last technique we were also able to investigate any asymmetry between left and right sides for the estimated optic radiation and pyramidal tracts.

Section snippets

Subjects

Twenty-one healthy volunteers were studied (11 females and 10 males). The mean age was 33 ± 9.7 years. Informed consent was obtained from all subjects before entering into the study.

MRI protocol

All scans were performed on a 1.5-T Signa Echospeed MRI system (GE Medical Systems, Milwaukee, WI). The following data sets were acquired: (1) conventional proton density- and T2-weighted spin echo imaging (TR 2000, TE 30/120ms, FOV 240 mm, matrix 256 × 256; 28 contiguous axial slices; 5-mm slice thickness) and (2)

Average group maps

The average group maps across the whole group of subjects for the white matter pathways under investigation were created and then overlaid onto a T1-weighted template image (Fig. 1). On visual inspection they qualitatively conform to the known anatomy of the tracts. The color intensity of these maps reflects the average connectivity value of each voxel to the seed point across all subjects.

Variability group maps

The variability group maps were constructed using the thresholded maps and then were overlaid onto a

Discussion

We have used mapping methods with FMT to investigate how the pyramidal tracts, the optic radiations, and the anterior callosal fibers are represented in a group of 21 normal volunteers. We have constructed group maps in a standard reference frame that show the interindividual variability of tract location and shape and have generated statistical images that represent the group effect for each tract. A few other studies have recently addressed the issue of making group specific inferences with

Acknowledgements

We are very grateful for John Ashburner’s help. We thank David MacManus, Chris Benton, and Ros Gordon for their technical assistance with the MR scanning and Phil Boulby for the diffusion sequence development. Olga Ciccarelli is supported by TEVA Pharmaceutical Ltd. and Ahmed Toosy by the Brain Research Trust. The scanning was funded by the MS Society of Great Britain and Northern Ireland.

References (42)

  • C. Poupon et al.

    Regularization of diffusion-based direction maps for the tracking of brain white matter fascicles

    Neuroimage

    (2000)
  • R.M. Steel et al.

    Diffusion tensor imaging (DTI) and proton magnetic resonance spectroscopy (1H MRS) in schizophrenic subjects and normal controls

    Psychiat. Res.

    (2001)
  • D. Xu et al.

    A framework for callosal fiber distribution analysis

    Neuroimage

    (2002)
  • D.C. Alexander et al.

    Spatial transformations of diffusion tensor magnetic resonance images

    IEEE Trans. Med. Imaging

    (2001)
  • P.J. Basser et al.

    In vivo fiber tractography using DT-MRI data

    Magn Reson. Med.

    (2000)
  • P.J. Basser

    Relationships between diffusion tensor and q-space MRI

    Magn Reson. Med.

    (2002)
  • A. Brodal

    Descending supraspinal pathways

  • T.E. Conturo et al.

    Tracking neuronal fiber pathways in the living human brain

    Proc. Natl. Acad. Sci. USA

    (1999)
  • M.C. de Lacoste et al.

    Topography of the human corpus callosum

    J. Neuropathol. Exp. Neurol.

    (1985)
  • F. Farzaneh et al.

    Analysis of T2 limitations and off-resonance effects on spatial resolution and artifacts in echo-planar imaging

    Magn. Reson. Med.

    (1990)
  • J. Foong et al.

    Neuropathological abnormalities of the corpus callosum in schizophreniaa diffusion tensor imaging study

    J. Neurol. Neurosurg. Psychiat.

    (2000)
  • Cited by (0)

    1

    These authors contributed equally to this work.

    View full text