Diffusion tensor imaging tractography and reliability analysis for limbic and paralimbic white matter tracts

https://doi.org/10.1016/j.pscychresns.2007.11.007Get rights and content

Abstract

Diffusion tensor imaging (DTI) provides the opportunity to study white matter tracts in vivo. The goal was to estimate the reliability of DTI tractography for the analysis of limbic and paralimbic white matter. Normative data from 24 healthy subjects and reliability data from four healthy and four depressed subjects were acquired at 1.5 Tesla, using twice-refocused spin-echo, echoplanar DTI and Fluid-Attenuated Inversion Recovery (FLAIR) DTI sequences. Fiber tracking was performed using the Fiber Assignment by Continuous Tracking algorithm. Fractional Anisotropy (FA), trace Apparent Diffusion Coefficient and tract volumes were calculated. The inter-rater (and intra-rater) intraclass correlation coefficients for FA values were as follows: rostral cingulum 0.89 (0.87), dorsal cingulum 0.85 (0.90), parahippocampal cingulum 0.85 (0.95), uncinate fasciculus 0.85 (0.87), medial prefrontal white matter 0.97 (0.99), ventromedial prefrontal white matter 0.92 (0.93), crus of fornix 0.80 (0.81). The reported DTI protocol provides a reliable method to analyze limbic and paralimbic white matter tracts relevant to psychiatric disorders.

Introduction

Diffusion tensor imaging (DTI) provides a method to study white matter organization in vivo on the sub-millimeter scale (Le Bihan, 2003, Mori and Zhang, 2006). DTI enables quantification of the diffusion of water in brain tissue, where water diffusion is not random, but restricted by barriers, including cell membranes, cytoskeletal elements and myelin (Beaulieu, 2002). In white matter tracts, where axons are aligned in bundles, water diffuses more readily along axonal paths than across them. DTI has been used to investigate the organization of white matter in aging and normal development (Sullivan and Pfefferbaum, 2006, Huppi and Dubois, 2006), psychiatric disorders (Kubicki et al., 2007, Lim and Helpern, 2002) and neurological disorders (Ge et al., 2005, Sundgren et al., 2004).

Three broad approaches have been applied to DTI data analysis: region of interest (ROI)-based, voxel-based statistical mapping and tractography. Most psychiatric studies have used ROI-based methods (Kubicki et al., 2007, Lim and Helpern, 2002), which have drawbacks that include potential inconsistency or bias in placing ROIs, contamination by the inclusion of CSF or grey matter in the ROIs, inclusion of white matter from different projections and relatively high measurement variance (Kanaan et al., 2006, Snook et al., 2007). Color-coded primary diffusion maps that show fiber orientation can assist with ROI definition (Hermoye et al., 2006) and partial volume effects can be reduced if the ROIs are tailored to the shape of the structure (Snook et al., 2007).

In voxel-based analyses, images are transformed into standard space and tested for voxel-wise group differences in DTI measures: such differences can be mapped across the entire brain and a priori hypotheses are not required. However, there are problems with methods based on image co-registration using non-diffusion weighted (i.e. b = 0 s/mm²) images (see Snook et al., 2007, Jones et al., 2005), and the use of study-specific Fractional Anisotropy (FA) templates (Kyriakopoulos et al., 2007), iterative template correction (Ardekani and Sinha, 2006) or Tract-Based Spatial Statistics (Smith et al., 2006) has been proposed to address such limitations. Nevertheless, although voxel-based approaches can identify local differences in white matter structure, they do not determine whether such differences fall within defined anatomical connections.

Diffusion tractography is an alternative or complementary method to ROI- and voxel-based approaches (Mori and van Zijl, 2002, Kanaan et al., 2006, Heiervang et al., 2006). The principle is to follow the maximal diffusion direction voxel by voxel to model fiber tracts. Advantages are that a large number of voxels are sampled from the tract, reducing measurement variance compared with 2D ROIs (Kanaan et al., 2006), and that measurements are localized to specific projections. For quantitative studies it is essential to use reliable DTI analyses, but fewer reliability studies have been reported for tractography than for 2D ROI studies. Intra- and inter-rater reliability have been reported for analyses of several tracts, from repeat analyses of the same images or from analyses of repeated scans, and reliabilities have been compared for different acquisition parameters, different methods of tract propagation and selection, and different levels of rater experience (Buchsbaum et al., 2006, Ciccarelli et al., 2003, Heiervang et al., 2006, Wakana et al., 2007). Some studies have reported the reliability of tract delineation and there are fewer data on the reliability of tract-specific measures, such as FA.

The goal of the present study was to estimate reliability for DTI-tractography measures of FA, trace Apparent Diffusion Coefficient (ADC) and tract volume, using DTI acquisitions with six diffusion gradient directions and a readily available analytical package for tractography (Jiang et al., 2006). White matter tracts were selected that are well-defined anatomically and relevant to psychiatric disorders, in which hippocampal, cingulate and prefrontal volumetric MRI abnormalities have repeatedly been reported (Campbell et al., 2004, Campbell and MacQueen, 2006, Karl et al., 2006, Steen et al., 2006). Two association tracts were examined, the uncinate fasciculus and cingulum, together with three commissural tracts, the fornix, medial prefrontal projections through the genu of the corpus callosum (forceps minor) and medial parieto-occipital projections through the splenium.

Section snippets

Subjects

DTI data sets were collected as part of clinical imaging studies in healthy right-handed controls and patients meeting DSM-IV criteria for a diagnosis of major depressive disorder, age range 21–50 years. Following the development of the protocols, data from four healthy volunteers and four depressed patients were analyzed for reliability, blind to diagnosis. The two groups were included so that reliability data would be applicable to clinical comparisons of these populations. Normative data

Results

Mean values and between-subject coefficients of variation for FA, trace ADC and tract volumes are shown in Table 1. FA values were 3.4% higher in the left rostral section of the cingulum and 6.7% higher in the dorsal section than on the right, with no differences in the parahippocampal section. Rostral cingulum tract volume was also 30% higher on the left side. Uncinate fasciculus tract volume was 42% greater on the right than left. Finally, trace ADC was 4.5% higher in the left than right crus

Reliability

The present report provides tractography protocols for two association and three commissural tracts for which FA, trace ADC and volume measurements were obtained using DTI acquisition with six gradient diffusion directions. A variety of methods can be used for DTI data acquisition and analysis, such that reliability cannot be extrapolated to other methods, raters or study populations. Nevertheless, reliability is a major issue for quantitative studies and the data indicate that reliable FA and

Conclusions

The DTI-tractography protocols provide reliable methods to quantify trace ADC and FA in specific limbic paralimbic tracts, with reliable measures also obtained for the volumes of several structures. Hemispheric comparisons in the normative sample were consistent with reported variations in symmetry. Although the DTI field is advancing rapidly, these data indicate that DTI-tractography methods can be applied to clinical studies using readily accessible methods with appropriate training.

Acknowledgments

This study was supported by the Canadian Institute of Health Research, personnel support from the Alberta Heritage Foundation for Medical Research and a Wyeth/CIHR Fellowship and infrastructure support from the Canadian Foundation for Innovation. The assistance of Rawle Carter, Umme Yasmin, Colin Cassault and Param Bhardwaj with subject recruitment is gratefully acknowledged.

References (52)

  • HusterR.J. et al.

    Morphologic asymmetry of the human anterior cingulate cortex

    NeuroImage

    (2007)
  • JiangH. et al.

    DtiStudio: resource program for diffusion tensor computation and fiber bundle tracking

    Computer Methods and Programs in Biomedicine

    (2006)
  • JonesD.K. et al.

    The effect of filter size on VBM analyses of DT-MRI data

    NeuroImage

    (2005)
  • KanaanR.A. et al.

    Tract-specific anisotropy measurements in diffusion tensor imaging

    Psychiatry Research. Neuroimaging

    (2006)
  • KarlA. et al.

    A meta-analysis of structural brain abnormalities in PTSD

    Neuroscience and Biobehavioral Reviews

    (2006)
  • KubickiM. et al.

    A review of diffusion tensor imaging studies in schizophrenia

    Journal of Psychiatric Research

    (2007)
  • LandmanB.A. et al.

    Effects of diffusion weighting schemes on the reproducibility of DTI-derived fractional anisotropy, mean diffusivity, and principal eigenvector measurements at 1.5 T

    NeuroImage

    (2007)
  • MalykhinN.V. et al.

    Three-dimensional volumetric analysis and reconstruction of amygdala and hippocampal head, body and tail

    Psychiatry Research

    (2007)
  • MoriS. et al.

    Principles of diffusion tensor imaging and its applications to basic neuroscience research

    Neuron

    (2006)
  • SkareS. et al.

    Condition number as a measure of noise performance of diffusion tensor data acquisition schemes with MRI

    Journal of Magnetic Resonance

    (2000)
  • SmithS.M. et al.

    Tract-based spatial statistics: voxelwise analysis of multi-subject diffusion data

    NeuroImage

    (2006)
  • SnookL. et al.

    Voxel based versus region of interest analysis in diffusion tensor imaging of neurodevelopment

    NeuroImage

    (2007)
  • SullivanE.V. et al.

    Diffusion tensor imaging and aging

    Neuroscience and Biobehavioral Reviews

    (2006)
  • ArdekaniS. et al.

    Statistical representation of mean diffusivity and fractional anisotropy brain maps of normal subjects

    Journal of Magnetic Resonance Imaging

    (2006)
  • BeaulieuC.

    The basis of anisotropic water diffusion in the nervous system—a technical review

    NMR in Biomedicine

    (2002)
  • BuchsbaumM.S. et al.

    Diffusion tensor imaging of frontal lobe white matter tracts in schizophrenia

    Annals of General Psychiatry

    (2006)
  • Cited by (88)

    View all citing articles on Scopus
    View full text