Share this article on:

Diffusion Tensor Imaging and Tractography: Have They Come of Age?

Sundgren, Pia C MD, PhD

Journal of Neuro-Ophthalmology: June 2009 - Volume 29 - Issue 2 - pp 93-95
doi: 10.1097/WNO.0b013e3181a594a1

Department of Radiology, University of Michigan Health Systems, Ann Arbor, Michigan.

Address correspondence to Pia C. Sundgren, MD, PhD, Professor, Department of Radiology, University of Michigan Health Systems, 1500 East Medical Center Drive, Ann Arbor, MI 48109. E-mail:

In this issue of the Journal of Neuro-Ophthalmology, Lanyon et al (1) have demonstrated that the V5/medial temporal (MT) complex defined by functional MRI can be activated by direct thalamic or collicular input without involving V1 (primary visual, calcarine, or striate cortex). The authors have shown this connection by means of diffusion tensor imaging (DTI) and tractography. Although interesting and supported by previous findings in animals, their findings are dependent on imaging methodology that produces beautiful pictures but has substantial technical limitations.

DTI is an MRI-based methodology originally described in 1994 by Basser et al (2) that has become widely used in clinical practice and in brain research. DTI allows direct in vivo examination of some aspects of tissue microstructure and yields quantitative measures reflecting the integrity of white matter fiber tracts by taking advantage of the intrinsic properties of directionality of water diffusion in human brain tissue.

The diffusion of water molecules is characterized by Brownian motion. When water molecules are unconstrained, the direction of motion of a given molecule is random. The displacements of water molecules over time are described by a Gaussian distribution. The diffusion is called isotropic when the motion is equal and unconstrained in all directions. However, the microstructure of brain tissue forms physical boundaries that limit the Brownian motion of water molecules, resulting in a restriction of the total amount of diffusion. In microstructures such as the white matter fibers, the diffusion of water molecules will be relatively more restricted in a direction perpendicular to the microstructural boundaries than parallel to them. Such diffusion is termed “anisotropic.”

Anisotropic diffusion is characterized not by a single coefficient, but by a second-order symmetric tensor of six unique elements or diffusivities requiring multiple measurements for complete determination. Once these six diffusion coefficients have been obtained, the degree of mobility of water protons in the system can be determined with an average principal diffusivity. The degree of anisotropy of this mobility can be found by calculating anisotropy indices such as fractional anisotropy (FA). FA is a measure of the portion of the magnitude of the diffusion tensor due to anisotropy and gives information about tissue organization, degree of myelination, and water mobility. The most widely used measure of anisotropy in the DTI literature, FA has a value that varies from zero, in the case of isotropic diffusion, to a maximum of 1, indicating perfectly linear diffusion occurring only along the primary eigenvector. The information can be presented in two dimensions by a widely accepted color scheme originally proposed by Pajevic and Pierpaoli in 1999 (3). It incorporates the anisotropy and directional information and attributes a specific color to each of the different orientations of the axonal fibers.

Almost a decade ago, several research groups began to use the three-dimensional (3D) information present in DTI to create 3D virtual trajectories in a method called “tractography.” The objective of DTI fiber tracking is to determine intervoxel connectivity on the basis of the anisotropic diffusion of water. In tractography, the local fiber orientation in each voxel is determined via DTI. Voxels are then connected to each other, starting at a seed point and propagating a tract by mathematically connecting the adjacent voxels based on information gleaned from the magnitude and directionality of diffusion anisotropy.

Hundreds of tractography studies of white matter connectivity have now been performed in healthy and diseased brain. The technique has become a particularly useful clinical tool in presurgical planning. It provides information about whether white matter tracts adjacent to a brain tumor are merely displaced or invaded (Fig. 1). DTI and tractography have also been popular in anatomic connectivity studies of healthy and diseased individuals and in the assessment of the immature brain in neonates, congenital brain malformations, demyelinating diseases, and epilepsy. For example, DTI and fiber-tracking studies have demonstrated abnormal white matter tracts within the cerebellum and cerebrum in holoprosencephaly that were not apparent on conventional MRI. DTI may prove invaluable in identifying potential epileptogenic foci as well as more optimally defining the extent of the lesion proposed for surgical resection.

There are several limitations to DTI and tractography. Background noise, patient movement, and distortion from imaging artifacts produce uncertainty. Accuracy of the constructed fiber tract is limited by the information contained in the diffusion tensor and in the different methods of constructing the tracts. DTI tractography cannot distinguish antegrade from retrograde information flow along a fiber pathway. Another limitation is the assumption that diffusion has Gaussian characteristics, which is not correct when diffusion is restricted. There is even experimental evidence that diffusion in normal white matter is non-Gaussian at high b-values (which represent the overall diffusion in an experiment). Even if the problems with non-Gaussian diffusion are solved, there is as yet no clinically applicable method of modeling non-Gaussian diffusion.

An important user limitation of DTI/tractography is that when selecting a specific fiber system, one has to know beforehand the anatomy of the white matter tracts. Furthermore, tractography is a mathematical estimation of white matter tracts rather than an anatomic depiction. Axonal fibers can appear disproportionately large because of higher FA. Further research is needed to reduce or eliminate the confounding problem created by crossing fibers and partial volume effects. Spatial resolution must improve to allow visualization of small fiber tracts.

Because these colorful images of the white matter bundles are so beautiful, it is easy to be swayed. Before doing so, we will need better methods to deal with non-Gaussian diffusion, crossing fibers, and partial volume effects. Only then will DTI and tractography provide robust and clinically useful information.

To learn more about the utility, limitations, and promise of DTI and tractography, consult excellent current reviews by Mori and van Zijl (4) and Mukherjee et al (5,6).

Back to Top | Article Outline


1. Lanyon LJ, Giaschi D, Young SA, et al. Combined functional MRI and diffusion tensor imaging analysis of visual motion pathways. J Neuroophthalmol 2009;29:96-103.
2. Basser PJ, Mattiello J, LeBihan D. MR diffusion tensor spectroscopy and imaging. Biophys J 1994;66:259-267.
3. Pajevic S, Pierpaoli C. Color schemes to represent the orientation of anisotropic tissues from diffusion tensor data: application to white matter fiber tract mapping in the human brain. Magn Reson Med 1999;42:526-540.
4. Mori S, van Zijl PC. Fiber tracking: principles and strategies a technical review. NMR Biomed 2002;15:468-480.
5. Mukherjee P, Berman JI, Chung SW, et al. Henry diffusion tensor MR imaging and fiber tractography: theoretic underpinnings. AJNR Am J Neuroradiol 2008;29:632-641,
6. Mukherjee P, Chung SW, Berman JI, et al. Diffusion tensor MR imaging and fiber tractography: technical considerations. AJNR Am J Neuroradiol 2008;29:843-852.
© 2009 Lippincott Williams & Wilkins, Inc.