Image Processing of Diffusion Tensor MRI Data - Advance Image Processing in Magnetic Resonance Imaging - page 419

FIGURE 13.1 Selected fiber traces (tractography) from a DT-MRI data set, shown in a

sagittal view. Note the presence of both long and short paths, and varying degrees of
curvature. These factors complicate automated extraction of tracts from diffusion MRI data.
This image was created with the 3-D Slicer DT-MRI visualization tool. (From Talos, I.-F.,
O'Donnell, L., Westin, C.-F., Warfield, S.K., Wells, W.M., Yoo, S.-S., Panych, L.P., Golby,
A., Mamata, H., Maier, S.S., Ratiu, P., Guttmann, C.R. , Black, P.M., Jolesz, F.A., and
Kikinis, R. (2003). Diffusion tensor and functional MRI fusion with anatomical MRI for
image guided neurosurgery. in Conference on Medical Image Computing and Computer-

Assisted Intervention (MICCAI)

. pp. 407-415. Toronto, Canada.)

One mathematical representation of this three-dimensional (3-D) diffusion
pattern is the diffusion tensor, a 3 3 symmetric, positive definite matrix. In

diffusion tensor MRI (DT-MRI) imagery of the brain, the eigensystem of the
diffusion tensor gives a local coordinate system that approximates the local neural
structure. The major eigenvector gives the direction of greatest diffusion (the
most probable fiber direction). The eigenvalues of the diffusion tensor represent

×

the diffusion coefficients in the principal directions of diffusion. Figure 13.3

shows the eigenvectors of a diffusion tensor in two dimensions, scaled by the

FIGURE 13.2

Idealized diagram showing the effect of axons on water diffusion, as
measured by diffusion MRI.