Image Processing of Diffusion Tensor MRI Data - Advance Image Processing in Magnetic Resonance Imaging - page 437

FIGURE 13.14 Tensor-warped distance map: this contour map shows metric distance
from an initial point located in the posterolateral part of the corpus callosum. The image
is a slice through a 3-D distance map, at the level of the initial point. The regions where
neighboring contours are widely separated indicate low metric distance, or high connec-
tivity. (From O'Donnell, L., Haker, S., and Westin, C.-F. (2002). New approaches to
estimation of white matter connectivity in diffusion tensor MRI: Elliptic PDEs and geo-
desics in a tensor-warped space. in Dohi, T. and Kikinis, R. (Eds.). Medical Image
Computing and Computer-Assisted Intervention (MICCAI) . pp. 459-466, Tokyo, Japan.).

Connectivity measure: We would like to quantify path quality, but both

long and short paths occur in the brain (see Figure 13.1 ). So we cannot

rank paths by Euclidean or metric length alone. We choose to use one
length to normalize the other, which is one way to address this issue.

By comparing the geodesic path length to the Euclidean length of the same
path, we produce a measure of the degree of connectivity between any two points.
We compute the ratio of Euclidean path length to geodesic path length for all

paths outward from the initial point. Figure 13.15 displays the connectivity

measure as calculated for the distance map shown in Figure 13.14.

13.8 CONCLUSION

In this chapter, we first gave background information on diffusion and diffusion
tensor MRI, then presented tensor shape measures that quantify anisotropy or the
lack thereof. Next, we described visualization techniques for diffusion tensor fields.