Introduction

Experimental evidence has shown that water diffusion is anisotropic in organized tissues such as muscles [1,2] and brain white matter.[3] Since the mid 1990s, the quantitative description of this anisotropy with diffusion tensor imaging (DTI) has become well established in the research environment, and its first applications in the clinic are now being reported.[4,5] For example, DTI is presently being explored as a research tool to study brain development,[6–8] multiple sclerosis,[9,10] amyotrophic lateral sclerosis,[11] stroke,[12–14] schizophrenia,[15,16] and reading disability.[17] Based on fiber orientation information obtained from DTI, it has been also shown that in vivo fiber tracking is possible.[18–33] In order to improve the utilization of this promising technology, it is important to understand the basis of the anisotropy contrast in DTI and the limitations imposed by using a macroscopic technique to visualize microscopic axonal structures. In this chapter, basic principles of the DTI-based tract reconstruction and its capability and limitations will be discussed.

Isotropic and anisotropic diffusion

It has been known that MRI can measure molecular diffusion constant. One of the unique and important features of the diffusion measurement by MR is that it always detects molecular movement along one predetermined axis (Fig. 5.1), which is determined by the resultant orientation of applied magnetic field gradients. Every MR scanner is equipped with three orthogonal x-, y-, and z-gradients. By combining these three-axis gradients, diffusion along any arbitrary axis can be measured. For example, if equal strength x- and y- gradients are applied simultaneously, diffusion along 45° from the x- and y-axes can be measured. The orientation of the diffusion measurement is not important if we are interested in freely diffusing water because the results are independent of the measurement orientations. Such orientation-independent diffusion is called isotropic diffusion (e.g., the lower compartment of Fig. 5.1A).