This article will briefly review the history of diffusion physics and diffusion imaging beginning with Brown's initial observations of the diffusion phenomenon in the early part of the 19th century, through the development of diffusion-weighted imaging, and diffusion tensor imaging. The basic principles and limitations of diffusion tensor imaging are discussed. We conclude by reflecting on some of the open interpretative questions in diffusion imaging.

Diffusion tensor magnetic resonance imaging is an imaging modality that measures the diffusion properties of water molecules in tissues noninvasively. Water diffusion is affected by tissue constituents, such as macromolecules, membranes, organelles, as well as by tissue microstructure, architecture, and organization. From the quantities measured with diffusion tensor-magnetic resonance imaging, one can infer information about brain tissue that cannot be obtained using conventional, compositional-based (eg, proton density and chemical spectroscopy), or relaxometry-based (eg, magnetization transfer, T1, T2, and T2*) magnetic resonance imaging methods. Understanding the relationship between a measured water diffusion pattern and the underlying histological features of the tissue, however, is not simple. In the absence of a robust and comprehensive model of water diffusivity in biological tissues, the biological interpretation of diffusion measurements relies on empirical evidence. Herein, the properties of several diffusion tensor-derived quantities are reviewed, together with the experimental evidence that helps clarify their relationship with the underlying properties of brain tissue.

White matter fiber pathways are key structural components of the brain and its functional organization. The limbic system carries a great deal of its anatomic connectivity via the cingulum bundle. By allowing the in vivo delineation of the stem of the major fiber pathway systems, diffusion tensor magnetic resonance imaging has opened a new window into the detailed structure of the white matter in health and disease. Topographic, biophysical, and volumetric information about fiber tracts will provide a more complete understanding of the brain. By appreciating its interconnections, the precise anatomical knowledge of the cingulum bundle will improve our understanding of the limbic system and may enable improvements in the assessment and treatment of neuropsychiatric disorders. In this study, the stem of the cingulum bundle was investigated and defined in terms of its trajectory, anisotropy, and volume, in four normal human subjects, using diffusion tensor imaging.

The raw diffusion tensor imaging data obtained after tensor calculations contain six tensor elements in each pixel. This unique data structure poses difficulties in visualizing and analyzing diffusion tensor imaging data. One of the most commonly used visualization techniques is the use of color-coded maps. This presents fiber orientation information as a mixture of three principal colors. These maps can reveal white matter substructures that may not be visible in conventional magnetic resonance imaging. By extending the fiber-orientation information into three-dimensional space, three-dimensional trajectories of white matter tracts can then be estimated. Once locations and trajectories of tracts of interest are identified, this technique allows for the utilization of tract-specific magnetic resonance analyses and/or macroscopic characterization of white matter anatomy. As an example, anatomical deformation of the white matter resultant of brain tumor is demonstrated. The potentials and limitations of the three-dimensional tract reconstruction techniques are also highlighted.

Diffusion-weighted imaging (DWI) represents a relatively novel magnetic resonance imaging (MRI) technique in which image contrast is related to differences in translational motion of water molecules within the tissue, rather than to differences in total water content. The rate of water motion is characterized by the apparent diffusion coefficient. In addition, full tensor DWI (diffusion tensor imaging, DT-MRI) samples the full diffusion tensor and therefore allows estimation of isotropic and anisotropic diffusion. The degree of anisotropic diffusion is thought to be determined by the local tissue characteristics or tissue architecture and can be quantified by DT-MRI. DWI has proven its clinical effectiveness in the early detection of acute cerebral ischemia. Multiple reports have discussed the value of DWI in a wide variety of other diseases of the central nervous system. DT-MRI appears to be especially promising in the evaluation of diseases that affect the integrity of white matter, in particular of white matter tracts. Herein, the current applications of DWI in clinical neurology are reviewed, with special attention to applications of DT-MRI.