Diffusion MRI techniques
Diffusion-weighted MRI (DW-MRI) is a type of MRI that measures water diffusion in living tissue. A DW-MRI scan can be made directionally sensitive, so that only diffusion in a particular direction affects the measurement. Since water in the brain tends to diffuse more easily along white matter axons than across them, the local orientation of bundles of axons may be inferred by the directional dependence of the diffusion signal at a given point. This orientation information and other values computed from the diffusion signal can be useful to brain scientists studying the effects of disease on the structure of the brain.
Our group processes and visualizes diffusion MRI data in the Diffusion MRI project. Although a complete description of the physics of DW-MRI and the plethora of processing techniques that have been developed is beyond the scope of this article (see Mori's review paper for a clear explanation), here we present a high-level overview of the information relevant to our work.
The Diffusion Signal
The direction along which a single DW-MRI scan measures diffusion is aligned with the gradient of a magnetic field, which we call the b vector. Water diffusion in the opposite direction of the b vector affects the diffusion signal in the same way as water diffusion along the b vector; therefore for a given vector g we consider g and -g to be equivalent. Imaging parameters related to the timing of various magnetic pulses are summarized in the so-called b value, which is measured in units of s/mm2.
The space of imaging parameters is therefore naturally described with a three-dimensional spherical coordinate system: the angular position of a point corresponds to the b vector, and the distance to the origin is the b value. This space is called Q space. A point in Q space corresponds to a specific set of imaging parameters and therefore a single image acquired with these parameters is natuarlly identified with that point. The origin in Q space corresponds to an "unweighted" image, with b=0s/mm2. Typical DW-MRI protocols acquire at least one unweighted image and a collection of images at different points in Q space, generally between six and eighty points distributed equally across a hemisphere at a fixed b value. Again, the points X and -X in Q space are equivalent; hence the hemisphere. A typical b value is 1000s/mm2.
At any given b value, the diffusion signal is inversely proportional to the extent of water diffusion in the gradient direction. By "extent of water diffusion" we mean the average distance that a water molecule travels in a given amount of time. The amount of time during which this diffusion is measured is proportional to b. Lower values (typically no lower than 100s/mm2) measure diffusion over smaller intervals of time and are therefore noisier but also more sensitive to tiny features in the tissue; higher values (as high as 10000s/mm2) measure diffusion over larger intervals of time and are therefore less noisy but also less sensitive to small details.
MRI scanners are rated by the strength of the magnetic field they can generate, which is measured in Teslas, or T. 1.5T scanners are common in clinical settings, 3T and 7T are becoming more common, and 11T and stronger machines are sometimes used in research with animals or ex vivo tissue. The stronger the magnet, the less noisy its image (typically), but the more expensive it is to operate. Magnets of 7T and above have also been reported to have unusual physiological effects on subjects, including dizziness, euphoria, and amnesia.
References
- S. Mori and J. Zhang. Principles of Diffusion Tensor Imaging and Its Applications to Basic Neuroscience Research. Neuron 51, pp. 527--539, September 2006.
- P.J. Basser and D.K. Jones. Diffusion-Tensor MRI: Theory, Experimental Design, and Data Analysis --- A Technical Review. NMR in Biomedicine 15, pp. 456--467, 2002.