Diffusion Magnetic Resonance Imaging in Neuroimaging

Abstract

Diffusion is one of the basic phenomena in nature and can be quantified with magnetic resonance imaging (MRI) in terms of signal attenuation. Diffusion MRI portrays the geometry of microstructures based on the fact that molecular diffusion varies in different directions and that the geometry can be faithfully reflected by the probability density function (PDF) of the molecular diffusion. The theory of MRI shows that the PDF can be obtained by Fourier transform of the MRI signal in the acquisition space (q-space). A technique that stemmed from this theory, called diffusion spectrum imaging (DSI), is capable of resolving complex axonal fiber geometry at each location in the brain, which consequently affords more accurate results of tractography and white matter integrity. Implementation of DSI, however, requires stringent hardware performance and long scan times, thus hampering its clinical application. To overcome these problems, many novel approaches have been proposed using fewer sampling numbers or lower gradient strengths. The advances in diffusion MRI open a door to the investigation of functional connectivity and white matter abnormalities in neuropsychiatric disorders. A future direction for diffusion MRI is to make this advanced technique more clinically feasible and to discover its clinical values in the diagnosis or treatment of brain disease.