Low b‐value diffusion tensor imaging for measuring pseudorandom flow of cerebrospinal fluid

Purpose Cerebrospinal fluid (CSF) plays an important role in the clearance system of the brain. Recently, low b ‐value diffusion tensor imaging (low‐ b DTI) has been reported to be useful in the observation of CSF flow; however, the precise flow property observed by low‐ b DTI has not been fully investigated. Accordingly, a mathematical framework of low‐ b DTI is proposed for investigating CSF and clarifying its pseudorandom flow. Theory The framework will show that the limit of the diffusion tensor as b ‐value decreases to zero approximately represents the covariance of the velocity distribution of the CSF’s pseudorandom flow. Methods The low b ‐value diffusion tensor (DT L ) of whole‐brain CSF was obtained using diffusion‐weighted echo‐planar imaging. Seven healthy volunteers were scanned for intersubject analysis; three of the volunteers was consecutively scanned for repeatability analysis. Obtained DT L was visually assessed by ellipsoid‐representation map and was statistically evaluated by calculating mean diffusivity (MD) and fractional anisotropy (FA) in regions of interest (ROIs) representing intensive pseudorandom flow. Results Obtained DT L consistently shows large and anisotropic diffusivity in some segments of CSF, typically the ROIs around the foramen of Monro, the aqueduct, the prepontine cistern, the middle cerebral artery, and the Sylvian fissure throughout the study. The statistical analysis shows high repeatability and consistently high MD and FA in all the ROIs for all the volunteers. Conclusion From the viewpoint of the proposed framework, the high and anisotropic DT L in the ROIs indicates large covariance of velocity distribution, which represents intensive pseudorandom flows of CSF.