Formulation and Simulation of Low b-value DTI of Pseudorandom Flow in CSF

Motivation: While low b-value (low-b) DTI is used to measure pseudorandom flow in CSF, spatial relationship of DT among multiple voxels has not been clearly shown. Goal(s): To reveal spatial features of pseudorandom flow with low-b DTI. Approach: A formula expressing the relationship between low-b DT map and flow-velocity vector-field was derived, and simulation of typical pseudorandom flows was performed using the formula with computational fluid dynamics and deep learning. Results: The relationship was successfully formulated for spatially gradually-varying flow-velocity, and the simulation revealed remarkable features such that flow direction can be inferred from DT map for some pseudorandom flows. Impact: Formulation of low-b DT map in relation to flow-velocity vector-field was derived, and simulation using the formula and computational fluid dynamics revealed remarkable features of pseudorandom flows, suggesting that it could help in understanding pseudorandom flow measured with low-b DTI.