Optimal Encoding Points for Diffusion Spectrum Imaging: Using Bi-Gaussian Extrapolation
Yeh C, Cho K, Li Y, Lin H, Lin C
Institute of Radiological Sciences, National Yang-Ming University
Diffusion spectrum imaging (DSI) is capable of defining complex orientational coherence within MR voxel. To establish the adequately sampling points over q-space, however, usually results in a long scan time. Additionally, diffusion q-value is generally limited by the available gradient strength in clinical machine and therefore produces the truncation error in 3D probability density function (PDF) reconstruction. To acquire the complex fiber orientations while avoiding these drawbacks, this study reduced DSI encoding points as well as maximum q-value by applying a bi-Gaussian curve fitting over the q-space. Both Monte-Carlo simulation and MR experiments showed that this method can decrease DSI acquisition time while preserving the PDF patterns and primary orientations.