A Least Difference Block Sharing (LDBS) Method for Optimizing the View-sharing iblocks-DTI

Motivation: View-sharing iblocks-DTI (VSiblocks-DTI) can substantially reduce the long scan time of iblocks-DTI while providing accurate DTI tensor calculations. However, its neighbor sharing method may limit its performance when using a randomized ordering of diffusion directions or small imaging matrix. Goal(s): This work aims to optimize the sharing method for VSiblocks-DTI. Approach: The least difference block sharing (LDBS) method was proposed and evaluated under different conditions. Results: The proposed LDBS method provided more accurate DTI tensor calculations than the previous neighbor sharing method under six different conditions, demonstrating its robustness to provide accurate DTI tensor calculation for VSiblocks-DTI. Impact: This study proposes a least difference block sharing (LDBS) method for optimizing view-sharing iblocks-DTI. It alleviates the limitation of previous sharing method on the ordering of diffusion directions and shows robust and accurate DTI tensor calculation with different matrix sizes.