Image artifact correction tools for brain DTI data: effects on test-retest reproducibility

Number of Sign. Lower Rep. Error Voxels Introduction This study investigated the effects of image co-registration on DTI derived scalar measurements (FA: fractional anisotropy, MD: mean, RD: radial and AD: axial diffusivity) and their across-session test/retest reproducibility. Specifically, we evaluated effects using the first vs averaged b0 volume for eddy current and motion correction in FSL [1]. This pipeline which used the averaged b0 was also compared to the automatic quality control tool DTIPrep Version 1.1.5 [2]. Additionally we tested effects of rotating the bmatrix [3] based on the 6 degrees and 12 degrees of Freedom (doF) correction performed by the FSL software FLIRT [4] and FNIRT [5].