Effective reduction of CSF partial volume effect in DTI by acquiring additional DWIs with smaller β-value

Abstract

One of the major limitations of DTI is its vulnerability to CSF contamination. To remove such effect, FLAIR preparation and the two-compartment tensor model have been proposed. However, they require substantially increased scan time (and lead to reduced brain tissue SNR in the former case). A simple and effective approach is proposed in the current study. Diffusion tensor (DT) was computed from diffusion-weighted images acquired with b-value=500 and 1000s/mm2. Experiments were performed in volunteers and rodents to compare this approach and conventional DTI, demonstrating effective reduction of CSF contamination.