Post-processing of diffusion-weighted MR data lowers the accuracy of the weighted linear least squares estimator

For clinically relevant SNR values (SNR>2), the weighted linear least squares estimator is theoretically expected to be as accurate as advanced estimators that incorporate prior knowledge of the data distribution in the estimation of DTI/DKI model parameters. However, one must bear in mind that the high accuracy vanishes if magnitude operations are applied prior to model fitting. After magnitude operations, which are generally included in the diffusion MRI processing pipeline, the prior knowledge of the noise parameter becomes a must in order to define an unbiased estimator.

This abstract and the presentation materials are available to members only; a login is required.

Join Here