Skip to Main Content
Diffusion Tensor Imaging (DTI) is the most used paradigm among the Diffusion Weighted Magnetic Resonance Imaging techniques, due to its inner simplicity and huge application potential. Least Squares has become the standard technique to estimate the Diffusion Tensor (DT) from Diffusion Weighted Images. This approach is known to be optimal when the acquired data follows Gaussian, Rician or non-central chi distributions. In this paper we study the effect of different acquisition schemes over the quality of the DT estimation. The following cases are considered: single coil, multiple coil fully sampled, and accelerated subsampled multiple coil acquisitions with SENSE and GRAPPA reconstructions. Reconstructed subsampled data will show an acceleration in the acquisition process, but also greater variance and bias in the DT estimation.