1Electrical Engineering, Stanford University, Fremont, CA, United States, 2FMRIB Centre, University of Oxford, Oxford, United Kingdom
Functional MRI has been slow to benefit from data acceleration techniques based on non-linear image reconstruction. We present a comparison of two non-linear image reconstruction methods based on sparsity and low-rank models of FMRI data. k-t FOCUSS uses an asymptotic L1 minimization program to solve for a sparse x-f reconstruction. In contrast, k-t FASTER solves for a spatio-temporally low-rank reconstruction using an iterative hard thresholding and matrix shrinkage algorithm, without requiring a pre-specified basis. We applied each algorithm to incoherently sampled FMRI data and demonstrate that the strict rank-constraint method outperforms spectral- and Karhunen-Loeve Transform (KLT)-sparsity across different metrics.