Purpose: In functional MRI (fMRI), faster sampling of data canprovide richer temporal information and increase temporaldegrees of freedom. However, acceleration is generally per-formed on a volume-by-volume basis, without consideration ofthe intrinsic spatio-temporal data structure. We present a novelmethod for accelerating fMRI data acquisition, k-t FASTER(FMRI Accelerated in Space-time via Truncation of EffectiveRank), which exploits the low-rank structure of fMRI data.Theory and Methods: Using matrix completion, 4.27 retro-spectively and prospectively under-sampled data were recon-structed (coil-independently) using an iterative nonlinearalgorithm, and compared with several different reconstructionstrategies. Matrix reconstruction error was evaluated; a dualregression analysis was performed to determine fidelity ofrecovered fMRI resting state networks (RSNs).Results: The retrospective sampling data showed that k-tFASTER produced the lowest error, approximately 3–4%, andthe highest quality RSNs. These results were validated in pro-spectively under-sampled experiments, with k-t FASTER pro-ducing better identification of RSNs than fully sampledacquisitions of the same duration.Conclusion: With k-t FASTER, incoherently under-sampledfMRI data can be robustly recovered using only rank con-straints. This technique can be used to improve the speed offMRI sampling, particularly for multivariate analyses such astemporal independent component analysis.
展开▼
