ASL perfusion fMRI has much less neurovascular effects than BOLD fMRI,
but its application in time-series analysis is still depreciated due to the low
signal-to-noise-ratio (SNR). Robust principal component analysis (RPCA) decomposing
the original data into a smoothly varying low-rank component and a residual
component with sparse signal. In this study, we used RPCA to denoise ASL MRI.
Our results showed that RPCA can markedly increase the sensitivity of ASL
MRI-based functional connectivity analysis.