Fig. 4.

Different undersampling patterns and corresponding point spread function (PSF). a Binary masks of k_y-t-space of phantom data at 0% heart rate variability (left) and 10% heart rate variability (middle). On the right is an ideal random sampling pattern with the same undersampling factor. b PSF after inverse fast fourier transform (FFT) in both spatial dimensions as function of time. The sampling pattern without heart rate variability (left) creates two side peaks, whereas the ideal random undersampling pattern (right) creates incoherent noise next to the main peak. A similar incoherent pattern is found with 10% heart rate variability (middle)