Cardiac pulsation is a physiological confound of functional magnetic resonance imaging (fMRI) time-series that introduces spurious signal fluctuations in proximity to blood vessels. fMRI alone is not sufficiently fast to resolve cardiac pulsation. Depending on the ratio between the instantaneous heart-rate and the acquisition sampling frequency (1/TR, with TR being the repetition time), the cardiac signal may alias into the frequency band of neural activation. The introduction of simultaneous multi-slice (SMS) imaging has significantly reduced the chances of cardiac aliasing. However, the necessity of covering the entire brain at high spatial resolution restrain the shortest TR to just over 0.5 seconds, which is in turn not sufficiently short to resolve cardiac pulsation beyond 60 beats per minute. Recently, hyper-sampling of the fMRI time-series has been introduced to overcome this issue. While each anatomical location is sampled every TR seconds, the time between consecutive excitations is shorter and thus adequate to resolve cardiac pulsation. In this study, we show that it is feasible to temporally and spatially resolve cardiac waveforms at each voxel location by combining a dedicated hyper-sampling decomposition scheme with SMS. We developed the technique on 774 healthy subjects selected from the Human Connectome Project (HCP) and validated the technology against the RETROICOR method. The proposed approach, which we name Data-driven WHOle-brain CArdiac signal REgression from highly accelerated simultaneous multi-Slice fMRI acquisitions (WHOCARES), is fully data-driven, does not make specific assumptions on cardiac pulsatility, and is independent from external physiological recordings so that the retrospective correction of fMRI data becomes possible when such measurements are not available. WHOCARES is freely available at https://github.com/gferrazzi/WHOCARES.