Severe variations in magnetic susceptibility are observed around the human heart. Especially, the multitude of interfaces between bone/tissue and air create spatial B₀ variations in the heart. These B₀ inhomogeneity are modulated by the heart and respiratory cycles and result in both strong spatial distortion artifacts and signal drop out in cardiac MRI. The conventional correction of B₀ inhomogeneity (so-called B₀ shimming) relies on spherical harmonic (SH) shapes [1] and static first or first and second order shapes are routinely applied in clinical routine. The benefits of multi-coil (MC) shimming, shimming with individual placed magnetic coils to modify the B₀ field, have demonstrated recently in the human brain [2,3]. This study entails a comprehensive, theoretical analysis of B₀ shimming capabilities in the human heart. Three-dimensional B₀ distributions over the in vivo human heart are addressed with various SH and MC shimming approach in a static, dynamic and a hybrid fashion.To show the benefit of the different shimming approaches we look at the phase evolution over heartbeat cycle to get the B₀ difference at every point. These three-dimensional B₀ field maps (Fig. 1) were obtained in five healthy volunteers with gradient echo MRI at four different echo times (TE = 2.00/3.17/4.37/5.57 ms) on a 3T MAGNETOM Prisma (Siemens AG, Erlangen, Germany). B₀ field maps were decomposed in (orthogonal) SH and (non-orthogonal) MC basis shapes with least squares optimization as described earlier [1] and applied at reversed polarity to simulate the B₀ shim capability. We introduced a quality parameter Q_G to analyze the quality of different shim methods. Q_G is calculated by the sum of the absolute difference of the field after shimming divided by the sum of the absolute difference of the field before shimming. For the SH shimming, boundary conditions were set to unlimited dynamic range. The MC setup in the simulation was a 60-channel MC setup (Fig. 2) on an elliptical cylinder (axes dimensions x/y/z = 520/300/555 mm) that is large enough to host the thorax of a typical male adult. Individual coil elements (diameter 75-155 mm) were simulated at 100 turns and a 1 A maximum current was assumed. The difference between the individual shim methods were 2nd and 3rd order SH shim (SH-2nd, SH-3rd), MC shim (MC), static shim (s), dynamic shim on heart cycle (dH), dynamic shim on heart cycle and slices (dHS), and for the dynamic SH shim methods, all parameter dynamic (A), only 1st order parameters were dynamically modified (1).The Q_G values of the aforementioned shim methods for all volunteers are shown in Fig. 3. The error of the simulated shim, represented by Q_G, is below 0.001%. The improvement of the average over the five volunteers for every individual shim method relative to SH-2nd-s (clinical standard) is shown in Fig. 4.The error of the improvement of Q_G value is below 0.3%. This demonstrates clearly the superiority of the dynamic shimming over static shimming. The third order static SH shimming (SH-3rd-s) is 29.5±0.3% better than the clinical standard, second order SH shimming (SH-2nd-s). The dynamic SH shimming (SH-3rd-dHS-A) with 56.5±0.2% improvement is clearly superior with respect to the static shimming of the same SH order. The comparison of MC shimming versus SH shimming demonstrates obvious advantage on the MC methods. The static MC shim (MC-s) with 43.8±0.2% is on 14.3 percentage points higher than the static SH shim (SH-3rd-s). Also the dynamic MC shim (MC-dHS), called DYNAMITE, (60.8±0.2%) outperforms dynamic third order SH shim (SH-3rd-dHS-A, 56.5±0.2%) and provides the best B₀ homogeneity of all methods considered in this study.The results of the study show that, as expected, the global standard static SH shimming are generally inferior in comparison with the results of specifically tailored and customized shimming methods based on dynamic and MC approaches. An essential improvement of B₀ homogeneity in the heart tissue can be gained by using dynamic shims with SH and, especially, with MC-basis. The most impressive improvements could be achieved with the most comprehensive (dynamic and MC) DYNAMITE approach. For the sake of clinically practicable and robust procedure, the analysis also demonstrates that a lot can be gained with partial dynamic B₀ shimming, i.e. a hybrid approach.