Unlike in the brain, fiber crossing in the myocardium of the heart is controversial despite some reports on the heart of animals [1] [2], and even more so when there is no consensus on the definition itself of fiber in the heart. Nevertheless, with a spatial resolution of the order of millimeter of clinical MRI systems, we can by the abuse of language use the term fiber to designate a group of myocytes having a privileged orientation in the heart. The purpose of this study is to investigate the fiber configuration complexity in a voxel, such as fiber crossing, in the human heart. The real data concerns an ex vivo human heart. It was acquired using a conventional diffusion tensor imaging (DTI) protocol on a Siemens 1.5T Magnetom Sonata with the following settings: TE=73ms, TR=6400ms, FOV=256×256mm, slice thickness=2 mm, slice spacing=2 mm, number of slices=54, image size=128×128, diffusion sensitivity b-value=1000s/mm², gradient directions=64 and a single T2-weighted (b-value=0) image. The heart was located in a plastic container and fixed by hydrophilic gel to maintain a diastolic shape. This setup has a low dielectric effect and eliminates unwanted susceptibility artifacts near the boundaries of the heart. To reconstruct the maps of fiber configurations, we employed two different techniques in high angular resolution diffusion imaging (HARDI): the constrained spherical deconvolution (CSD) [3] and the analytical q-ball imaging (AQBI) [4][5], which are commonly used for brain, but until now never applied to heart. Strictly speaking, the CSD generates the fiber orientation distribution (FOD) map, and the AQBI the orientation distribution function (ODF) map. For simplicity, we call both of them the ODF. The ODF maps reconstructed using CSD and AQBI are shown in Fig. 1 (21^th slice). The zones marked by the red rectangle contain the voxels with complex fiber configurations. To show more clearly the complex fiber configurations revealed by CSD and AQBI, in Fig. 2(a) to (c) are visualized their zoomed versions. The ODFs obtained from both methods show more or less the existence of intravoxel complex fiber configurations in the middle of the red rectangular region. The fiber crossings are more pronounced by CSD, but with AQBI, we can still observe that it does not concern simple single fiber configurations. As an illustration, a clearly crossing fiber (reconstructed by CSD) in the voxel circled by the black box is shown in Fig. 2(c). Other fiber crossing voxels were also found near this voxel. By exploring the neighboring slices around the 21th slice, at the same region, we also observed the existence of complex fiber configurations, as shown in Fig. 3 (ODFs obtained using CSD).On the same DTI datasets, two different reconstruction methods led to the same observation that the myocardium of the human heart presents complex fiber configurations. This result is not due to the partial volume since the voxel is in the middle of the myocardium. It is not due to noise either since the fiber crossing occurs also at exactly the same position in the neighboring slices around the current slice. However, the two methods did not yield the same degree of resolving the fiber crossing. It would then be necessary to validate such findings either via simulation or via some physical measurements like polarized light imaging.This work has investigated for the first time the fiber configuration complexity of the human heart. The results consistently demonstrated that the human myocardium exhibits clearly some complex fiber configurations in a voxel such as fiber crossing when using both CSD and AQBI reconstruction techniques, which is particularly interesting since the results are obtained on conventional DTI data (so shorter acquisition time). Since the two different reconstruction methods led to slightly different complex fiber configuration results, further study both on data acquisition schemes such as HARDI and on the reconstruction of fiber configurations would be interesting for confirming these preliminary findings or searching fiber configuration information other than fiber orientation.