PURPOSE: Although gated first-pass contrast-enhanced sequences are the clinical standard for cardiovascular MR perfusion, some patient conditions necessitate using ungated steady-state sequences. However, through-plane cardiac motion and blood flow into the left ventricle can disrupt the magnetization steady state of the tissue, and perfusion quantification based on a steady-state assumption will contain errors. The tissue magnetization steady-state disruption can be eliminated with a proposed sequence modification that simultaneously excites transition bands with no change in the sequence resolution or timing parameters. THEORY AND METHODS: The proposed sequence modification simultaneously excites two transition bands adjacent to the imaged region. The transition bands experience the same consistent excitation history as the imaged slices. Thus, any tissue that moves into an imaged slice location from a transition band location will not disrupt the magnetization steady state. Gradient dephasing and radiofrequency spoiling are used to null the transition band signal so that it does not contribute to the reconstructed images. Transition bands were added to a two-dimensional ungated steady-state radial FLASH (fast low-angle shot) sequence with simultaneous multiband imaging on a PRISMA 3T MRI scanner. Phantom, normal canine, and human subject data are presented. RESULTS: Transition bands reduce the amount of tissue magnetization disruption to the steady state without adding artifacts to the imaged slices. Myocardial blood flow maps from a selected normal canine study and a normal human subject show good uniformity and consistency to literature values for all slices. Perfusion estimates with the proposed method also demonstrate good consistency with saturation-recovery methods commonly used for myocardial perfusion imaging. CONCLUSION: We have demonstrated that the proposed transition bands can reduce quantification errors resulting from blood flow into the left ventricle and through-plane cardiac and respiratory motion. There is no loss of image-acquisition efficiency, and temporal resolution is unchanged with this technique.