A method is disclosed for generating a motion-corrected PET image of an examination area in a combined MR-PET system. In an embodiment, the method includes recording PET events from the examination area in a first recording time frame; recording a number of MR images of the examination area in at least the first recording time frame; calculating an at least two-dimensional movement information of the examination area on the basis of the number of MR images, wherein the movement information describes the movement information of the examination area during the first recording time frame, and determining the motion-corrected PET image from the PET events using the calculated movement information.