There is a long-standing debate as to whether visual mental imagery relies entirely on symbolic (language-like) representations or also relies on depictive (picture-like) representations. We sought to discover whether visual mental imagery could evoke cortical activity with precise visual field topography (retinotopy). Participants received three conditions: the perception condition consisted of a standard retinotopic mapping procedure, where two flickering checkerboard wedges rotated around a central fixation point. The imagery and attention conditions consisted of the same stimulus, but only the outer arcs of the wedges were visible. During imagery, participants mentally reproduced the stimulus wedges, using the stimulus arcs as a guide. The attention condition required either distributed attention or focused attention to where the stimulus wedges would have been. Event-related analysis revealed that the imagery (greater than either form of attention) retinotopic maps were similar to the perception maps. Moreover, blocked analysis revealed similar perception and imagery effects in human motion processing region MT+. These results support the depictive view of visual mental imagery.