Using video stimuli and fMRI, we found retinotopic responses in parietal, temporal, and frontal cortex in humans (Sereno et al, SFN 2004). Here, we study the effect of stimulus content vs. attention on those maps. We used structured motion (e.g., point-light biological motion) in a polar angle mapping paradigm. In contrast to standard mapping, we also filled the background with either scrambled versions of the structured motion (a subtle stimulus contrast), or with additional copies of it (no stimulus contrast). Subjects either attended to the rotating wedge (task performed on wedge contents) or centrally (unrelated attentionally demanding foveal task). Each subject was scanned in 4-6 runs of 3 conditions: Attention+Stimulus, Stimulus, Attention. Data were analysed using Freesurfer and averaged across subjects using spherical surface morphing. Significant retinotopic responses were found in a number of areas: For the Attention+Stimulus condition, we saw maps in primary visual, lateral and ventral temporal,...