Abstract
The flashed face distortion effect (FFDE) is a visual illusion in which faces presented rapidly in the visual periphery appear grotesquely distorted. Previous studies have suggested that the FFDE may involve higher-level mechanisms of face perception that integrate information across temporally and spatially adjacent stimuli. However, the effect may also arise from lower-level visual processes, such as encoding local edges or orientations. To provide further insight, the present study examined whether the FFDE is represented in retinotopic or spatiotopic coordinates, which are often associated with different stages of visual processing. Participants made a saccadic eye movement during the sequential presentation of face stimuli. Across the saccade, faces were presented either at the same retinal location (retinotopic coordinates) or at the same screen location (spatiotopic coordinates). The subjective strength of the FFDE for postsaccadic faces was measured using a seven-point rating scale. The results showed that the FFDE was pronounced in retinotopic coordinates but was substantially reduced in spatiotopic coordinates. These findings suggest that the FFDE is primarily mediated by relatively low-level visual processing mechanisms, similar to classical visual aftereffects.