Abstract
Purpose. When a figure moves behind a stationary slit, observers report seeing the figure as an integrated whole although there is no extended retinotopic image (anorthoscopic perception). Using visual masking, we tested whether these figures are of perceptual origin or a result of a cognitive process (e.g., observers can infer shape using salient features such as the wheel of a car) and whether any part of the figure is represented non-retinotopically beyond the boundaries of the slit.
Methods. In the main experiment, a 5×5 square grid array moved (5.7 deg/sec) behind a slit (21 arcmin wide) in the presence or absence of a briefly flashed mask which overlapped neither with the slit nor with the path of the moving grid. Observers were required to make a same/different judgment between the moving grid and a subsequently presented stationary comparison grid. In the control experiment, rather than moving the grid, each column of the grid was sequentially flashed in the slit.
Results. In the main experiment, performance was significantly degraded in the presence of the mask. The analysis of the effect of the mask on individual columns of the grid reveals that the mask was most effective for a column presented 60 ms prior to the presentation of the mask. Our control experiment showed that these results are not contaminated by attentional capture, masking of the slit area, and/or cognitive strategies. Given that the mask appears to exert its effect on spatial regions beyond the slit, our results support the idea that the grid stimulus is represented non-retinotopically beyond the boundaries of the slit.
Conclusions. Anorthoscopic percepts are not of a cognitive origin and are represented in a spatially extended manner in a non-retinotopic coordinate system. Our results also show that masking can occur in non-retinotopic coordinates.