Abstract
With each saccade the position of visual information falling on the retina shifts. Despite these transsaccadic disruptions, we perceive a stable visual world. This stability persists even when stimuli move during saccades, as shown by poor transsaccadic displacement detection (e.g., Mack, 1970). In simple stimuli, displacement detection improves dramatically when stimuli briefly disappear at the conclusion of the saccade (e.g., Deubel et al., 1994). This “blank” effect is taken as evidence that the visual system assumes stability unless provided contrary evidence like that introduced by the blank period. The present study examined whether a post-saccade blank similarly improves displacement detection in real-world scenes, as the stability assumption hypothesis predicts. In two experiments, participants were cued to move their eyes away from central fixation to the left or right while viewing a real-world scene. On some trials, the entire scene moved 2° (E1) or 4° (E2) while their eyes were in motion. On half of the trials (i.e., the blank condition), the scene was briefly absent at the conclusion of the eye movement before returning to view. Participants responded whether or not they detected a scene displacement. In contrast with previous work and with our own control experiment (E3), the post-saccade blank screen impaired displacement detection in scenes. Participants were less sensitive (d’) to scene displacements in the blank condition (E1: p < 0.001, d = 0.99; E2: p < 0.001, d = 1.45). In contrast, participants were more sensitive to simple stimuli displacements with the post-saccade blank (p = 0.002, d = 0.91), replicating the classic effect. These results suggest that the blank effect does not scale up to complex scene images, and calls into question the generality of the stability assumption hypothesis that has found support in studies of simple visual stimuli.