Visual stability in naturalistic scenes

Abstract

The current study examines how visual stability is established for naturalistic scenes. Previous studies have shown that detection of position shifts is significantly better when the shift occurs for the saccade target object compared to either a background shift or a whole image shift (Currie et al., 2000). Other studies have shown that removing the target object from the screen for a short period of time (i.e., blanking) significantly improves shift detection (Deubel et al., 1996). Here we tested whether the blanking effect would similarly improve shift detection for background and whole image shifts for naturalistic scenes. Participants were presented with scene images and instructed to execute a saccade to a highlighted target object. During the saccade, there could be three different shifts: only the saccade target shifted, the whole image shifted, or the background shifted but the target remained stationary. We also included control trials where no shift occurred. Half of the trials in each condition had a 250ms blank (target blank, context blank, or all blank) that occurred as soon as the saccade was detected. Participants were asked to indicate whether they detected a move. We found a significant effect of condition where saccade target shifts resulted in highest detection rate, and the background shifts had the lowest detection rate. More importantly, we found that blanking significantly interacted with condition: Blanking only improved shift detection in target shift condition but not in background or whole image shifts. These results suggest that the visual system uses a localized solution for establishing object correspondence across saccades where it mainly relies on the saccade target to determine stability.