Journal of experimental psychology-learning memory and cognition vol:23 issue:3 pages:739-760
Participants had to detect saccade-contingent displacements of a point-light walker. When the walker was the only visible object, displacements were relatively hard to detect, indicating that transsaccadic memory for an object's egocentric position is inaccurate. Adding a landmark point-light figure (allowing allocentric coding) decreased detection thresholds only slightly. When the target walker was flanked by 2 companion walkers, accuracy increased, provided that processes of perceptual organization were triggered. First, closer objects functioned as more efficient landmarks. Second, detection was best when the configuration of objects afforded grouping by proximity. The finding that allocentric coding aids transsaccadic memory only when organizational processes are mobilized suggests that transsaccadic integration is not achieved exclusively by a low-level specialized mechanism, but that more general perceptual processes play a prominent role.