Abstract
Humans are very good at remembering large numbers of scenes over substantial periods of time. How good are they at remembering changes to scenes? In this study, we compared the long-term effect on scene memory and subsequent change detection of a previous visual search for change compared to intentional memorization of scenes. First, participants either performed a change detection task in natural scenes or explicitly memorized those scenes. Two weeks later, they were shown single scenes and asked if they were old or new. Then they were asked to recall changes (if the initial task was change detection) or to detect a change between the initial and current presentations (if the initial task was memorization). Finally, they searched for that change by alternating between the two versions of the scene. Incidental memory for the scenes, viewed during an initial change blindness was the same as memory after intentional memorization (~67-70% correct recognition and ~13-15% false recognition), with better memory for scenes where changes had been found. Unsurprisingly, observers were fast to find the change the second time if they remembered the change. More surprising, observers were faster to find changes that they had previously found but did not remember finding. This result was replicated in two of three change detection experiments. Time required to find changes in the first and second phases of the experiment were correlated in scenes where the observers didn't remember the change, perhaps showing some similarity between search strategies. Observers showed better memory for changes that had been hard to find. Control experiments rule out boring accounts for this effect (e.g. initial exposure duration). We conclude that scenes can be encoded incidentally as well as explicitly and that changes in those scenes can leave measurable traces even if they are not explicitly recalled.
Meeting abstract presented at VSS 2016