Abstract
The current study further examined the hypothesis that post-cues can encourage the retrieval of long-term memory (LTM) representations and lead to improved change detection performance (Beck & van Lamsweerde, 2011) using a gaze contingent change detection task. Nine objects were presented and after seven objects were fixated, a brief blank screen was presented followed by the test image in which one item changed identity or all items were unchanged. The potential change object was either cued in the test array or not cued. The change object could be any of the seven items fixated prior to the change (lags 0-6) or one of the two items never fixated. The availability of identity information in the periphery was manipulated by using a moving window technique for half of the participants. When an object was not being fixated, it was blurred so that its location was visible but its features were not. Change detection performance was better for the cue trials than the no-cue trials at all lags for both the moving-window and no-moving-window conditions. When there was no moving window, the cue benefit was also found for objects that were never fixated. However, when the moving window was used, performance on the cue trials was not better than performance on the no-cue trials for items that were never fixated. Therefore, cues do aid change detection performance for items no longer stored in visual working memory (~lags 3-6), and cues improve change detection performance for items that were never fixated if features can be identified in the periphery. Furthermore, fixating the change object in the test array was predictive of accurate performance for the cue trials, but not for the no-cue trials, suggesting that the cue does encourage LTM retrieval that does not necessarily occur without a cue.
Meeting abstract presented at VSS 2012