The Effect of Task-Irrelevant Location Information in Visual Working Memory of Feature Binding: Manipulation in Test Location

Previous research has demonstrated that in visual working memory task, participants tend to process each item’s location information even when it is task-irrelevant. This is demonstrated in the evidence showing that, compared with testing an item at its original encoding location, testing in a new location impairs memory. For the location-change manipulation, previous studies manipulated the item’s location during the testing phase but did not distinguish between two independent mechanisms: displacement from the encoding location (displacement effect) and overlap with another item’s location during encoding (displacement + interference effect). Note that the latter effect could include interference from memory of the previous item in that location. In the present study, we employed a change-detection task requiring memory of color–orientation bindings (OLD vs NEW judgment). We presented multiple items sequentially during the encoding phase and in isolation during the testing phase. During testing, the item’s relative location between the testing and encoding phases was either congruent or incongruent. When the location was incongruent, by varying the probe location with or without overlapping the location of another item during encoding, we separated the effects of displacement and displacement plus interference on binding memory. For the OLD probes, performance for test locations congruent with the original encoding location was much better than for the incongruent locations. For the two types of incongruent locations, the location occupied by an item during encoding did not cause additional impairment. For the NEW probes, accuracy did not differ significantly between empty and occupied locations when the probe shared only orientation or neither feature with the previous occupant. However, NEW probes sharing the previous occupant’s color showed lower accuracy and faster correct responses at occupied than at empty locations. This pattern may reflect preferential encoding or use of color–location information, although feature-dependent interference during retrieval remains another possible explanation.