Gestalt Grouping Cues Can Improve Filtering Performance in Visual Working Memory

[摘要] We investigated compensation mechanisms aimed at preventing task-irrelevant items from entering the limited Visual Working Memory (VWM) workspace. Previous studies have shown that Gestalt grouping cues improve VWM performance. Yet, it is unknown whether Gestalt grouping cues can be used to ameliorate filtering performance in VWM. In Experiment 1 participants preformed a change-detection task in which they viewed memory arrays with either three targets, six targets, or three targets and three distractors (i.e., the filtering condition), and where asked to remember only the targets' colors and to ignore the distractors (based on their shape). In the filtering condition, the targets formed a Kanitza triangle (closure), appeared in a triangle-like configuration (proximity), or appeared at random positions. We found that the proximity condition was able to ameliorate filtering performance. However, the closure condition did not further improved filtering performance (beyond proximity) when the targets formed a Kanitza triangle. Presumably, grouping the targets biased their selection over the distractors, resulting in a better performance. Previous studies have shown that different processes underlie selection of targets and rejection of distractors. In Experiment 2, participants preformed the same task as in Experiment 1, however, this time we implemented the Gestalt grouping cues on the distractors. Similar to Experiment 1, filtering performance was better in the proximity condition and the closure condition did not further improve filtering performance when the distractors formed a Kanitza triangle. The results of this experiment demonstrate that grouping the distractors served to ignore them, either by directly inhibiting distractors selection or by indirectly reinforcing targets' encoding. We conclude that grouping affects filtering performance via targets selection or distractors de-selection processes.

[学科分类] 眼科学