Abstract
Visual working memory (VWM) allows us to actively represent a limited amount of visual information in mind. A classic study by Luck and Vogel (1997) demonstrated that visual features are represented as integrated objects in VWM. On the other hand, recent studies (e.g., Fougnie and Alvarez, 2011) had participants represent multi-feature objects (e.g., colored triangles) in their VWM and had them recall all the features of the same object sequentially. The result revealed that multiple features composing the same object (e.g., color and orientation) are prone to independent and probabilistic representational failures, thus questioning the integrated object account of VWM representations. To reconcile this discrepancy, we invented a novel VWM recall paradigm that enabled simultaneous recall of multiple features. More precisely, participants encoded one or two colored triangles in their VWM, and after a short retention interval, a test probe was presented at one of the stimulus locations to prompt participants’ recall. Critically, this test probe was composed of 180 concentric color rings whose radius covaried as its color changed continuously on a circular CIELab color space. Thus, by clicking on a specific color ring at a specific angular direction, participants reported the color and the orientation of the tested item simultaneously. Using this novel approach, we found that 1) remembering two features of one object is less taxing to our VWM than remembering one feature each from two objects and 2) the representational quality of multiple features of the same object tended to covary. Furthermore in Experiment 2, we replicated these results for two objects presented at the same location, thus demonstrating that the shared location is not enough to explain the object benefit. Taken together, our results provide a strong support for the integrated object account of VWM representations.