Holistic Representation as an Alternative Format of Visual Working Memory
Poster Presentation: Tuesday, May 19, 2026, 8:30 am – 12:30 pm, Banyan Breezeway
Session: Visual Working Memory: Models, neural
Schedule of Events | Search Abstracts | Symposia | Talk Sessions | Poster Sessions
Ziyi Su1, Brad Wyble1; 1Penn State University
Working memory representations are commonly assumed to store information as separate objects. However, recent work emphasizes the flexible nature of working memory representation such that different types of representations might be formed according to distinct task requirements (Lecerf & Ribaupierre, 2010; Gonthier, 2021). One alternative format of representation is a holistic representation: representing the whole display as a single object without decomposing it into distinct objects or features. This type of representation could be strategic in tasks that favor retrieving or manipulating the entire display. Change detection, for example, can be conceivably accomplished by comparing a holistic representation of the remembered array against the test display. Studies of memorizing complex scenes reported a similar strategy of encoding “the gist” as one part of the memory (e.g., Rensink, 2000). However, holistic representations have not been widely recognized as a format readily applied for other visual working memory tasks. Therefore, this study probes evidence for holistic representation with colored square arrays. Participants encoded 6 colored squares for each trial and completed a change detection test in which 1, 3 or 6 of the squares would be probed, unpredictably. The squares were initially presented simultaneously during the encoding phase for one group, which encouraged holistic representation, and were presented sequentially for the other group, which encouraged item-based encoding. We predicted better performance for probe 6 exclusively in the simultaneous condition because a holistic memory could be used in its entirety while the probe 3 condition required selecting a subset of the squares. Results confirmed this: Simultaneous presentation elicited higher performance when probing 6 rather than 3 items; but for sequential presentation, subjects had similar performance for probing 3 and 6 squares. These results suggest strategic use of holistic representations in certain task configurations, and support an alternative to object-based storage.