Constructing Gestalt in Visual Working Memory
M. Shen,Q. Gao,N. Tang,R. Shui,S. Chen,Z. Gao
DOI: https://doi.org/10.1167/14.10.34
2014-01-01
Journal of Vision
Abstract:So far ample studies have demonstrated that VWM plays a critical role in several fundamental cognitive processes, such as perception, language processing, and planning. A critical factor that makes VWM so important is that VWM could "actively" maintain and manipulate the incoming information. However, so far most of the studies focus on a relatively "static" aspect of VWM, for instance, capacity, representation resolution, etc. Few studies have attempted to explore the active aspect of VWM. Here we investigated the active part of VWM by asking whether a Gestalt could be constructed in VWM based on the incoming information. Particularly, in a modified change detection task, we sequentially presented the memorized objects. Importantly, in 50% of trials these objects could form a virtual rectangle or triangle (i.e., a Gestalt) when they were presented simultaneously. We predicted that if the VWM could actively hold the visual information, then it will detect the relationship among the objects and construct a Gestalt based on the stored objects, which will help reduce the memory load. In line with this prediction, in 5 experiments we consistently found that when a potential Gestalt could be constructed among the memorized 3 or 4 objects, VWM performance was significantly improved. These results suggest that VWM indeed is actively, instead of passively, involved in holding visual information. Meeting abstract presented at VSS 2014