Abstract
Visual working memory has very strict capacity limitations; on average, people are able to hold in their memory only between 3 to 4 independent objects. Although most studies examined the maintenance of visual information over time, capacity limitations might originate at the earlier, encoding phase. In the current study, we tested this hypothesis directly; we modified the classical change detection paradigm by eliminating the maintenance phase. In change detection, two stimulus arrays are presented sequentially, with a temporal gap and observers are asked to report a change between two arrays. We have eliminated the temporal gap and presented two arrays simultaneously, side by side, creating a parallel change detection. In the first experiment, we compared detection of change in object orientation in both classical and parallel change detection paradigms. Observers' performance dropped dramatically with increasing set size, even in parallel change detection task, in which two arrays were presented simultaneously. In the second experiment, we tested two visual features, orientation and size and their conjunction in parallel change detection task. We replicated results from the first experiment; furthermore, observers showed similar drop in performance with increasing set size for both, object size and conjunction of size and orientation. Our results suggest that visual working memory capacity limitations originate at the early, encoding phase of visual information processing and are therefore limitations of visual perception and not limitations of memory per se.
Meeting abstract presented at VSS 2016