Purchase this article with an account.
Lilian Azer, Weiwei Zhang; The Effects of Structural Regularity on Working Memory Representations. Journal of Vision 2018;18(10):683. doi: 10.1167/18.10.683.
Download citation file:
© ARVO (1962-2015); The Authors (2016-present)
Previous experience and long-term memory can influence various aspects of working memory representation and processing. The present study has assessed how structural regularity of to-be-remembered information affects item and configural encoding in visual working memory (VWM) using Xie and Zhang's (2017) dual-trace Signal Detection Theory model. With this model, item and configural information in VWM can be characterized as discrete and continuous components, respectively. Two types of structural regularity, face versus non-face and goodness-of pattern, were tested in two experiments. Experiment 1 tested VWM for a set of stimuli that had matched low-level physical attributes but differed continuously on faceness rating in a modified change detection task. In this task, observers were required to retain four face or non-face stimuli over a 1-s retention interval and then reported whether a cued item in the test display was an old or new stimulus on a 6-point confidence scale. The resulting Receiver Operating Characteristics (ROC) curves were fit with the dual-trace Signal Detection Theory model, producing estimates of item and configural encoding. No significant difference was found in either measure for face and non-face stimuli. Nonetheless, there was a more liberal bias for face than non-face stimuli, leading to more false positives for faces. Experiment 2 replaced face versus non-face stimuli with dot-array patterns that differed in pattern goodness. Good patterns were remembered better than poor patterns, largely due to increased configural encoding for good patterns. Together these findings suggest that structural regularity of memory stimulus at different levels (category versus exemplar) have dissociable effects on VWM storage.
Meeting abstract presented at VSS 2018
This PDF is available to Subscribers Only