Abstract
Previous work suggests that the similarity of distracting information can differentially alter how visual representations are forgotten. Though these effects have been shown for color memory, it is unclear if they reflect more general principles and will extend to other object features such as shape. In Experiment 1, we describe the creation of the Validated Circular Shape Space (VCS space), a "Shape Wheel" whereby 2D line drawings were morphed together to create an array of 360 shapes, corresponding to 360 degrees on a circle. We iteratively developed VCS space by designing prototype shapes, morphing the shapes to create a circular space, collecting pairwise similarity ratings, constructing subjective space using multidimensional scaling (MDS), then redesigning the prototype shapes that were problematic. This procedure required seven validation steps, with the final validation step ensuring that angular distance on VCS space was a proxy for subjective similarity. In Experiment 2, we then used VCS space to assess how shape memory was impacted by distracting information that varied in subjective similarity relative to a study shape. A mixture model was used to operationalize memory as two separate components: the probability that study items were successfully remembered, defined as accuracy, and the level of detail, defined as precision. Relative to a baseline scrambled-line condition, we found that subjectively dissimilar distractors decreased accuracy but not precision, while subjectively similar distractors decreased precision but not accuracy. These findings extend previous literature by demonstrating the nature of visual interference differentially impacted shape memory. Subjectively dissimilar interference erased memory, whereas subjectively similar interference blurred representations. As these effects were consistent across studies, we suggest they provide converging evidence for a set of general principles regarding how interference impacts visual representations and the features therein.
Meeting abstract presented at VSS 2018