Abstract
This study investigated whether objects are represented as individual features or as integrated objects in visual memory. Participants were presented with pictures of real-world objects and were later tested for their memory. There were four options for each object, created by combining two different exemplars (e.g., two different bottles), and two different variations of another feature. For different objects, the other feature was either state (e.g., open or closed), orientation (e.g., straight or rotated) or material (e.g., glass or metal). Participants were asked to choose the object they saw in the first part. Three days after the first session, participants returned for a second memory test phase. As expected, we observed a decrease in the number of remembered items in the long-term test as compared to the immediate memory test. The objects-based theory states that we represent objects in memory as integrated objects and not as individual features. Therefore, the object-based model predicts that the number of forgotten items between the short and long term test should match the increase in the guessing rate, because the object should be forgotten completely. Furthermore, this model predicts that there would be no increase in the partially-correct answers (i.e., objects for which only one feature was remembered) between the short and long-term test. The results verified these predictions: when comparing the short and long-term test, the decrease in the number of fully correct answers was equivalent to the increase in guessing rate (t(19)=0.64, p=0.53). In addition, there was no significant difference in the partially-correct answers (t(19)=0.55,p=0.58) comparing the short and long-term test. Several following experiments replicated these results with a shortened presentation time and with an explicit memory study phase. These results support the object-based theory that claims memory hold integrated objects instead of represented separated features.
Meeting abstract presented at VSS 2018