Abstract
The unit of visual working memory (VWM) has been widely discussed in last decades. Previous studies often treated the object as a unit because of the same-object advantage effect. This traditional assumption was challenged by a recent study (Huang, 2019), in which the behavior evidence consistently showed a significant same-Boolean map advantage but no same-object advantage effect, suggesting Boolean map as the unit of VWM. However, it remains unknown whether the information has been maintained in the form of a Boolean map since it entered VWM. Therefore, we adopted the contralateral delay activity (CDA) in two experiments, tracking the stored information on-line to test the unit of VWM directly. In Experiment 1, participants performed a lateralized change-detection task, wherein memory items varied in the number of objects and Boolean maps. There were 4 conditions: one single-color circle (single Boolean map and single object, 1B1O), two circles in different colors (2B2O), one two-color circle that was composed of two semicircles in different colors (2B1O), and two same-color circles (1B2O, not analysis because of participants’ strategy). The results showed that there was a dissociation in the distribution of activity between conditions. At P7/P8, the CDA in 2B1O was significantly higher than 1B1O, and was not significantly different from 2B2O, supporting the Boolean map theory. At P3/P4, the early CDA was more sensitive to the number of objects, reflecting in that the CDA in 2B1O was significantly lower than 2B2O, and was not significantly different from 1B1O. In the later period, there was no difference between 2B1O and 2B2O, biasing to the Boolean map theory. In Experiment 2, we doubled the number of stimuli and replicated Experiment 1. Overall, these findings indicated that the Boolean map provided a better explanation for the unit of VWM than the object.