Abstract
Visual perception operates in an object-based manner and working memory allows a flexible access to a limited number of object representations, even when they are physically no longer present. Recently, it has been shown that we compensate for small changes of an object's content features over memory episodes, which supports its perceptual stability. This phenomenon was termed 'serial dependence' and has mostly been studied in situations containing only a single relevant object. However, since we are typically confronted with several relevant objects, the central question of how we selectively create temporal stability of several objects has remained unsolved. As different objects can be distinguished by their context features, in our case color, temporal position or spatial position, we tested whether serial dependence is supported by the congruence of context features across trials. We asked participants to remember the content features of two objects (i.e., motion directions of colored dot fields) per trial, which were presented sequentially (Experiments 1 and 2) or simultaneously (Experiments 3 and 4). At the end of a trial, one motion direction was cued for continuous report either by its color (Experiments 1 and 4), serial position (Experiment 2) or spatial position (Experiment 3). Across all experiments we consistently observed serial dependence, i.e., an attractive bias of currently toward previously memorized motion directions that was clearly enhanced when objects shared the same context feature. These findings demonstrate that the binding of an object’s content (here, motion direction) and context features (e.g. color in Experiment 1 and 4) is not completely erased after a trial, but carried over to subsequent episodes. As this reflects temporal dependencies in natural settings, the present findings reveal a mechanism that supports stable representations of individualized objects over time.