We perceive our environment by breaking it down into segments known as events. Event segmentation influences memory by enhancing the retention of information at boundaries as compared to information that is contained within the boundaries of an event (the event boundary advantage). This effect has been attributed to changes in attention during perception of events. Prior studies have demonstrated greater attention while perceiving event boundaries but have failed to demonstrate attention as the underlying mechanism for the event-boundary advantage. Two behavioral experiments were conducted to investigate, a) whether the event boundary advantage is observed even for events that are perceived while performing a concurrent task? and b) Is there a decrease in the boundary advantage when the concurrent task complexity is increased? In both experiments, participants watched videos related to performance of daily tasks, while simultaneously performing a probe detection task; either a simple dot detection (Experiment 1) or a go/ no-go task (Experiment 2). The probe was presented either at an event boundary or at pre-defined non-boundary time point and the memory for both temporal locations was measured after the completion of the detection task. A mixed effects logistic regression revealed an interactive effect for both detection accuracy and the boundary advantage; probe detection at event boundaries remained unaffected throughout an event irrespective of the level of the task complexity while, contrary to prediction, a boundary advantage in memory was also observed. But detection and memory accuracy for non-boundaries decreased successively for both low and high secondary task complexity suggesting greater interference for processing non-boundary information. These results indicate that greater attention may not be the only predictor of better memory for event boundaries as postulated by Event Segmentation theory.

Event boundary advantage (EBA) refers to greater memorability of information at boundaries than for any other part of an event. Recent studies have identified post-encoding processes as a likely source of EBA. The current study investigated whether boundaries are distinctly remembered because they act as gateways for retrieval of associated event-elements by using a trace modification paradigm where memory for the last item of an encoded triplet (A-B-C) is suppressed by replacing it with a novel item on re-exposure (A-B-D). Two hierarchical Bayesian models tested whether the immediate associate, boundary item or only the category difference between old and new item of the triplet predicts suppression. Results indicate suppression is predicted by an interaction between memory of A and category, but successful updating is predicted by an interaction between memory of B and category. We discuss the implications of the result for understanding role of event boundaries in trace updating.