Conceptual and perceptual chunking of real-world objects in visual working memory

We investigated the impact of chunking on visual working memory (VWM) capacity and explored whether its benefits depend on the overlap between studied objects and retrieval probes. Informed by Fuzzy Trace Theory, we examined whether chunking real-world objects is a fully gist-reliant process or whether it preserves coarse- or fine-grained visual detail. To test this, participants completed four experiments involving object pairs that varied in conceptual and/or perceptual similarity. At test, they judged whether a probe was an "old" (studied) or "novel" (unseen) item. Some distractors were "related lures", similar to targets. We compared hit rates and false alarms across conditions (related vs. unrelated pairs), using a signal detection-inspired model. Item-specific and gist-based indices of memory detection showed that chunking enhances object memory for related object pairs, but probing memory with lures related to studied objects revealed that this benefit was largely attributable to gist-based and coarse-grained visual detail memory traces. The findings suggest that chunking based on object relatedness can increase VWM capacity, but at the cost of representing fine-grained visual details of chunked objects.