Visual working memory can selectively reset a subset of its representations

Balaban, Halely; Drew, Trafton; Luria, Roy · 2017 · Crossref

DOI: 10.3758/s13423-017-1400-y

archive: archived pipeline: cataloged verified

Get this paper ↗ (DOI — opens at the source; we link to it, we don't host it)

Summary

This study investigates the mechanism of resetting in visual working memory (VWM), specifically determining whether the process is "global" (discarding all active representations when one mapping is lost) or "local" (selectively removing only the representations whose object-to-representation mappings are no longer valid). The authors were motivated by the need to understand how VWM manages online processing when environmental objects change in ways that break their correspondence with stored representations, such as when a single object splits into two independent parts. To test this, the researchers conducted two experiments using an online change-detection task. Participants monitored moving polygons and reported if any shape changed during movement. In the critical condition, one or more polygons separated into two halves that moved independently, triggering a resetting process because the original single-object mapping became irrelevant. Experiment 1 manipulated the number of separating polygons (one or two out of three) to see if losing more mappings increased the performance cost. Experiment 2 manipulated the total set size (two or three items) while keeping the number of separating items constant (one) to determine if the cost depended on the total load in VWM. The results consistently supported the local resetting hypothesis. In Experiment 1, participants missed shape changes occurring simultaneously with the separation of a polygon, but accurately detected changes in polygons that did not separate. Crucially, the performance cost for the separated item increased when two polygons separated compared to one, indicating that resetting operations are additive and specific to the affected items. In Experiment 2, the cost associated with resetting a single item was unaffected by whether the total set size was two or three, demonstrating that the resetting process does not depend on the overall number of items held in memory. Furthermore, changes in unseparated items were detected with high accuracy, confirming that intact mappings remained accessible even while other representations were being reset. These findings imply that VWM resetting is a highly efficient, item-specific process rather than a global wipeout of memory contents. This selective mechanism allows VWM to maintain stable representations of relevant objects while discarding obsolete ones, which is essential for functioning in dynamic environments. The results provide strong behavioral support for the object file theory, suggesting that VWM relies on ongoing correspondence between representations and environmental objects, and that this correspondence can be updated or reset locally without disrupting the entire memory state.

Provenance

The full processing record for this entry. Every stage of this paper's journey through the pipeline is logged — what ran, with which tool and model, how many attempts it took, and when it last completed.

Summary generated by qwen3.6-27b-prismaquant on 2026-06-26; verification: verified.

Topics

Ranked by relevance to this paper. Hover a topic for its definition.

• working memory