Matching of visual input to only one item at any one time

Houtkamp, Roos; Roelfsema, Pieter R. · 2008 · OpenAlex-citations

DOI: 10.1007/s00426-008-0157-3

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 whether human observers can simultaneously match visual input to multiple items held in working memory, addressing a gap in understanding the capacity limits of visual search. While previous research established that working memory can store approximately three to four items, it remained unclear if multiple stored items could act as active "search-templates" in parallel. The authors hypothesized that matching visual input is fundamentally limited to one item at a time, regardless of the number of items stored in memory. To test this, the researchers employed a rapid serial visual presentation (RSVP) paradigm across three experiments. Participants viewed a stream of 30 colored shapes and were tasked with detecting if a specific target appeared. In Experiment 1 (Shapes), subjects searched for one or two target shapes. In Experiment 2 (Colors), they searched for one or two target colors. In Experiment 3 (Combined), subjects searched for one shape, one color, or one of each simultaneously. Performance was measured using a staircase procedure to determine detection thresholds. The authors compared subject performance against two theoretical models: a "two-template model," which assumes parallel matching processes, and a "one-template model," which assumes only a single active template can match visual input at any given moment. The results demonstrated a significant drop in performance when subjects searched for two targets compared to one. In the Shapes experiment, accuracy fell from 90% to 65%; in the Colors experiment, it dropped from 82% to 69%; and in the Combined experiment, it decreased from approximately 81% to 67%. Statistical analysis revealed that these performance levels were consistent with the one-template model but significantly lower than the predictions of the two-template model. The estimated number of active templates averaged close to 1.0 across all experiments and subjects, indicating that observers effectively utilized only a single search-template at a time. This limitation persisted even when targets differed in feature domain (shape vs. color), ruling out interference due to shared feature dimensions. The findings conclude that the matching process in visual search has a strict capacity limitation, allowing only one item from working memory to actively guide attention at any moment. This suggests a distinction between the storage capacity of working memory, which can hold multiple items, and the executive matching process, which is serial. The results imply that when searching for multiple items, observers must switch between templates sequentially rather than processing them in parallel. This provides a parsimonious explanation for previous findings regarding the weak influence of accessory memory items on visual search and supports theories distinguishing between passive storage and active executive selection in working memory.

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 2026-06-25; verification: verified.

Topics

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

• visual search