Combined conceptual and perceptual control of visual attention in search for real-world objects

Bahle, Brett; Winsler, Kurt; Kiat, John E.; Luck, Steven J. · 2025 · OpenAlex-citations

DOI: 10.3758/s13414-025-03116-4

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 how visual attention is guided during the search for real-world objects, specifically examining the interplay between perceptual features (visual appearance) and conceptual information (semantic meaning). While traditional models of visual search, such as Guided Search, posit that attention is driven solely by low-level perceptual features, this research addresses the gap in understanding how search operates when specific visual details are unknown. The authors hypothesize that conceptual information can guide attention, particularly when observers know only the category of a target object rather than its specific appearance. To test this, the researchers conducted experiments using eye-tracking to measure overt attention during a two-item visual search task. Participants searched for a target among a distractor, cued either by a picture of the specific target or by a category label (e.g., "grasshopper"). Crucially, the study quantified both perceptual and conceptual similarity between targets and distractors using continuous metrics. Perceptual similarity was derived from the THINGS database, which maps objects into a 49-dimensional embedding space based on human categorization judgments. Conceptual similarity was derived from ConceptNet, a semantic network trained on large text corpora, mapping objects into a 300-dimensional semantic space. This allowed the authors to statistically control for one type of similarity while assessing the predictive power of the other. The design compared performance on the first exposure to a target-distractor pair (when perceptual features were unknown) versus the twelfth exposure (when features were learned). The results demonstrated that the guidance of visual attention shifts from conceptual to perceptual control as observers gain experience. When participants searched for a target for the first time using only a category cue, their eye movements were strongly predicted by conceptual similarity, with perceptual features showing no significant predictive ability once conceptual information was controlled. Conversely, when participants had previously seen the specific target (picture cue or later exposures), search behavior was strongly predicted by perceptual similarity, with conceptual features losing their predictive power. This indicates that conceptual information effectively guides attention when precise perceptual templates are unavailable. These findings provide novel evidence that conceptual information plays a significant role in guiding visual search, challenging the view that search is driven exclusively by low-level perceptual features. The study implies that the brain utilizes semantic networks to prioritize attention in naturalistic scenarios where specific object appearances are unknown. Furthermore, it highlights a dynamic transition in attentional control: as observers acquire specific perceptual knowledge of a target, they progressively rely less on conceptual guidance and more on perceptual matching. This has important implications for understanding how humans navigate complex visual environments and for refining cognitive models of attention to include semantic influences.

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