Contrasting gist-based and template-based guidance during real-world visual search

Area of Science:

• Cognitive Psychology
• Visual Perception
• Neuroscience

Background:

• Visual search in complex scenes is influenced by target features and scene semantics (gist recognition).
• The interplay between feature-based template guidance and semantic scene-level guidance requires further investigation.

Purpose of the Study:

• To compare the relative roles of feature-based template guidance and semantic scene gist recognition in guiding visual search.
• To determine the temporal dynamics and independence of these two guidance mechanisms.

Main Methods:

• Three experiments involved participants searching for a target object amidst distractors.
• Distractor color was manipulated to match or mismatch target color or a visual working memory item.
• Gaze capture by matching distractors was measured as an index of template guidance.

Main Results:

• A distractor match effect, indicating template guidance, was present from the initial gaze saccade.
• This effect was independent of scene gist information and distractor location plausibility.
• Semantic, gist-based guidance was slower to influence gaze compared to feature-based template guidance.

Conclusions:

• Feature-based template guidance operates early in visual search and is not initially constrained by semantic scene analysis.
• Semantic scene gist guidance influences gaze later than feature-based guidance.
• These findings suggest that feature-based guidance can guide attention to non-plausible regions before scene-level semantic analysis is fully utilized.