Meaning maps predict reaction time in change detection

Area of Science:

• Cognitive Neuroscience
• Visual Perception
• Computational Vision

Background:

• Change detection is crucial for adapting to dynamic environments.
• Prior studies suggest high-level semantic information guides change detection in natural scenes.
• Working memory actively maintains representations for comparison.

Purpose of the Study:

• To investigate the role of semantic features in visual change detection.
• To determine if semantic information predicts change detection accuracy in a flicker paradigm.
• To explore the influence of scene meaning on visual processing.

Main Methods:

• Utilized meaning maps, a data-driven method, to quantify semantic information in naturalistic scenes.
• Employed a flicker paradigm to present visual stimuli and record change detection performance.
• Conducted two experiments manipulating scene meaning and orientation (Experiment 2 used scene inversion).

Main Results:

• Changes in highly meaningful regions were detected more readily than in non-meaningful regions, even when controlling for low-level saliency.
• Scene inversion significantly reduced the advantage for detecting changes in meaningful regions.
• Semantic information demonstrably influences visual change detection.

Conclusions:

• The visual system actively leverages semantic features for effective change detection.
• Meaningful content plays a significant role in how we perceive and react to changes in our environment.
• These findings highlight the importance of semantic processing in visual working memory and scene understanding.