Statistical learning of distractor co-occurrences facilitates visual search

Area of Science:

• Cognitive Psychology
• Visual Perception

Background:

• Visual search relies on target-distractor relationships.
• The role of distractor-only statistical structure in visual search is unclear.

Purpose of the Study:

• To investigate if implicit learning of distractor co-occurrence statistics benefits visual search.
• To determine if structured distractor arrays enhance search efficiency compared to unstructured ones.

Main Methods:

• Participants performed visual search tasks with novel shapes.
• Two conditions were used: structured scenes (co-occurring distractor pairs) and unstructured scenes (randomly paired distractors).
• Five online experiments (N=1,140) involved search training.

Main Results:

• Search efficiency was significantly higher in structured scenes than unstructured scenes after training.
• This structure benefit was observed regardless of fixed or variable shape positions within pairs.
• Participants showed improved search efficiency without explicit knowledge of distractor pairings.

Conclusions:

• Implicitly learned co-occurrence statistics among distractors enhance visual search efficiency.
• This suggests that the brain can leverage statistical regularities in distractor arrangements.
• This finding may explain the efficiency of visual search in natural environments rich with regularities.