Statistical learning in the past modulates contextual cueing in the future.

Area of Science:

• Cognitive psychology
• Visual perception
• Attention and visual search

Background:

• Visual search is enhanced by statistical regularities, such as repeated target locations (probability learning) and familiar nontarget arrangements (contextual cueing).
• Contextual cueing, as defined by Chun and Jiang (1998), uses familiar spatial contexts to guide attention to target locations.

Purpose of the Study:

• To investigate how probability learning influences the adaptation of contextual cueing when target locations change.
• To determine if probability learning alone is sufficient for adaptive contextual cueing under relocated target conditions.

Main Methods:

• An initial learning phase established probability learning and contextual cueing effects.
• Targets were subsequently relocated within their contexts to new positions, which had prior exposure in different contexts.
• Observed contextual cueing effects for relocated targets based on their origin (old vs. new contexts).

Main Results:

• Contextual cueing was present for relocated targets that originated from previously learned (old) contexts.
• Contextual cueing turned into a performance cost when relocated targets originated from unlearned (new) contexts.
• Probability learning alone did not ensure adaptive contextual cueing for relocated targets.

Conclusions:

• Adaptive contextual cueing depends on the contextual history of target locations, not solely on probability learning.
• Observers integrate multiple statistical cues, including the contextual past of target locations, to predict future stimulus occurrences.
• The findings suggest a complex interplay between different statistical learning mechanisms in guiding visual attention.