Statistical learning and relearning in visual attention: transfer effects in distractor suppression

Abstract

Effective visual attention requires suppressing distracting information while prioritizing relevant content. Statistical learning of regularities in our environment can optimize this process. The present study investigated whether a previously learned suppression mechanism for high-probability distractor locations can transfer and be retrieved when these regularities change. In two experiments, using an additional singleton paradigm, we manipulated distractor location probabilities across different phases of visual search tasks. Experiment 1 examined whether suppression transfers when the high-probability distractor location changes, while Experiment 2 additionally tested whether relearning occurs when the original location is reinstated. Results showed that participants developed location-specific suppression, demonstrated by faster response times for distractors in high-probability versus low-probability locations. When these locations changed, the relative performance pattern transferred to new configurations. When the initial locations were reinstated in Experiment 2, participants were able to suppress the location with a greater magnitude, suggesting memory retrieval despite intervening experiences. We conclude that statistical learning in attentional suppression operates through adaptive mechanisms that balance stability and flexibility, with implications for understanding how attentional control optimizes performance in dynamic visual environments.