Isaac Savelson¹ (), Christopher Hauck², Mei-Ching Lien², Eric Ruthruff³, Andrew B. Leber¹; ¹The Ohio State University, ²Oregon State University, ³The University of New Mexico

Humans’ ability to resist attentional capture by distracting stimuli appears to rely on learning the specific features of a distractor over the course of repeated encounters with the item. A seminal study by Vatterott and Vecera (2012) found substantial attentional capture when first encountering salient distractor, followed by a complete attenuation of capture as experience with the color singleton increased. Critically, this pattern of capture-then-rejection was repeatedly observed throughout the experiment following changes in the distractor color. Despite the apparent robustness of this effect in the original study, we report three experiments in which we appeared to fail to replicate Vatterott and Vecera’s findings. Importantly, these experiments were conducted independently by two groups of researchers using much larger sample sizes than the original study. Using the conventional split-block analysis, we found no evidence for experience-based distractor rejection. That is, we found no rebound of attentional capture when presenting a novel distractor and were unable to detect a significant decrease in capture across block halves. Critically, the apparent replication failure did not appear to be due to a lack of distractor rejection; instead, minimal attentional capture was observed overall, leaving little room for further reduction. However, a finer-grained trial-by-trial analysis revealed that capture did occur at the start of each block, but complete distractor rejection appeared after only ~2 encounters with the distractor – much faster than Vatterott and Vecera originally reported. Further, rejection seemed to be partially robust to changes in distractor color: the initial capture in blocks 2-4 was greatly reduced relative to the first block. These results necessitate an update to our understanding and study of learned distractor rejection, especially regarding the speed at which distractor rejection is learned. Overall, we conclude that learned distractor rejection is more robust than initially believed but can be difficult to measure.

Acknowledgements: Funding: NSF BCS-2021038 to ABL