Li, Sha2019-08-202019-08-202019-05https://hdl.handle.net/11299/206234University of Minnesota Ph.D. dissertation. May 2019. Major: Psychology. Advisor: Yuhong Jiang. 1 computer file (PDF); x, 155 pages.Extensive research has shown that prior experience and selection history modulate visual selective attention. Humans are able to learn various types of statistical regularities and use them to optimize allocation of attention. For example, people typically respond faster to visual properties that are predictive of important stimuli (e.g., rewards and search targets) in the past. However, much less is known about whether established attentional modulation persists when the task or visual statistics change. This dissertation aims to understand the effects of visual statistical learning on attention and how such learning adapts to changes in the environment. The first study focuses on reward learning. The results show that participants respond faster to targets that lead to higher reward. But when presented as distractors in a subsequent task, previously high-reward targets do not capture more attention. These findings suggest that monetary reward enhances attentional priority for high-reward targets, but the enhancement dissipates when the prioritized items become distractors in a different task. The second study shows that people are sensitive to occurrence rates of visual features such as color. Participants respond faster to colors that are more frequently associated with the target. However, when colors become equally frequent, people do not always continue prioritizing the previously high-frequency colors. Changes in attentional priority depend on the type of statistical regularities people have learned. The third study examines transfer of the location probability effect across different tasks. Behavioral and eye tracking data suggest that participants prioritize spatial locations that are more likely to contain the target. However, established spatial priority does not transfer to a novel task that requires a different oculomotor search procedure. This dissertation demonstrates that various types of statistical learning affect visual selective attention. However, constraints exist in whether the learning-induced attentional modulation persists in a dynamic environment.enFeatural FrequencyLocation Probability CueingSelective AttentionVisual SearchVisual Statistical LearningThesis or Dissertation