Abstract
There is growing consensus that selection history modulates attentional priority and is distinct from current goals and physical salience. Here, we focus on one of the selection history phenomena, statistical learning (SL). Attentional guidance and suppression induced by SL is thought to be long-lasting, implicit, and inflexible. We examined these claims with regards to SL of target location (4 experiments, total n=146) and distractor color (1 experiment, n=160). We measured search slopes as an index of attentional guidance and distractor interference as an index of distractor suppression. To determine whether SL is long-lasting, we used the two-phase learning-extinction design where a probability imbalance was present in the learning but absent in the extinction. Unlike past research that relied on one-shot awareness test to determine whether SL is implicit, we examined the correlation between the magnitude of SL and participants’ probability estimates. To increase sensitivity, we tested awareness both early (immediately after learning) and late (after extinction). Finally, to determine whether SL is inflexible, we used the between-group comparison where one group was informed of the transition between the two phases of the experiment (and even given an alternative source of attentional guidance) and the other group was not. We found attentional guidance induced by SL of target location to be long-lasting, implicit, and inflexible. However, though suppression induced by SL of distractor color was also found to be implicit, it was short-lived.