Statistical regularities bias memory decisions without enhancing working memory encoding: Insights from attribute amnesia.

While previous studies have shown memory enhancement for items with statistical regularities, it remains unclear whether this advantage persists when people are not anticipating the need to recall that information. Here, we used the attribute amnesia paradigm to examine whether statistical regularities influence working memory encoding in the absence of intentional memorization. In Experiment 1, participants reported the location of a colored target that appeared more frequently in one color. On a surprise trial probing target color, participants who saw the target in the frequent color were significantly more likely to answer correctly than those who saw it in a less frequent color. More importantly, regardless of which color was actually shown, participants across both groups tended to choose the frequent color as target color, suggesting a response bias, rather than enhanced encoding, driven by statistical regularities. Experiment 2 inserted a separate visual search task with equalized color probabilities and found an attentional bias toward the frequent color, confirming its attentional prioritization. Experiment 3 extended the above findings to task-irrelevant, yet physically salient and attention-grabbing distractors. Together, these findings indicate that although statistical regularities do not enhance working memory encoding, participants implicitly extract summary statistics of attended item attributes across trials, which in turn shapes their subsequent decisions.