Data from: History-Driven Attention as a Distinctive Mechanism: Domain Specificity and Dynamic Characteristics

This dataset includes the data and analysis scripts of the experiments in the current research.Attentional selection has traditionally been attributed to goal-driven and stimulus-driven mechanisms. More recently, a third, history-driven mechanism has been proposed, underscoring the role of statistical regularities in guiding attention. However, whether this mechanism is domain-general and constitutes a distinct component of selective attention remains controversial. The present study investigated whether history-driven attention extends to perceptual and semantic regularities while delineating its distinctive characteristics. Using a dynamic probe detection task, we found a significant attention bias toward perceptually structured visual streams (i.e., periodic shape sequences) compared to random counterparts. By contrast, no such bias was evident for semantically structured streams (i.e., Chinese idiom sequences). Remarkably, the attention bias toward perceptual regularities emerged gradually, persisted despite environmental changes yet adapted after a delay, contrasting with the instantaneity of stimulus-driven selection. Furthermore, this attention bias arose spontaneously and implicitly, even when regularities were task-irrelevant and unnoticed by observers, with its pattern rather than strength altered by awareness and task relevance, distinguishing it qualitatively from goal-driven selection. These findings elucidate the domain-specific influence of statistical learning on attention deployment and establish history-driven selection as a distinct mechanism beyond the conventional goal- and stimulus-driven dichotomy of selective attention.