Predictability associated with an event influences its perceived time. The two forms of predictions that are often discussed and have a dissociable influence on perceived time are repetition and expectation. However, predictions based on expectation can be seen at multiple levels, potentially leading to an inconsistency in the pattern in which expectation influences perceived time. Therefore, the present study aimed to investigate how different levels of predictions impact perceived time. In two separate experiments utilizing visual and auditory stimuli, we used a hierarchical violation-of-expectation paradigm that can dissociate two types of predictions based on local and global rules. Results from analysis of variance computed with local and global predictions revealed a pattern of local and global predictions having a distinct influence on perceived time. More specifically, while the local predictions that consider the immediate stimulus exposure reduced the perceived time, the global predictions that consider the overall regularities of a given context increased the perceived time. These results integrate well with the recent theoretical models rooted in a predictive coding framework that emphasizes the opposing effects of the first order and second order predictions on perceived time.

与事件相关的可预测性会影响其感知时间。经常讨论并对感知时间产生不可分离影响的两种预测形式是重复和期望。然而，基于期望的预测可以在多个层面上看到，这可能导致期望影响感知时间的模式不一致。因此，本研究旨在调查不同级别的预测如何影响感知时间。在利用视觉和听觉刺激的两个独立实验中，我们使用了分层违反预期范式，该范式可以根据局部和全局规则分离两种类型的预测。用局部和全局预测计算的方差分析结果揭示了局部和全局预测模式对感知时间有明显影响。更具体地说，虽然考虑直接刺激暴露的局部预测减少了感知时间，但考虑给定上下文的整体规律性的全局预测增加了感知时间。这些结果与最近植根于预测编码框架的理论模型很好地结合在一起，该框架强调一阶和二阶预测对感知时间的相反影响。