In filled-duration illusion, a continuous (long) tone or an auditory sequence with multiple clicks is typically perceived as longer than the same physical duration (i.e., empty interval) enclosed by two auditory clicks. The auditory sequence is composed of multiple empty intervals. However, the individual empty interval in an auditory sequence, compared with the empty interval presented alone, could be biased in duration perception. In the current study, we implemented five experiments to reveal that the time perception of a single empty interval versus that of (mean) empty intervals in an auditory sequence depends on the task demands and contextual information. Specifically, we observed that the empty interval (140 ms) was perceived as longer than the same physical inter-stimulus interval in a sound sequence (Experiments 1 and 3). However, the empty interval (140 ms) was perceived as shorter than a continuous beep (i.e., filled duration of 140 ms) (Experiment 2). We observed a robust compression effect, in which the target empty interval (bounded by two oddball clicks) was perceived as shorter than the other physically equivalent intervals in a sound sequence (Experiment 4). In addition to the 'compression', perception of the target empty interval was assimilated by the other, task-irrelevant empty intervals in the sound sequence (Experiment 5). We explained the observed contextually modulated temporal illusions within a Bayesian inference framework.

中文翻译：

---

听觉持续时间的幻觉取决于任务和情境。

在持续时间错觉中，连续的（长）音或具有多次咔嗒声的听觉序列通常被认为比两个听觉咔嗒声所包围的相同物理持续时间（即空间隔）更长。听觉序列由多个空间隔组成。但是，与单独显示的空音间隔相比，听觉序列中的单个空音间隔可能会在持续时间感知中产生偏差。在当前的研究中，我们实施了五个实验，以揭示一个听觉序列中单个空间隔与（平均）空间隔的时间感知取决于任务需求和上下文信息。具体而言，我们观察到空间隔（140 ms）被认为比声音序列中相同的物理刺激间隔更长（实验1和3）。然而，空间隔（140 ms）比连续蜂鸣声（即140 ms的持续时间）短（实验2）。我们观察到了强大的压缩效果，在该效果中，目标空间隔（由两次奇怪的击球声限制）被认为比声音序列中其他物理等效间隔短（实验4）。除“压缩”外，声音序列中其他与任务无关的空间隔也吸收了对目标空间隔的感知（实验5）。我们解释了在贝叶斯推理框架内观察到的上下文调制的时间错觉。其中目标空间隔（由两次奇数球咔嗒声界定）比声音序列中其他物理等效间隔短（实验4）。除“压缩”外，声音序列中其他与任务无关的空间隔也吸收了对目标空间隔的感知（实验5）。我们解释了在贝叶斯推理框架内观察到的上下文调制的时间错觉。其中目标空间隔（由两次奇数球咔嗒声界定）比声音序列中其他物理等效间隔短（实验4）。除“压缩”外，声音序列中其他与任务无关的空间隔也吸收了对目标空间隔的感知（实验5）。我们解释了在贝叶斯推理框架内观察到的上下文调制的时间错觉。