Introduction
The world around us is relatively stable and predictable over short periods, our experiences are thus useful to guide decisions because the past and the present often correlate. Large amounts of research using visual stimuli, including orientation, color, and motion direction \cite{Bae2020,Barbosa2020,Fischer2014} have demonstrated that our current perception is influenced by recent events. Fewer research focuses on time perception, and our understanding remains limited. Time perception is inherently subjective and susceptible to various factors, including attention, memory, sensory modality, psychophysical task, and temporal context \cite{Lapid2008,Penney2000,Shi2013,Wittmann2009,Treisman_1984}. For instance, the subjective duration can be biased toward recent history \cite{Burr2009,Jazayeri2010,Nakajima1992}, leading to an underestimation of long durations and an overestimation of short durations, known as the central tendency effect \cite{Glasauer2021,Hollingworth_1910}. The other phenomenon in time perception that is frequently studied is the sequential effect, which is different from the central tendency effect, for it refers to the influence of recent past stimuli on the perception of the current stimuli, rather than a tendency to the average or mean value of a set of stimuli. Only a handful of recent studies have explored trial-to-trial sequential effects on timing \cite{Glasauer2022,Togoli2021,Wehrman2020,Wiener2014}, and the mechanism underlying temporal sequential effects remains unclear based on the relatively limited research evidence.
The sequential effect, seen as carryover effects, can be categorized into sensory carryover and decisional carryover. Sensory carryover reflects how past stimuli influence current perception, often under the broader term “serial dependence” \cite{Fischer2014}. This influence manifests as assimilation or repulsion biases: assimilation makes successive stimuli seem more alike than they are, while repulsion biases the current percept away from the preceding one. Decisional carryover, on the other hand, is shaped by prior responses. Mechanisms underlying sequential dependence remain debatable. Some studies primarily focus on the preceding sensory input \cite{Cicchini2017,Fornaciai2018,Togoli2021}, while others attribute the sequential bias mainly to prior responses \cite{Feigin2021,Wehrman2020,Wiener2014}. For instance, research has identified behavioral or neural signatures of serial dependence effects without a response \cite{Czoschke2019,Fornaciai2018} or with a 'flipped' response \cite{Cicchini2017}, suggesting that the serial dependence is a perceptual mechanism independent of task-related responses \cite{Liberman2016}. In contrast, recent studies found that sequential bias strongly depends on the prior response \cite{Feigin2021}. For instance, serial dependence for motion direction is only observed when the preceding trials involved the same direction-report response rather than a different task response \cite{Bae2020}, underscoring the significance of post-perceptual, response-related factors \cite{Pascucci2019,Ranieri2022,Suárez-Pinilla2018}. Furthermore, \citet{Wiener2014} showed that participants tended to repeat their previous trial's decision, highlighting a decisional carry-over effect.
Many studies on the sequential effect have focused on scenarios involving a single type of stimulus. Typically, participants report a single target feature in most trials, but sometimes no response is required \cite{Czoschke2019,Fischer2014}. These studies showed serial dependence even without a response in prior trials, suggesting that a task-relevant response isn't essential for sequential effects. However, focusing on a single feature might blur the lines between perceiving and reporting it. The frequent need to report a target feature might prime participants to prepare responses, even when responses are not needed, potentially skewing the observed sequential dependence.