Episodic memory allows individuals to mentally reconstruct past experiences in a temporally compressed form, such that remembering typically unfolds faster than the original event. This compression is not uniform: recent evidence indicates that negative events are remembered with less temporal compression than neutral ones, suggesting enhanced preservation of event unfolding. The present study investigated the role of encoding processes in this emotional memory advantage. In Experiment 1, we used a divided-attention manipulation to test whether the availability of attentional resources at encoding modulates the effect of emotion on temporal compression. Divided attention reduced overall memory quality—elevating compression rates and reducing detail and dynamism—but did not alter the emotional memory advantage. Experiment 2 examined whether negative emotion influences event segmentation. Results showed that emotion and event segmentation exerted independent effects on temporal compression. Experiment 3 collected online descriptions of unfolding events as a proxy for the amount and type of information attended to during encoding. Results showed that participants attended to comparable amounts of information during negative and neutral events. However, similarity analyses revealed that the correspondence between online descriptions and memory narratives from Experiment 1 decreased for neutral videos encoded under divided attention, whereas it remained stable for negative videos. Together, these findings indicate that negative events produce richer and less compressed memory representations through mechanisms that operate beyond attentional allocation and event segmentation, potentially involving rapid replay or early consolidation processes at event offset.

Mind-wandering (MW) refers to task-unrelated thoughts that are decoupled from the surrounding environment. Although often considered disruptive, MW has been proposed to support functions such as creativity, problem solving, and planning, which presuppose that its content is encoded in memory and later retrieved when relevant. Yet, factors determining whether MW episodes are remembered or forgotten remain poorly understood. To address this gap, participants completed a 10-min Think-Aloud Protocol, during which they continuously verbalized their thoughts, followed by an unexpected free-recall test after a one-day delay. After recall, participants rated the phenomenological characteristics of each MW episode, allowing direct comparisons between recalled and forgotten episodes to identify content-related and temporal predictors of recall. Our results showed that recall probability increased with episode length and for episodes involving other people, rated as more important, goal-related, structured, deliberate, routine, and recurrent in daily life. Recall probability was also higher for episodes with temporal and functional future orientation, whereas episodes with no apparent function were less likely to be recalled. In a combined model including all significant predictors, only episode length, involvement of other people, future temporal orientation, and absence of function uniquely predicted recall. In addition, recall exhibited primacy and recency effects, as well as a tendency to follow the original chronological order of MW episodes. Together, these findings indicate that memory for MW episodes is not random but rather shaped by their phenomenological characteristics and temporal dynamics, suggesting that it follows organizational principles similar to those observed for externally perceived stimuli.

Episodic memories are temporally compressed summaries of past experience, with the degree of compression varying with event characteristics. This study examined how two such characteristics—environmental familiarity and action involvement—influence temporal compression, objective recall content, and subjective memory quality. Across three experiments, participants viewed first-person videos of navigation (walking) in familiar or unfamiliar places, with goal-directed action videos (e.g., buying a notebook) also included in two of the experiments. Participants then mentally replayed each video to derive a compression measure, alongside subjective ratings and recall descriptions. Action involvement consistently reduced compression and enriched both subjective and objective aspects of remembering. Environmental familiarity effects on compression were context-dependent: absent when videos were recalled immediately after viewing (Experiment 1), present — with unfamiliar videos remembered in less compressed form — when familiarity was blocked across separate lists (Experiment 2), and eliminated when familiar and unfamiliar videos were mixed within the same lists (Experiment 3). Familiar videos were nonetheless consistently rated as more detailed and easier to remember, independent of list composition. Overall, environmental familiarity shaped the subjective experience of remembering without altering objective recall content, revealing a partial dissociation between subjective and objective memory measures, whereas action involvement consistently benefited both, underscoring the structuring role of actions in event memory. These results are discussed in relation to schema-based encoding, working-memory resource, and scene-construction accounts of temporal compression.

Narrative identity hinges on the construction of a coherent life story. While the cognitive and neural bases of past-oriented narratives are well documented, much less is known about their future-oriented counterparts. This study investigates whether autobiographical reasoning—defined as reflecting on the personal meaning of events—operates similarly or differently depending on temporal orientation. In Study 1, we assessed the narrative content and functions of autobiographical reasoning. For both past and future events, autobiographical reasoning involved linking represented events to other personal experiences and characteristics. However, past-oriented reasoning relied more on personal characteristics, whereas future-oriented reasoning included more references to personal goals. In Study 2, participants evoked the concrete content of past and future self-defining experiences, or engaged in autobiographical reasoning for the same events, during fMRI scanning. Autobiographical reasoning showed a consistent neural signature across temporal orientations, involving the left dorsomedial prefrontal cortex, inferior frontal gyrus, and lateral temporal cortex. These findings highlight both shared and distinct aspects of autobiographical reasoning across the past and future. Overall, autobiographical reasoning appears to draw primarily on semantic processing and conceptual self-knowledge, enabling individuals to derive meaning and value from personal events by linking them to other experiences, goals, and traits.

A hallmark of the human mind is its tendency to generate spontaneous thoughts, whether during tasks or in idle moments. This phenomenon is typically studied in the laboratory using the Thought-Probe Protocol (TPP), in which participants report the content of their thoughts when prompted at various intervals. Although well validated, the TPP nonetheless suffers from several limitations, such as its inability to track the flow of thoughts between probes. To address these issues, researchers have recently revisited the Think-Aloud Protocol (TAP), which involves the continuous verbalization of spontaneous thoughts. While the TAP offers access to the ongoing flow of thoughts, its validity relative to other methods has not yet been fully established. In this study, we compared four methods for assessing spontaneous thoughts: the TAP, TPP, Daily Life Experience Sampling Protocol (DLESP), and retrospective thought listing. We focused on the phenomenological characteristics of thoughts and features that predicted their recall after a one-day delay. Our results revealed minimal differences between the TAP and TPP in terms of thought characteristics and memory predictors. However, thoughts reported with these two methods differed from those assessed more ecologically with the DLESP, and certain thought features were overrepresented in retrospective thought listing. Overall, our findings suggest that the TAP is as valid as the TPP for investigating spontaneous thought, although thought characteristics may differ between laboratory and real-world settings. They also suggest that concurrent reporting methods, such as the TAP and TPP, provide a more representative view of spontaneous thought features than retrospective assessments.

Remembering the unfolding of past episodes usually takes less time than their actual duration. Recent work measuring the time required to mentally replay videos suggests that such temporal compression emerges when continuous events become too long to be fully held in working memory. However, in this previous research, the time needed to initiate the mental replay (i.e., to mentally visualize the starting point of events) was not taken into account. In the current study, we developed a corrected measure of remembering duration that controls for this initiation time to better characterize working memory capacity in representing continuous event. Using this corrected measure, we assessed the time needed by participants to mentally replay videos depicting continuous events (e.g., turning a car jack) lasting 3, 6, 9, 12, or 15 s. The results showed that events longer that 3-s were remembered in less time than their actual duration, suggesting that working memory is limited in its capacity to represent continuous events, leading to the emergence of temporal compression when they exceed 3 s. Exploratory analyses further revealed that faster mental replay tended to be judged less dynamic but not less clear, suggesting that the temporal compression of continuous events likely results from an incomplete remembering of event’s unfolding rather than a complete but imprecise mental replay.

Naturalistic events are represented in episodic memory as sequences of experience units (i.e., segments of prior experience) that are separated by temporal discontinuities, so that events are temporally compressed during memory replay. In the current study, we aimed to shed light on the interplay between the segmental structure of events and working memory (WM) resources in shaping this temporal structure of episodic memories. Participants watched a series of 1-min videos that contained many or few event boundaries (EBs), while for half of the videos, a concurrent task was simultaneously performed to reduce the availability of WM resources. After each video, the time needed to mentally replay the events was measured, and then the content of memories was verbally described. Results showed that the number of recalled experience units was higher for videos that contained many EBs, leading to a lower temporal compression rate. In addition, the number of recalled experience units was lower when a concurrent task was performed during event perception, particularly for videos that contained many EBs. These results suggest that both the segmental structure of events and WM resources contribute to the formation of experience units that constitute episodic memories, thereby shaping the temporal compression of events.

When recalling real-world events, people typically remember a sequence of key moments rather than a continuous stream, often omitting portions of their previous experience. It remains unclear whether such omissions reflect gaps in memory encoding or whether the corresponding moments are available in memory but not accessed during retrieval. To investigate this, the present study assessed recognition memory for recalled versus omitted segments. Participants walked around their university campus while wearing eye-tracking glasses that recorded their experience. Twenty-four hours later, they freely recalled the events and completed a recognition task, discriminating between 5-s video clips from their own walk and those from other participants. Recognition accuracy was lower for unrecalled than for recalled moments, but nevertheless above chance. A second experiment replicated these results and tested whether overlaying participants’ original eye movements on the clips during recognition would enhance performance—it did not. These results suggest that omissions in the recall of events result from both encoding and retrieval processes: while some moments may not be stored, others are available but not accessed during recall. We discuss how the dynamics of event perception and memory reconstruction contribute to the selective recall of real-world experiences.

Recent studies have revealed that the continuous flow of information that characterises naturalistic events is temporally compressed in episodic memory, so that remembering an event generally takes less time than the duration of the past episode. However, the specific characteristics of an event that influence its temporal compression in memory remain poorly understood. In the present study, we examined the extent to which the negative valence of events impacts their rate of compression in memory representations. We conducted two experiments in which participants were instructed to mentally replay a series of videos depicting negative or neutral events. The results showed that the time taken to mentally replay a video, relative to the actual video duration, was significantly longer for negative than for neutral videos. These results suggest that negative emotion increases the sampling rate of units of experience that represent the course of events, leading to a lower compression of events in memory representations.

How do people mentally replay real-life events, and what shapes the time it takes to remember them? In this study, we investigated the temporal compression of memories by examining how long it takes participants to recall everyday events they recorded using wearable cameras. While remembering duration increased with the actual length of events, this relationship was nonlinear: recall duration rose steeply for events lasting up to ~10 min, then plateaued, suggesting scale-invariant retrieval beyond this threshold. Crucially, various event characteristics also influenced remembering duration, with events that were more unusual, unpredictable, emotionally positive, socially engaging, or marked by greater change showing less temporal compression. These effects were not explained by retrieval difficulty, but rather reflected the richness of memory representations, including greater detail and stronger sense of reliving. Together, these findings suggest that memory compression depends not only on the event’s actual duration, but also on how it was subjectively experienced and structured in memory. By linking event features to the tempo of recall, this study offers novel insight into the dynamics of episodic memory and the mechanisms that shape how we mentally replay real-life experiences

This study examined the influence of time and self-relevance on collective memories. Participants recalled their memories of two events from 2020 that differed in self-relevance: the COVID-19 pandemic and a political event. Furthermore, each event was recalled at two time points: in 2021 and 2022. In 2021, participants also imagined a future pandemic and a future political event (the dissolution of the EU) to assess the extent to which representations of the collective future rely on representations of the collective past. Given the wide and complex nature of collective memory, we measured the type of information people remembered and imagined (personal vs. collective) and the common themes discussed by participants. Results showed that, in 2021, people recalled more personal information about the pandemic than the political event. By 2022, pandemic memories were recalled with more collective than personal information, similarly to political event memories, which remained primarily collective. For imagined future events, participants reported more collective than personal elements. Moreover, the themes described when imagining a future pandemic were similar to the ones recalled about the past pandemic. Themes about the pandemic also evolved into a more comprehensive view of the events over time. Overall, the findings suggest that the content of collective memories, especially self-relevant ones, evolves over time, emphasizing the constructive nature of collective memories and how it shapes collective future thinking.

Remembering past events usually takes less time than their actual duration - events are temporally compressed in memory. A recent study found that this compression is not systematic but emerges when continuous events exceed approximately 9 s. Unexpectedly, however, remembering shorter events (3-6 s) took more time than their actual duration. Here, we aimed to investigate the mechanisms behind this increased replay duration of short events. In Experiment 1, we developed a corrected measure accounting for recall initiation time - the time needed to access the beginning of the event. With this correction, the longer replay times for short events disappeared, suggesting the effect was partly due to unmeasured recall initiation time. In Experiment 2, we examined the potential role of a central tendency bias by exposing participants to different ranges of event durations. Replay duration was influenced by the event's relative position within the duration range, consistent with a central tendency bias. However, for events longer than 9 s, temporal compression occurred consistently across all conditions. Together, these findings suggest that while central tendency influences replay duration, temporal compression systematically emerges when events exceed a few seconds, likely reflecting memory capacity limits in representing continuous experiences.

Remembering the unfolding of past experiences usually takes less time than their actual duration. In this study, we examined the extent to which this temporal compression in memory depends on the number and duration of events that need to be maintained in a sequence. Participants were asked to watch and then mentally replay short videos depicting one, two, or three continuous events (i.e., people performing continuous actions in an uninterrupted way), each lasting 3, 6, 9, or 12 s. Across two experiments, we computed indices of remembering duration and temporal compression for each event. Results showed that event remembering duration was close to the actual event duration for short events, but smaller for longer ones (i.e., temporal compression was not systematic but occurred selectively depending on event duration). Furthermore, events were mentally replayed more quickly when they were part of a sequence of several events than when they were presented alone, and this decrease in the duration of event recall with the number of events was more pronounced for longer events. Exploratory analyses revealed that individual differences in memory compression were predicted by visual imagery capacity. These results suggest that working memory capacity in representing naturalistic events is limited by both the number and duration of events to be retained, which may in part explain why the unfolding of events is temporally compressed in episodic memory.

We live in uncertain times and how this pervasive sense of uncertainty affects our ability to think about the future remains largely unexplored. This study aims to investigate the effects of uncertainty salience on episodic future thinking—the ability to mentally represent specific future events. Experiment 1 assessed the impact of uncertainty on the accessibility of episodic future thoughts using an event fluency task. Participants were randomly assigned to either an uncertainty induction or control condition, and then were asked to imagine as many future events as possible that could happen in different time periods. The results showed that participants in the uncertainty condition produced fewer events, suggesting that uncertainty salience reduced the accessibility of episodic future thoughts. Experiment 2 investigated in further detail the mechanisms of production of episodic future thoughts that are affected by uncertainty. The results showed that uncertainty primarily reduced the accessibility of previously formed future thoughts (i.e., memories of the future) rather than affecting the ability to generatively think about the future and construct events. These findings shed new light on the impact of uncertainty on episodic future thinking, paving the way to further investigation into its implications for decision-making and future-oriented behavior.

While previous studies have highlighted the role of episodic future thinking in goal pursuit, the underlying cognitive mechanisms remain unexplored. Episodic future thinking may promote goal pursuit by shaping the feeling that imagined events will (or will not) happen in the future – referred to as belief in future occurrence. We investigated whether goal self-concordance (Experiment 1) and other goal characteristics identified as influential in goal pursuit (Experiment 2) modulate belief in the future occurrence of goal-related events and predict the actual occurrence of these events. Results showed that goal self-concordance, engagement, and expectancy had an indirect effect on the actual occurrence of events, which was (partially) mediated by belief in future occurrence. The mediating role of belief supports the view that belief in future occurrence when imagining events conveys useful information, allowing us to make informed decisions and undertake adaptive actions in the process of goal pursuit.

Remembering the unfolding of past episodes usually takes less time than their actual duration. In this study, we evaluated whether such temporal compression emerges when continuous events are too long to be fully held in working memory. To do so, we asked 90 young adults to watch and mentally replay video clips showing people performing a continuous action (e.g., turning a car jack) that lasted 3, 6, 9, 12, or 15 s. For each clip, participants had to carefully watch the event and then to mentally replay it as accurately and precisely as possible. Results showed that mental replay durations increased with event duration but in a non-linear manner: they were close to the actual event duration for short videos (3-9 s), but significantly smaller for longer videos (12 and 15 s). These results suggest that working memory is temporally limited in its capacity to represent continuous events, which could in part explain why the unfolding of events is temporally compressed in episodic memory.

Remembering the unfolding of past experiences usually takes less time than their actual duration. In this study, we examined the extent to which this temporal compression in memory depends on the number and duration of events that need to be maintained in a sequence. Participants were asked to watch and then mentally replay short videos depicting one, two, or three continuous events (i.e., people performing continuous actions in an uninterrupted way), each lasting 3, 6, 9, or 12 s. We computed remembering duration and a temporal compression index for each event. Results showed that event remembering duration was close to the actual event duration for short events, but smaller for longer ones (i.e., temporal compression was not systematic but occurred selectively depending on event duration). Furthermore, events were mentally replayed more quickly when they were part of a sequence of several events than when they were presented alone, and this decrease in the duration of event recall with the number of events was more pronounced for longer events. These results suggest that working memory capacity in representing naturalistic events is limited by both the number and duration of events to be retained, which may in part explain why the unfolding of events is temporally compressed in episodic memory.