Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2023 Oct 2.
Published in final edited form as: Curr Dir Psychol Sci. 2023 May 6;32(4):284–292. doi: 10.1177/09637214221143053

Psychological Momentum

Christopher J Honey 1, Abhijt Mahabal 2, Buddhika Bellana 3
PMCID: PMC10545321  NIHMSID: NIHMS1851399  PMID: 37786409

Abstract

Our mental experience is largely continuous on the scale of seconds and minutes. However, this continuity does not always arise from a volitional carrying forward of ideas. Instead, recent actions, thoughts, dispositions, and emotions can persist in mind, continually shaping our later experience. Aspects of this fundamental property of human cognition – psychological momentum – have been studied under the rubrics of memory, task set, mood, mind-wandering, and mindset. Reviewing these largely independent threads of research, we argue that psychological momentum is best understood from an integrated perspective, as an adaptation that helps us meet the current demands of our environment and to form lasting memories.

Our thoughts are connected, not only on the scale of seconds as one thought flows into another, but also on the scale of many minutes, as recent thoughts and associations return to mind. For example, we might review relevant material to ensure it is “fresh in mind” before an interview or test. But the persistence of mental content does not require such volition. We may struggle to focus on a work meeting if an earlier personal conversation is “stuck in our head” or is “on our minds”. Similarly, we may read a story or watch a movie and then find that themes, characters, or moods effortlessly linger in our minds after the narrative ends (Figure 1). Readers have coined the term “book hangover” to describe the lingering immersion that shapes their thoughts and feelings for minutes (or days) after finishing a novel (v1ncetta, 2017).

Figure 1.

Figure 1.

Open in a new tab

Our thoughts and actions at one moment (e.g., watching a movie; left panel) effortlessly persist to shape thoughts and actions minutes later, even when those former ideas are not obviously relevant to the later context (e.g., outside of the movie theatre; right panel). Illustrated by Rita Terra (@ritzz_ritzz_).

What is shared across these examples? We posit that the common thread is a fundamental property of human psychology: that contentful mental states can effortlessly persist and return to mind for several minutes or more, even in the absence of external cues or overt goals. We will refer to this phenomenon as “psychological momentum”, and ask: does it have an adaptive or rational explanation? How can it be prevented or magnified? After reviewing research on memory, task-switching, and mind-wandering, we sketch the outlines of a model that could generate testable predictions.

In a recent study, we observed psychological momentum by measuring the content of spontaneous thought after reading a story (Bellana et al., 2022). Each participant in our study generated free associate word-chains, both before and after reading a 2,000-3,000 word narrative (Figure 2A). General story themes (e.g. “death” or “love”) and specific narrative entities (e.g. “river” or “spy”) were over-expressed in free-associate chains generated after reading each story, as compared to before (Figure 2B). The lingering phenomenon increased amongst participants who felt “transported” by the text that they read; the lingering was reduced amongst those who read a version of the story with sentence-order randomly scrambled; and the lingering was eliminated in participants who read the story’s words in a scrambled order in the context of a verbal working memory task (Figure 2C). Finally, most participants reported that the narrative lingering occurred without volition, with many describing it as an unwanted interference in their subsequent thought (Figure 2D). For example, one participant wrote: “I think I almost tried to not use words/themes that were in the text as I didn’t want to have been influenced by the text. I realised I was coming up with blanks/dead ends a little more the second time as I didn’t want to go towards darker themes or water based themes [from the story]”; another participant wrote: “In the first round, the words I typed were considerably more organic than those in the second round, as I could not really get the story out of my head after reading it, so many of the associations were related to extraneous thoughts or associations with the story itself.”

Figure 2. Immersive stories linger in mind.

Figure 2.

Open in a new tab

(A) Schematic of the experimental paradigm from Bellana et al., (2022). Online participants performed free association (i.e., typing whatever words came to mind, as they came to mind) for 5-minutes before and after reading a short story. To manipulate the coherence of the story, participants were randomly assigned to read an intact version, a version with sentence-order randomly scrambled, or a version with word-order randomly scrambled. Participants in the word-scrambled condition were intermittently probed with a verbal working memory task concerning the sentence they had just read. (B) Specific themes and words from the story were over-expressed in post-story free association. So Much Water So Close to Home is about a wife who suspects her husband may have murdered a young girl while on a fishing trip. Roy Spivey was about a woman who sits beside a celebrity (a famous actor in spy movies) on an airplane. The two find they like one another, and the celebrity shares his phone number – withholding one number that he asks she commits to memory: four. The ‘overexpression’ of a word in free association, or bias, was calculated by taking the proportion of post-story free association chains that contained a given word [p(Post)] minus the proportion of pre-story free association chains containing the same word [p(Pre)]. Positive values reflect words that are more likely to occur in post-story free association as compared to pre-story. Negative values reflect words that are more likely to occur in pre-story free association as compared to post-story. Asterisks denote theme words that showed the strongest semantic similarity with participant-generated story themes for each story, calculated using word embeddings. For more details about how ‘theme similarity’ was calculated, please see Bellana et al. (2022). For legibility, only free associates that occurred in at least 16% of free association chains or showed a 10% bias for pre-or post-story are displayed. Larger plotted points had a larger value of p(Pre). (C) The extent to which an individual feels as if the story they read ‘lingers’ in their mind after reading it is strongly predicted by the extent to which they found themselves ‘transported’ into the world of the story. Each point represents one participant. Data are collapsed across all stories in Experiment 1 from Bellana et al., (). (D) A separate group of participants (n=240) performed this paradigm and were asked to describe the intentionality of lingering, if they experienced any. Most participants endorsed lingering as unintentional. Figure panels adapted from Bellana et al. (2022).

With these results in mind, a working definition of psychological momentum is: the lingering of contentful mental states over several minutes, even in the absence of external cues or goals. By “contentful” we mean that a lingering mental state must be related to specific features of a recent experience (e.g., thoughts of “spying” and “spies” after reading a story involving these topics).

Why is psychological momentum worthy of study? Consider that each waking moment of our lives is embedded in a stream of thought, whose content shapes what we think about next. As human thought is marked by this history-dependence, psychological momentum implies that the degree of history-dependence is not fixed. Instead, some inputs have the potential to shape our subsequent thought and behavior more than others (Faber & D’Mello, 2018).

Characterizing the kinds of inputs, and ways of thinking, that give rise to this momentum will be critical for any complete understanding of human cognition. Moreover, because the persistence of negative mental content (e.g., rumination) is a hallmark of post-traumatic stress, depression and anxiety disorders (Spinhoven et al., 2018), a model of how such processes arise and can be modulated will be relevant to advances in mental health. On the other hand, mental persistence can also be beneficial, enabling us to find creative solution to open-ended problems, even when we are ostensibly focused on a separate task (Gable et al., 2019). If we wish to develop intelligent machines that can learn and solve open-ended problems, it may help to understand how humans revisit their recent thoughts and experiences (Wittkuhn et al., 2021).

We intend the term “psychological momentum” to operate by rough analogy with physical momentum. Though this terminology is imprecise, we hope it captures the intuitive notion that thoughts can possess direction (toward a particular question or region of semantic space) and intensity (enabling them to resist control and to out-compete other thoughts). When a ball is set in motion in a direction, it acquires momentum and continues to move in the original direction after the initiating force is removed. Similarly, when we are cued to consider a particular idea or event, we continue to consider that idea or event even after the initial cue is removed. And just as the momentum of a ball is increased when its mass or velocity is increased, psychological momentum is increased when the number of thoughts or the intensity of each thought is increased in a particular direction.

Below, we will argue that psychological momentum is under-studied and that understanding it requires the integration of psychological, neuroscientific and computational ideas. First, we ask how psychological momentum can be understood in relation to existing ideas about mindsets, memory, task-sets, and mind-wandering. Second, we consider the functional consequences of psychological momentum, and the key ingredients required for a testable process-level model of this aspect of our thinking.

Momentum as a mindset

Can psychological momentum be understood as an induced “mindset”? It appears that mindsets can be induced, and they do persist over time. For example, Herz et al. (2020) marshaled a large literature to argue that we continually shift between “narrow” and “broad” states of mind, in which our perception varies between sensory-driven and expectation-driven; our attention from global and local processing; our thoughts from expansive to constrained; and our moods from positive to negative. Similarly, in the domain of memory, instructing people to recall and describe precise details can induce a specific mode of recollection, which lasts for minutes and affects the divergence of subsequent thought (Madore et al. 2015). Finally, it is also possible to induce affective states that linger over time, in turn shaping the affective quality of spontaneous thought (Andrews-Hanna et al., 2021), and the way that unrelated material is encoded into memory (Tambini et al., 2017).

Although taking on a mindset can be a part of our experience of psychological momentum, a mindset (e.g., broad vs. narrow) does not involve the persistence of specific content from a past event. After two children see their favorite fantasy action movie in the theater (Figure 1, left), they may run around the lobby, happy and energetic, and with an expansive mindset. However, psychological momentum is meant to capture not only this induced mindset, but also how the specific episodic and conceptual content of the film persists in mind. For example, the children may repeat their favorite character’s catchphrase; they may find themselves thinking about bravery, about shields, about the brilliant crimson of dragon-fire; they may spend time imagining what it is like to possess magical powers; and they may play-act counterfactual variations of their favorite scenes (Figure 1, right). In sum, although mindsets and moods do persist and are a part of any model of psychological momentum, we must look to further neural and psychological processes to understand how specific events and concepts spontaneously persist in mind.

Momentum as memory

Because psychological momentum involves the persistence of specific mental content (i.e., thoughts, memories, emotions, and dispositions) over time, it is natural to think of it as a form of memory. Should we understand it as a manifestation of working memory, long-term memory or another mnemonic process?

If we define working memory as the maintenance of a limited amount of information in a state of heightened availability for ongoing processing (Cowan, 2017), then working memory appears to be an appropriate repository for lingering mental content. However, verbal and visual working memory are highly capacity limited and sensitive to interference (Oberauer et al., 2016), which does not seem to be compatible with detailed episodic content that lingers for minutes and forces itself to mind. In fact, our experiments with narrative lingering revealed little lingering of content when the words constituting the story were processed in a verbal working memory task; instead, the lingering effects were most prominent when participants were able to extract situation-level meaning from the words (Bellana et al., 2022). Concretely, this means that if a participant read the word “bread” and focused on its rhyming properties, then the notion of bread would not linger; if they focused on the semantics of the word, then it might linger in mind somewhat; but if they understood bread as the object of desire of a hungry family, then the notion of bread was very likely to linger in mind.

Because processing information in a “deep” and meaningful way is important for psychological momentum, we should consider how neural and cognitive memory processes are modulated by meaningfulness. Cognitively, Craik & Lockhart (1972) introduced a “levels of processing” framework which distinguished deep from shallow processing, and proposed that memory is improved when participants attend to what an input means rather than its surface-level physical properties. More recently, we have learned that situationally meaningful information is represented in a large-scale brain network that includes the medial prefrontal cortex, the posterior cingulate cortex/precuneus and the angular gyrus. These high-level brain regions, generally known as the default mode network (DMN), are thought to represent and simulate “scenarios”, composed of agents (self and other), the goals and outcomes of those agents, and the environments in which they interact. Within the “process memory” framework, DMN regions are thought to spontaneously integrate past and present information over tens of seconds, up to minutes of time (Hasson, Chen, & Honey, 2015). Thus, psychological momentum could arise from processing within DMN regions, as new information spontaneously interacts with, and is contextualized by, our recent past.

The processes underlying psychological momentum may be intertwined with those supporting memory consolidation. It has been known for more than a century (Müller & Pilzecker, 1900) that after we study a series of sounds or images (such as pairs of syllables) they exhibit a “perseveration tendency”, spontaneously re-entering awareness for several minutes. If we perform another task during this period of perseveration, our memory for the original materials (syllable pairings) is impaired. These seminal observations motivated the idea that information is consolidated into more durable memory in the minutes immediately following the original experience, while the information is perseverating (Dewar et al., 2007).

Advances in our understanding of memory consolidation continue to support a potential link between consolidation and momentum. For example, the consolidation of memories into durable storage is thought to rely on memory “reactivation” in association with high frequency rhythmic events (“sharp-wave ripples”) in the hippocampus (Buzsaki, 2015; Wittkuhn et al., 2021). Importantly, the spontaneous reactivation of the neural ensembles associated with past experiences (often referred to as “replay”) is not limited to sleep: brain states are reactivated in the awake human brain for minutes after an original experience (Higgins et al., 2021; also see potential benefits of post-encoding quiescence on memory: Humiston et al., 2019). Moreover, these replay events may not only assist in consolidation and generalization of memories, but may also involve reinstatement of lingering mental contexts (Howard & Kahana, 2002), which can bias subsequent decision making (Bornstein & Norman, 2017; Mattar & Daw, 2018; Schuck & Niv, 2019) and might also return previous ideas into our ongoing stream of thought.

It is unclear whether psychological momentum requires the hippocampus. One hippocampal amnesic participant was able to update and integrate episodic information in DMN regions on the scale of a minute (Zuo et al., 2020). Therefore, there may be reverberatory processes within the DMN that can support lingering episodic contexts in the absence of the hippocampus. However, it is also likely that “process memory” in real world settings involves a continual interplay between DMN regions and episodic memory systems, such that particular episodic memories are cued by high-level situational contexts, and the retrieved information from episodic memory serves to refresh and update the context (Howard & Kahana, 2002). Therefore, measuring lingering mental contexts in amnesic participants will provide critical constraints on theories of psychological momentum.

Thus, on the scale of seconds to minutes, psychological momentum may be mediated by the continual integration and reverberation of episodic information within long-timescale cortical regions of the DMN. On the timescales of minutes and hours, the same processes that mediate memory consolidation (i.e. tagging and hippocampal replay of memories) may lead to spontaneous re-entry of information into DMN systems.

Momentum of task-sets and goals

In parallel with the literature on human memory, scientists studying goal-directed thinking and task switching have measured how earlier tasks affect later behavior. A consistent finding is that people are slower and less accurate in task performance immediately after the switch from one task to another. These “switch costs” have been attributed to two sources: first, “task set reconfiguration” (i.e., the time required for control processes to establish the mental contingencies to perform the new task) and, secondly, “task-set inertia” (i.e., the interference of prior input-output contingencies with new ones) (Wylie & Allport, 2000). Task-set inertia denotes our tendency to carry forward a specific disposition to act, cognitively or motorically, which resonates with the notion of psychological momentum. However, the switch costs reported in the literature decay within a few trials following the switch, and tasks are not usually thought to carry forward specific mental content (i.e., thoughts, memories, and emotions). Still, perhaps task-set inertia could apply to persistent “internal” tasks that are not yoked to exogenous cues, but instead to our persistent goals as self-interested agents (e.g. deriving coherence; finding affection; preserving our self-image).

Of special importance to models of psychological momentum is the proposal that spontaneous human thoughts center on (and return to) a collection of “current concerns” or not-yet-achieved goals (Klinger, 1978). Klinger argued that an individual’s current concerns (e.g., working towards a career in health care) can be understood to upweight the features of our external environment or internal thoughts that are relevant to this goal (Klinger & Cox, 2011). These goal-relevant features are then more likely to be noticed, trigger related thoughts or dreams, or be remembered. This pioneering work addresses the question of why some experiences are more likely to induce psychological momentum than others (i.e., personal goal-relevance), in the tradition of long-standing theories from Gestalt psychology concerning the resolution of psychic tension and “unfinished business” (e.g. Zeigarnik, 1928). This framework could provide the basis of a model of psychological momentum, but it faces difficulties in concretely predicting what content can and will linger. For example, to explain why semantic content lingers after reading a short story, we need to be able to concretely specify the “concerns” that will become post-story thoughts: are they the concerns for some particular characters in the story? Or are they the reader’s concern for closure and understanding? Or might they be non-specific semantics from the narrative?

Momentum as Mind-Wandering and Rumination

The literature on mind-wandering and rumination has also proposed accounts of how past thoughts influence those of the future (Christoff et al., 2016). For example, Amir & Bernstein (2021) have modeled the trajectories of cognitive states as arising from transactions between working memory, emotion, and internal/external attentional orientations. Although such models do not predict precisely what will linger in mind, they can explain how patterns of thought can recursively interact, so that negative affect at one moment can lead to selfreinforcing patterns of ruminative negative thinking (Andrews-Hanna et al., 2021). But rumination may not be limited to basic affective biases and semantic associations: Bargh (2011) and Dijksterhuis & Strick (2016) have argued that higher cognitive processes can proceed outside of awareness to elaborate and answer unfinished thoughts. Importantly, multiple empirical findings in support of this view have failed to replicate (Nieuwenstein et al., 2015). Still, in light of the anecdotal and empirical evidence for the creative and practical benefits of “incubating” an idea (Gable et al., 2019), we need to build concrete models of how thoughts and goals can persist in mind, even when we do not overtly intend or attend to those thoughts and goals.

What is the purpose of psychological momentum?

If we want to understand how mindsets, memories and goal states interact to generate psychological momentum, it can be helpful to first consider the problem normatively: what are the adaptive costs and benefits of lingering? We are familiar with settings in which psychological momentum seems to be maladaptive, as when a conversation has moved on to a new topic, but our thinking is pulled back to an earlier topic. In such settings, past information interferes with present processing. Indeed, a key principle of event segmentation theory is that only the current “event model” is in working memory (Radvansky & Zacks, 2017), which minimizes interference between past information and current processing . So in what ways can psychological momentum be adaptive?

If the past properties of the world are likely to persist into the future, then it could be adaptive for us to allow previously important information to persist in mind. Anderson & Schooler (1991) argued that items in memory should be kept “available” proportional to the likelihood that they will be needed again, and that the likelihood of needing them can be predicted by the rate at which they were encountered in the past. We can extend this rational argument to apply not only to the persistence of mental content, but also to broad dispositions and emotions: if the world was dangerous in the past few minutes, then it is likely to be dangerous again in the next few minutes. In this view, then, it becomes adaptive for our mental states to persist in a world where the statistics are slowly changing, so that frequently expressed past information and behavioral needs are likely to recur. Conversely, if we are operating in a world of sharply shifting and unrelated contexts (e.g., speed dating, or an unbroken series of unrelated work meetings) then it may be maladaptive for information to linger in mind (see also Dubrow et al., 2017).

The idea that lingering mental content is adaptive has not, to our knowledge, been empirically tested in humans. But to determine which information should (ideally) persist in mind, and under what conditions, it may be possible to make progress using simulated agents. For example, Lu et al. (2022) modeled memory encoding and retrieval as “actions” available to an agent learning an optimal policy for making successful predictions about its environment. In this way, their model connects properties of the environment (e.g., how and when the statistics of the environment shift) with optimal policies on memory maintenance and storage. When the model was exposed to sequences of unrelated events, it learned to selectively retrieve past information at moments of higher uncertainty about the immediate future. Future work can connect such a reinforcement learning model to the brain, by mapping the regions of the DMN and their situation-representation circuitry) as subcomponents of a high-level control agent endowed with an episodic memory system (Gershman & Daw, 2017; Dohmatob et al., 2020). Critically, when memory is modeled as a resource available to an agent, then we can understand how the agent uses memory not only to predict states of the world, but also to act in the world (Goyal et al., 2022). In this way, we may eventually understand how persisting and replayed mental content not only supports memory consolidation (in the service of learning), but can also provide immediate advantages for problem solving and decision making.

In addition to developing the agentive model sketched above, we will briefly mention some important future directions for understanding psychological momentum. First, we must develop interventions that can block or interfere with lingering mental states, not only to properly characterize causal relationships, but also to help people manage unwanted lingering in their lives. Second, given the variability in which individuals experience psychological momentum (e.g., Figure 2C), probing individual differences is crucial (Andrews-Hanna et al. 2021; Yeung & Fernandes, 2021). Third, we must understand whether human language is important for psychological momentum: do the semantics and syntax of language provide a scaffold that allows latent mental states to persist and elaborate (Clark & Toribio, 2012). Finally, we must determine whether psychological momentum depends on a “self”-related task representation (e.g., Klinger, 1978), possibly distinct from other kinds of task-sets, and implemented in DMN circuitry.

Conclusion

The phenomenon of psychological momentum (Figure 1) can be understood through the lens of memory; in relation to consolidation and generalization; as a kind of persistent task-set or mindset; or as a form of rumination around current concerns. Each of these lenses reveals a part of this basic feature of human experience, but future work should strive toward a more integrated view. To this end, we must theoretically and empirically determine how learning agents manage the demands of competing memories and tasks, as they try to solve problems and achieve their goals in a continually changing world.

Acknowledgments

The authors gratefully acknowledge the support of the National Institutes of Mental Health (grant R01MH119099 to C.J.H). The authors would also like to thank Simon Brown, Aidan Horner, Morris Moscovitch, Robert Goldstone and our reviewers for their thoughtful suggestions on earlier versions of this manuscript.

Footnotes

Conflict of interest: The authors declare no competing financial interests.

Contributor Information

Christopher J. Honey, Department of Psychological & Brain Sciences, Johns Hopkins University

Abhijt Mahabal, Content and Discovery, Pinterest.

Buddhika Bellana, Department of Psychology, Glendon Campus, York University.

References

  1. Amir I, & Bernstein A (2021). Dynamics of internal attention and internally-directed cognition: The attention-to-thoughts (a2t) model.
  2. Anderson JR, & Schooler LJ (1991). Reflections of the environment in memory. Psychological science, 2(6), 396–408. [Google Scholar]
  3. Andrews-Hanna JR, Woo CW, Wilcox R, Eisenbarth H, Kim B, Han J, … & Wager TD. (2021). The conceptual building blocks of everyday thought: Tracking the emergence and dynamics of ruminative and nonruminative thinking. Journal of Experimental Psychology: General. [DOI] [PMC free article] [PubMed] [Google Scholar]; ** Using a free association paradigm, Andrews-Hanna et al. demonstrated that our current concerns from our daily lives shape the trajectory of our spontaneous generated semantic associations. Critically, they demonstrate that a propensity to move towards and remain in negatively-valenced points in free association ‘space’ can predict individual differences in rumination.
  4. Bargh JA, 2011. Unconscious thought theory and its discontents: A critique of the critiques. Social Cognition, 29(6), pp.629–647. [Google Scholar]
  5. Bellana B, Mahabal A, & Honey CJ (2022). Narrative thinking lingers in spontaneous thought. Nature Communications, 13(1), 4585. [DOI] [PMC free article] [PubMed] [Google Scholar]; ** Using both objective measures (i.e., quantifying differences between pre-and post-story free association chains) and subjective measures (i.e., self-report), we show that stories persist in our spontaneous thoughts for several minutes after reading. We conclude that the extent to which an input lingers may be a function of the depth of ‘meaning’ extracted from that input.
  6. Bornstein AM and Norman KA, 2017. Reinstated episodic context guides sampling-based decisions for reward. Nature neuroscience, 20(7), pp.997–1003. [DOI] [PubMed] [Google Scholar]
  7. Buzsáki G (2015). Hippocampal sharp wave-ripple: A cognitive biomarker for episodic memory and planning. Hippocampus, 25(10), 1073–1188. [DOI] [PMC free article] [PubMed] [Google Scholar]
  8. Christoff K, Irving ZC, Fox KC, Spreng RN, & Andrews-Hanna JR (2016). Mindwandering as spontaneous thought: a dynamic framework. Nature Reviews Neuroscience, 17(11), 718–731. [DOI] [PubMed] [Google Scholar]
  9. Clark A, & Toribio J (2012). Magic words: how language augments hum an computation.In Language and Meaning in Cognitive Science (pp. 33–51). Routledge. [Google Scholar]
  10. Cowan N (2017). The many faces of working memory and short-term storage. Psychonomic bulletin & review, 24(4), 1158–1170. [DOI] [PubMed] [Google Scholar]
  11. Craik FI, & Lockhart RS (1972). Levels of processing: A framework for memory research. Journal of verbal learning and verbal behavior, 11(6), 671–684. [Google Scholar]
  12. Dewar MT, Cowan N, & Della Sala S (2007). Forgetting due to retroactive interference: A fusion of MüHer and Pilzecker’s (1900) early insights into everyday forgetting and recent research on anterograde amnesia. Cortex, 43(5), 616–634. [DOI] [PMC free article] [PubMed] [Google Scholar]
  13. Dijksterhuis A and Strick M, 2016. A case for thinking without consciousness. Perspectives on Psychological Science, 11(1), pp.117–132. [DOI] [PubMed] [Google Scholar]
  14. Dohmatob E, Dumas G, & Bzdok D (2020). Dark control: The default mode network as a reinforcement learning agent. Human Brain Mapping, 41(12) [DOI] [PMC free article] [PubMed] [Google Scholar]
  15. DuBrow S, Rouhani N, Niv Y, Norman KA. Does mental context drift or shift?. Current opinion in behavioral sciences. 2017. Oct 1;17:141–6. [DOI] [PMC free article] [PubMed] [Google Scholar]
  16. Faber M, & D’Mello SK (2018). How the stimulus influences mind wandering in semantically rich task contexts. Cognitive research: principles and implications, 3(1), 1–14. [DOI] [PMC free article] [PubMed] [Google Scholar]; ** Faber & D’Mello highlight the importance of the complexity and richness of a stimulus when examining patterns of spontaneous thought. Whereas mind wandering in constrained, lab-based tasks tends to be characterized by prospection and introspection, semantically-rich tasks elicit more mind wandering events that are autobiographical or semantic in nature, and that are related to the content of the eliciting stimulus.
  17. Gable SL, Hopper EA, & Schooler JW (2019). When the muses strike: Creative ideas of physicists and writers routinely occur during mind wandering. Psychological science, 30(3), 396–404. [DOI] [PubMed] [Google Scholar]
  18. Gershman SJ, Daw ND. Reinforcement learning and episodic memory in humans and animals: an integrative framework. Annual review of psychology. 2017. Jan 3;68:101–28. [DOI] [PMC free article] [PubMed] [Google Scholar]
  19. Goyal A, Friesen A, Banino A, Weber T, Ke NR, Badia AP, Guez A, Mirza M, Humphreys PC, Konyushova K, Valko M, Osindero S, Lillicrap T, Heess N, Blundell C. Retrieval-augmented reinforcement learning. International Conference on Machine Learning pp. 7740–7765. PMLR. [Google Scholar]
  20. Goyal A, Friesen A, Banino A, Weber T, Ke NR, Badia AP, … & Blundell C (2022, June). Retrieval-augmented reinforcement learning. In International Conference on Machine Learning (pp. 7740–7765). PMLR. [Google Scholar]
  21. Hasson U, Chen J, Honey CJ. Hierarchical process memory: memory as an integral component of information processing. Trends in cognitive sciences. 2015. Jun 1;19(6):304–13. [DOI] [PMC free article] [PubMed] [Google Scholar]
  22. Herz N, Baror S, & Bar M (2020). Overarching states of mind. Trends in Cognitive Sciences, 24(3), 184–199. [DOI] [PubMed] [Google Scholar]; ** Herz et al. argue for the central cognitive role of ‘states of mind’ or overarching dispositions that fluctuate over time and context. Understanding of how states of mind arise and shape ongoing cognition will be a necessary step towards understanding why some experiences have a tendency to exert a persistent influence.
  23. Higgins C, Liu Y, Vidaurre D, Kurth-Nelson Z, Dolan R, Behrens T, & Woolrich M (2021). Replay bursts in humans coincide with activation of the default mode and parietal alpha networks. Neuron, 109(5), 882–893. [DOI] [PMC free article] [PubMed] [Google Scholar]
  24. Howard MW, & Kahana MJ (2002). A distributed representation of temporal context. Journal of mathematical psychology, 46(3), 269–299. [Google Scholar]
  25. Humiston GB, Tucker MA, Summer T, & Wamsley EJ (2019). Resting states and memory consolidation: A preregistered replication and meta-analysis. Scientific reports, 9(1), 1–9. [DOI] [PMC free article] [PubMed] [Google Scholar]
  26. Klinger E, 1978. Modes of normal conscious flow. In The stream of consciousness (pp. 225–258). Springer, Boston, MA. [Google Scholar]
  27. Klinger E, and Cox WM (2011). “Motivation and the goal theory of current concerns,” in Handbook of Motivational Counseling, 2nd Edn, eds Cox WM and Klinger E (Chichester: Wiley; ), 3–47. [Google Scholar]
  28. Lu Q, Hasson U, & Norman KA (2022). A neural network model of when to retrieve and encode episodic memories. eLife, 11, e74445. [DOI] [PMC free article] [PubMed] [Google Scholar]; ** Lu et al. computationally study when it would be optimal to encode and retrieve memories. In this modelling framework, memory is a tool available to an agent performing a task. This framing enables scientists to quantitatively determine the conditions under which task performance is aided, or hindered, by lingering mental states.
  29. Madore KP, Addis DR, & Schacter DL (2015). Creativity and memory: Effects of an episodic-specificity induction on divergent thinking. Psychological science, 26(9), 1461–1468. [DOI] [PMC free article] [PubMed] [Google Scholar]
  30. Mattar MG, & Daw ND (2018). Prioritized memory access explains planning and hippocampal replay. Nature neuroscience, 21(11), 1609–1617. [DOI] [PMC free article] [PubMed] [Google Scholar]
  31. Müller GE and Pilzecker A. (1900). Experimentelle Beiträge zur Lehre vom Gedächtnis. Z. Psychol. Ergänzungsband 1: 1–300. [Google Scholar]
  32. Nieuwenstein MR, Wierenga T, Morey RD, Wicherts JM, Blom TN, Wagenmakers EJ and van Rijn H, 2015. On making the right choice: A meta-analysis and large-scale replication attempt of the unconscious thought advantage. Judgment and Decision Making, 10(1), pp.1–17. [Google Scholar]
  33. Oberauer K, Farrell S, Jarrold C, & Lewandowsky S (2016). What limits working memory capacity?. Psychological bulletin, 142(7), 758. [DOI] [PubMed] [Google Scholar]
  34. Radvansky GA, & Zacks JM (2017). Event boundaries in memory and cognition. Current opinion in behavioral sciences, 17, 133–140. [DOI] [PMC free article] [PubMed] [Google Scholar]
  35. Schuck NW, & Niv Y (2019). Sequential replay of nonspatial task states in the human hippocampus. Science, 364(6447), eaaw5181. [DOI] [PMC free article] [PubMed] [Google Scholar]
  36. Spinhoven P, van Hemert AM, & Penninx BW (2018). Repetitive negative thinking as a predictor of depression and anxiety: A longitudinal cohort study. Journal of Affective Disorders, 241, 216–225. [DOI] [PubMed] [Google Scholar]
  37. Tambini A, Rimmele U, Phelps EA, & Davachi L (2017). Emotional brain states carry over and enhance future memory formation. Nature neuroscience, 20(2), 271–278. [DOI] [PMC free article] [PubMed] [Google Scholar]
  38. Tambini A, & Davachi L (2019). Awake reactivation of prior experiences consolidates memories and biases cognition. Trends in cognitive sciences, 23(10), 876–890. [DOI] [PMC free article] [PubMed] [Google Scholar]
  39. [v1ncetta]. (2017, October 25). What book gave you the biggest book hangover [Online forum post]. Reddit.https://www.reddit.com/r/books/comments/78mmsg/what_book_gave_you_the_biggest_book_hangover/ [Google Scholar]
  40. Wittkuhn L, Chien S, Hall-McMaster S, Schuck NW. Replay in minds and machines. Neuroscience & Biobehavioral Reviews 2021. Oct 1;129:367–88. [DOI] [PubMed] [Google Scholar]; ** In this review, Wittkuhn et al. highlight the parallels between replay and learning in neuroscience and in artificial intelligence (AI). Of particular importance, the use of replay in AI provides compelling examples of how replaying past experiences can benefit new learning.
  41. Wylie G, & Allport A (2000). Task switching and the measurement of “switch costs”. Psychological research, 63(3), 212–233. [DOI] [PubMed] [Google Scholar]
  42. Yeung RC, & Fernandes MA (2021). Recurrent involuntary memories are modulated by age and linked to mental health. Psychology and aging, 36(7), 883. [DOI] [PubMed] [Google Scholar]
  43. Zeigarnik B (1927). Das Behalten erledigter und unerledigter Handlungen [Retention of completed and uncompleted actions.]. Psychologische Forschung, 9, 1–85. [Google Scholar]
  44. Zuo X, Honey CJ, Barense MD, Crombie D, Norman KA, Hasson U, & Chen J (2020). Temporal integration of narrative information in a hippocampal amnesic patient. Neuroimage, 213, 116658. [DOI] [PMC free article] [PubMed] [Google Scholar]

RESOURCES