Reward prediction errors create event boundaries in memory

Locus coeruleus activation 'resets' hippocampal event representations and separates adjacent memories

Memories reflect the ebb and flow of experiences, capturing unique and meaningful events from our lives. Using a combination of functional magnetic resonance imaging (fMRI), neuromelanin imaging, and pupillometry, we show that arousal and locus coeruleus (LC) activation transform otherwise continuous experiences into distinct episodic memories. As sequences unfold, encountering a context shift, or event boundary, triggers arousal and LC processes that predict later memory separation. Boundaries furthermore promote temporal pattern separation within left hippocampal dentate gyrus, which correlates with heightened LC responses to those same transition points. We also find that a neurochemical index of prolonged LC activation correlates with diminished arousal responses at boundaries, suggesting a connection between elevated LC output and impaired event processing. These findings align with the idea that arousal processes initiate a neural and memory 'reset' in response to significant changes, constructing the very episodes that define everyday memory.

The neuroscience of active learning and direct instruction

Throughout the educational system, students experiencing active learning pedagogy perform better and fail less than those taught through direct instruction. Can this be ascribed to differences in learning from a neuroscientific perspective? This review examines mechanistic, neuroscientific evidence that might explain differences in cognitive engagement contributing to learning outcomes between these instructional approaches. In classrooms, direct instruction comprehensively describes academic content, while active learning provides structured opportunities for learners to explore, apply, and manipulate content. Synaptic plasticity and its modulation by arousal or novelty are central to all learning and both approaches. As a form of social learning, direct instruction relies upon working memory. The reinforcement learning circuit, associated agency, curiosity, and peer-to-peer social interactions combine to enhance motivation, improve retention, and build higher-order-thinking skills in active learning environments. When working memory becomes overwhelmed, additionally engaging the reinforcement learning circuit improves retention, providing an explanation for the benefits of active learning. This analysis provides a mechanistic examination of how emerging neuroscience principles might inform pedagogical choices at all educational levels.

The influence of emotion on temporal context models

Temporal context models (TCMs) have been influential in understanding episodic memory and its neural underpinnings. Recently, TCMs have been extended to explain emotional memory effects, one of the most clinically important findings in the field of memory research. This review covers recent advances in hypotheses for the neural representation of spatiotemporal context through the lens of TCMs, including their ability to explain the influence of emotion on episodic and temporal memory. In recent years, simplifying assumptions of "classical" TCMs - with exponential trace decay and the mechanism by which temporal context is recovered - have become increasingly clear. The review also outlines how recent advances could be incorporated into a future TCM, beyond classical assumptions, to integrate emotional modulation.

Learning to express reward prediction error-like dopaminergic activity requires plastic representations of time

The dominant theoretical framework to account for reinforcement learning in the brain is temporal difference learning (TD) learning, whereby certain units signal reward prediction errors (RPE). The TD algorithm has been traditionally mapped onto the dopaminergic system, as firing properties of dopamine neurons can resemble RPEs. However, certain predictions of TD learning are inconsistent with experimental results, and previous implementations of the algorithm have made unscalable assumptions regarding stimulus-specific fixed temporal bases. We propose an alternate framework to describe dopamine signaling in the brain, FLEX (Flexibly Learned Errors in Expected Reward). In FLEX, dopamine release is similar, but not identical to RPE, leading to predictions that contrast to those of TD. While FLEX itself is a general theoretical framework, we describe a specific, biophysically plausible implementation, the results of which are consistent with a preponderance of both existing and reanalyzed experimental data.

Bridging Neuroscience and AI: Environmental Enrichment as a Model for Forward Knowledge Transfer

Continual learning (CL) refers to an agent's capability to learn from a continuous stream of data and transfer knowledge without forgetting old information. One crucial aspect of CL is forward transfer, i.e., improved and faster learning on a new task by leveraging information from prior knowledge. While this ability comes naturally to biological brains, it poses a significant challenge for artificial intelligence (AI). Here, we suggest that environmental enrichment (EE) can be used as a biological model for studying forward transfer, inspiring human-like AI development. EE refers to animal studies that enhance cognitive, social, motor, and sensory stimulation and is a model for what, in humans, is referred to as 'cognitive reserve'. Enriched animals show significant improvement in learning speed and performance on new tasks, typically exhibiting forward transfer. We explore anatomical, molecular, and neuronal changes post-EE and discuss how artificial neural networks (ANNs) can be used to predict neural computation changes after enriched experiences. Finally, we provide a synergistic way of combining neuroscience and AI research that paves the path toward developing AI capable of rapid and efficient new task learning.

Prediction error and event segmentation in episodic memory

Organizing the continuous flow of experiences into meaningful events is a crucial prerequisite for episodic memory. Prediction error and event segmentation both play important roles in supporting the genesis of meaningful mnemonic representations of events. We review theoretical contributions discussing the relationship between prediction error and event segmentation, as well as literature on episodic memory related to prediction error and event segmentation. We discuss the extent of overlap of mechanisms underlying memory emergence through prediction error and event segmentation, with a specific focus on attention and working memory. Finally, we identify areas in research that are currently developing and suggest future directions. We provide an overview of mechanisms underlying memory formation through predictions, violations of predictions, and event segmentation.

Multiple timescales of temporal context in risky choice: Behavioral identification and relationships to physiological arousal

Context-dependence is fundamental to risky monetary decision-making. A growing body of evidence suggests that temporal context, or recent events, alters risk-taking at a minimum of three timescales: immediate (e.g. trial-by-trial), neighborhood (e.g. a group of consecutive trials), and global (e.g. task-level). To examine context effects, we created a novel monetary choice set with intentional temporal structure in which option values shifted between multiple levels of value magnitude ("contexts") several times over the course of the task. This structure allowed us to examine whether effects of each timescale were simultaneously present in risky choice behavior and the potential mechanistic role of arousal, an established correlate of risk-taking, in context-dependency. We found that risk-taking was sensitive to immediate, neighborhood, and global timescales: risk-taking decreased following large (vs. small) outcome amounts, increased following large positive (but not negative) shifts in context, and increased when cumulative earnings exceeded expectations. We quantified arousal with skin conductance responses, which were related to the global timescale, increasing with cumulative earnings, suggesting that physiological arousal captures a task-level assessment of performance. Our results both replicate and extend prior research by demonstrating that risky decision-making is consistently dynamic at multiple timescales and that the role of arousal in risk-taking extends to some, but not all timescales of context-dependence.

Interval Timing as a Computational Pathway From Early Life Adversity to Affective Disorders

Adverse early life experiences can have remarkably enduring negative consequences on mental health, with numerous, varied psychiatric conditions sharing this developmental origin. Yet, the mechanisms linking adverse experiences to these conditions remain poorly understood. Here, we draw on a principled model of interval timing to propose that statistically optimal adaptation of temporal representations to an unpredictable early life environment can produce key characteristics of anhedonia, a transdiagnostic symptom associated with affective disorders like depression and anxiety. The core observation is that early temporal unpredictability produces broader, more imprecise temporal expectations. As a result, reward anticipation is diminished, and associative learning is slowed. When agents with such representations are later introduced to more stable environments, they demonstrate a negativity bias, responding more to the omission of reward than its receipt. Increased encoding of negative events has been proposed to contribute to disorders with anhedonia as a symptom. We then examined how unpredictability interacts with another form of adversity, low reward availability. We found that unpredictability's effect was most strongly felt in richer environments, potentially leading to categorically different phenotypic expressions. In sum, our formalization suggests a single mechanism can help to link early life adversity to a range of behaviors associated with anhedonia, and offers novel insights into the interactive impacts of multiple adversities.

More than a moment: What does it mean to call something an 'event'?

Experiences are stored in the mind as discrete mental units, or 'events,' which influence-and are influenced by-attention, learning, and memory. In this way, the notion of an 'event' is foundational to cognitive science. However, despite tremendous progress in understanding the behavioral and neural signatures of events, there is no agreed-upon definition of an event. Here, we discuss different theoretical frameworks of event perception and memory, noting what they can and cannot account for in the literature. We then highlight key aspects of events that we believe should be accounted for in theories of event processing--in particular, we argue that the structure and substance of events should be better reflected in our theories and paradigms. Finally, we discuss empirical gaps in the event cognition literature and what the future of event cognition research may look like.

Long-term, multi-event surprise correlates with enhanced autobiographical memory

Neurobiological and psychological models of learning emphasize the importance of prediction errors (surprises) for memory formation. This relationship has been shown for individual momentary surprising events; however, it is less clear whether surprise that unfolds across multiple events and timescales is also linked with better memory of those events. We asked basketball fans about their most positive and negative autobiographical memories of individual plays, games and seasons, allowing surprise measurements spanning seconds, hours and months. We used advanced analytics on National Basketball Association play-by-play data and betting odds spanning 17 seasons, more than 22,000 games and more than 5.6 million plays to compute and align the estimated surprise value of each memory. We found that surprising events were associated with better recall of positive memories on the scale of seconds and months and negative memories across all three timescales. Game and season memories could not be explained by surprise at shorter timescales, suggesting that long-term, multi-event surprise correlates with memory. These results expand notions of surprise in models of learning and reinforce its relevance in real-world domains.

Bayesian Surprise Predicts Human Event Segmentation in Story Listening

Event segmentation theory posits that people segment continuous experience into discrete events and that event boundaries occur when there are large transient increases in prediction error. Here, we set out to test this theory in the context of story listening, by using a deep learning language model (GPT-2) to compute the predicted probability distribution of the next word, at each point in the story. For three stories, we used the probability distributions generated by GPT-2 to compute the time series of prediction error. We also asked participants to listen to these stories while marking event boundaries. We used regression models to relate the GPT-2 measures to the human segmentation data. We found that event boundaries are associated with transient increases in Bayesian surprise but not with a simpler measure of prediction error (surprisal) that tracks, for each word in the story, how strongly that word was predicted at the previous time point. These results support the hypothesis that prediction error serves as a control mechanism governing event segmentation and point to important differences between operational definitions of prediction error.

Predictions transform memories: How expected versus unexpected events are integrated or separated in memory

Our brains constantly generate predictions about the environment based on prior knowledge. Many of the events we experience are consistent with these predictions, while others might be inconsistent with prior knowledge and thus violate our predictions. To guide future behavior, the memory system must be able to strengthen, transform, or add to existing knowledge based on the accuracy of our predictions. We synthesize recent evidence suggesting that when an event is consistent with our predictions, it leads to neural integration between related memories, which is associated with enhanced associative memory, as well as memory biases. Prediction errors, in turn, can promote both neural integration and separation, and lead to multiple mnemonic outcomes. We review these findings and how they interact with factors such as memory reactivation, prediction error strength, and task goals, to offer insight into what determines memory for events that violate our predictions. In doing so, this review brings together recent neural and behavioral research to advance our understanding of how predictions shape memory, and why.

Learning to Express Reward Prediction Error-like Dopaminergic Activity Requires Plastic Representations of Time

The dominant theoretical framework to account for reinforcement learning in the brain is temporal difference (TD) reinforcement learning. The TD framework predicts that some neuronal elements should represent the reward prediction error (RPE), which means they signal the difference between the expected future rewards and the actual rewards. The prominence of the TD theory arises from the observation that firing properties of dopaminergic neurons in the ventral tegmental area appear similar to those of RPE model-neurons in TD learning. Previous implementations of TD learning assume a fixed temporal basis for each stimulus that might eventually predict a reward. Here we show that such a fixed temporal basis is implausible and that certain predictions of TD learning are inconsistent with experiments. We propose instead an alternative theoretical framework, coined FLEX (Flexibly Learned Errors in Expected Reward). In FLEX, feature specific representations of time are learned, allowing for neural representations of stimuli to adjust their timing and relation to rewards in an online manner. In FLEX dopamine acts as an instructive signal which helps build temporal models of the environment. FLEX is a general theoretical framework that has many possible biophysical implementations. In order to show that FLEX is a feasible approach, we present a specific biophysically plausible model which implements the principles of FLEX. We show that this implementation can account for various reinforcement learning paradigms, and that its results and predictions are consistent with a preponderance of both existing and reanalyzed experimental data.

Multiple routes to enhanced memory for emotionally relevant events

Events associated with aversive or rewarding outcomes are prioritized in memory. This memory boost is commonly attributed to the elicited affective response, closely linked to noradrenergic and dopaminergic modulation of hippocampal plasticity. Herein we review and compare this 'affect' mechanism to an additional, recently discovered, 'prediction' mechanism whereby memories are strengthened by the extent to which outcomes deviate from expectations, that is, by prediction errors (PEs). The mnemonic impact of PEs is separate from the affective outcome itself and has a distinct neural signature. While both routes enhance memory, these mechanisms are linked to different - and sometimes opposing - predictions for memory integration. We discuss new findings that highlight mechanisms by which emotional events strengthen, integrate, and segment memory.

Strong but Fragmented Memory of a Stressful Episode

While it is commonly assumed that stressful events are vividly remembered, it remains largely unknown whether all aspects of memory for a stressful episode are enhanced. In this preregistered study, we tested whether stress enhances later remembering of individual elements of a stressful episode at the cost of impaired processing of the association between these elements. Therefore, male and female participants (N = 122) underwent a stressful (or control) episode during which they encoded a series of stimuli. To investigate stress effects on the memory for individual events and the links between these, we used temporal sequence effects in recognition memory tested 24 h after encoding. Specifically, we tested whether stress would affect the memory enhancement for a target item if this is preceded by another item that also preceded the target during encoding (recognition priming). Our results showed that participants recalled single events encoded under stress better than those encoded under nonstressful conditions, but were less able to leverage the temporal sequence of events encoded under stress to cue memory at delayed recall, reflected in reduced memory for items preceded by the item that preceded them also during encoding. Functional near-infrared spectroscopy further revealed that encoding under stress was accompanied by opposite changes in inferotemporal and dorsolateral prefrontal areas. Together, our data suggest that acute stress induces a mode of memory formation that results in strong but less integrated memories.

Collective events and individual affect shape autobiographical memory

How do collective events shape how we remember our lives? We leveraged advances in natural language processing as well as a rich, longitudinal assessment of 1,000 Americans throughout 2020 to examine how memory is influenced by two prominent factors: surprise and emotion. Autobiographical memory for 2020 displayed a unique signature: There was a substantial bump in March, aligning with pandemic onset and lockdowns, consistent across three memory collections 1 y apart. We further investigated how emotion, using both immediate and retrieved measures, predicted the amount and content of autobiographical memory: Negative affect increased recall across all measures, whereas its more clinical indices, depression and posttraumatic stress disorder, selectively increased nonepisodic recall. Finally, in a separate cohort, we found pandemic news to be better remembered, surprising, and negative, while lockdowns compressed remembered time. Our work connects laboratory findings to the real world and delineates the effects of acute versus clinical signatures of negative emotion on memory.

Inferring danger with minimal aversive experience

Learning about threats is crucial for survival and fundamentally rests upon Pavlovian conditioning. However, Pavlovian threat learning is largely limited to detecting known (or similar) threats and involves first-hand exposure to danger, which inevitably poses a risk of harm. We discuss how individuals leverage a rich repertoire of mnemonic processes that operate largely in safety and significantly expand our ability to recognize danger beyond Pavlovian threat associations. These processes result in complementary memories - acquired individually or through social interactions - that represent potential threats and the relational structure of our environment. The interplay between these memories allows danger to be inferred rather than directly learned, thereby flexibly protecting us from potential harm in novel situations despite minimal prior aversive experience.

Evidence That Event Boundaries Are Access Points for Memory Retrieval

When recalling memories, we often scan information-rich continuous episodes, for example, to find our keys. How does our brain access and search through those memories? We suggest that high-level structure, marked by event boundaries, guides us through this process: In our computational model, memory scanning is sped up by skipping ahead to the next event boundary upon reaching a decision threshold. In adult Mechanical Turk workers from the United States, we used a movie (normed for event boundaries; Study 1, N = 203) to prompt memory scanning of movie segments for answers (Study 2, N = 298) and mental simulation (Study 3, N = 100) of these segments. Confirming model predictions, we found that memory-scanning times varied as a function of the number of event boundaries within a segment and the distance of the search target to the previous boundary (the key diagnostic parameter). Mental simulation times were also described by a skipping process with a higher skipping threshold than memory scanning. These findings identify event boundaries as access points to memory.

Mnemonic Content and Hippocampal Patterns Shape Judgments of Time

Our experience of time can feel dilated or compressed, rather than reflecting true "clock time." Although many contextual factors influence the subjective perception of time, it is unclear how memory accessibility plays a role in constructing our experience of and memory for time. Here, we used a combination of behavioral and functional MRI measures in healthy young adults (N = 147) to ask the question of how memory is incorporated into temporal duration judgments. Behaviorally, we found that event boundaries, which have been shown to disrupt ongoing memory integration processes, result in the temporal compression of duration judgments. Additionally, using a multivoxel pattern similarity analysis of functional MRI data, we found that greater temporal pattern change in the left hippocampus within individual trials was associated with longer duration judgments. Together, these data suggest that mnemonic processes play a role in constructing representations of time.

Aberrant memory and delusional ideation: A pernicious partnership?

Delusions can be conceptualized as beliefs that are both at odds with consensus reality and espoused with high conviction. While delusions represent a cardinal symptom of schizophrenia, delusion-like beliefs can be found in the general population. Do similar cognitive mechanisms support delusionality across this spectrum? If so, what are they? Here, we examine evidence for a mechanistic role of the (associative) memory system in the formation and maintenance of delusions and delusion-like beliefs. While general neurocognitive metrics do not tend to associate with delusionality, our scoping review of the clinical and subclinical literature reveals several subdomains of memory function that do. These include a propensity to commit errors of commission (i.e., false alarms and intrusions), source memory biases, and metamemory impairment. We discuss how several of these effects may stem from aberrant associative memory function and offer recommendations for future research. Further, we propose a state/trait interaction model in which underlying traits (i.e., impaired associative and metamemory function) may become coupled with delusionality during states of acute psychosis, when memory function is particularly challenged by aberrant salience attribution and noisy perceptual input. According to this model, delusions may arise as explanations to high-salience (but low-source) mnemonic content that is endorsed with high confidence.

Contextual inference in learning and memory

Context is widely regarded as a major determinant of learning and memory across numerous domains, including classical and instrumental conditioning, episodic memory, economic decision-making, and motor learning. However, studies across these domains remain disconnected due to the lack of a unifying framework formalizing the concept of context and its role in learning. Here, we develop a unified vernacular allowing direct comparisons between different domains of contextual learning. This leads to a Bayesian model positing that context is unobserved and needs to be inferred. Contextual inference then controls the creation, expression, and updating of memories. This theoretical approach reveals two distinct components that underlie adaptation, proper and apparent learning, respectively referring to the creation and updating of memories versus time-varying adjustments in their expression. We review a number of extensions of the basic Bayesian model that allow it to account for increasingly complex forms of contextual learning.

The subjective feeling of a gap between conceptual and perceptual fluency is interpreted as a metacognitive signal of pastness

The present study aimed to address the following question: does the discrepancy between an expected word and its readability enhances or impair its memorability? We used an adaptation of the sentence stem paradigm (Whittlesea in J Exp Psycol 19:1235-1253, 1993) and manipulated the perceptual clarity of the words by introducing some Gaussian noise (Reber in Psycol Sci 9:45-48, 1998). The target words were semantically predictable or otherwise (conceptual fluency) or were easy or difficult to read (perceptual fluency). The first experiment was conducted to ensure that the two manipulated factors had an impact on the readability of the words. In particular, results showed that when the words were written against a noisy background their predictability enhanced the judgement of readability. The second experiment aimed to test the hypothesis that recognition would be influenced by the discrepancy between conceptual and perceptual fluency. The results showed that with a noisy background, the predictability of the target words had an impact on recognition judgement; with a clear background, the effect on the recognition judgement was caused by the non-predictability of the target words. Conversely, confidence in judgement increased when the two factors went in the same direction, that is, predictability with clarity and non-predictability with low clarity. The results showed that (a) depending on the task, the effects of conceptual and perceptual fluency did not go in the same direction; (b) the kinds of fluency (conceptual and perceptual) were not independent; and (c) recognition judgements were affected by the gap between conceptual and perceptual fluency.

Reward Uncertainty and Expected Value Enhance Generalization of Episodic Memory

Previous research has revealed some mechanisms underlying the generalization of reward expectation of generalization stimuli, but little is known about the generalization of episodic memory for rewarding events, its consolidation, and how reward components such as expected value and reward uncertainty affect it. Participants underwent a Pavlovian reward-conditioning task to test whether reward conditioning would enhance episodic memory generalization and which reward components would directly affect it. Counterbalanced across participants, one semantic category was paired with a reward, while the other was never paired. Following a delay of either 5 min or 24 h, participants took a memory test consisting of old, highly similar, and new items. We found that participants were more likely to falsely recognize lure items as old in the reward-paired category after 5 min and 24 h delays. These results indicate that reward conditioning enhanced the generalization of episodic memory, but this effect was not necessarily dependent on consolidation. The composite score and raw data of generalization further showed that the uncertainty and expected value enhanced generalization. Together, these findings revealed an effect of reward conditioning on episodic memory generalization and supported the enhancement effects of expected value and uncertainty.

Generalized Simultaneous Localization and Mapping (G-SLAM) as unification framework for natural and artificial intelligences: towards reverse engineering the hippocampal/entorhinal system and principles of high-level cognition

Simultaneous localization and mapping (SLAM) represents a fundamental problem for autonomous embodied systems, for which the hippocampal/entorhinal system (H/E-S) has been optimized over the course of evolution. We have developed a biologically-inspired SLAM architecture based on latent variable generative modeling within the Free Energy Principle and Active Inference (FEP-AI) framework, which affords flexible navigation and planning in mobile robots. We have primarily focused on attempting to reverse engineer H/E-S "design" properties, but here we consider ways in which SLAM principles from robotics may help us better understand nervous systems and emergent minds. After reviewing LatentSLAM and notable features of this control architecture, we consider how the H/E-S may realize these functional properties not only for physical navigation, but also with respect to high-level cognition understood as generalized simultaneous localization and mapping (G-SLAM). We focus on loop-closure, graph-relaxation, and node duplication as particularly impactful architectural features, suggesting these computational phenomena may contribute to understanding cognitive insight (as proto-causal-inference), accommodation (as integration into existing schemas), and assimilation (as category formation). All these operations can similarly be describable in terms of structure/category learning on multiple levels of abstraction. However, here we adopt an ecological rationality perspective, framing H/E-S functions as orchestrating SLAM processes within both concrete and abstract hypothesis spaces. In this navigation/search process, adaptive cognitive equilibration between assimilation and accommodation involves balancing tradeoffs between exploration and exploitation; this dynamic equilibrium may be near optimally realized in FEP-AI, wherein control systems governed by expected free energy objective functions naturally balance model simplicity and accuracy. With respect to structure learning, such a balance would involve constructing models and categories that are neither too inclusive nor exclusive. We propose these (generalized) SLAM phenomena may represent some of the most impactful sources of variation in cognition both within and between individuals, suggesting that modulators of H/E-S functioning may potentially illuminate their adaptive significances as fundamental cybernetic control parameters. Finally, we discuss how understanding H/E-S contributions to G-SLAM may provide a unifying framework for high-level cognition and its potential realization in artificial intelligences.

Grounding the Attentional Boost Effect in Events and the Efficient Brain

Attention and memory for everyday experiences vary over time, wherein some moments are better attended and subsequently better remembered than others. These effects have been demonstrated in naturalistic viewing tasks with complex and relatively uncontrolled stimuli, as well as in more controlled laboratory tasks with simpler stimuli. For example, in the attentional boost effect (ABE), participants perform two tasks at once: memorizing a series of briefly presented stimuli (e.g., pictures of outdoor scenes) for a later memory test, and responding to other concurrently presented cues that meet pre-defined criteria (e.g., participants press a button for a blue target square and do nothing for a red distractor square). However, rather than increasing dual-task interference, attending to a target cue boosts, rather than impairs, subsequent memory for concurrently presented information. In this review we describe current data on the extent and limitations of the attentional boost effect and whether it may be related to activity in the locus coeruleus neuromodulatory system. We suggest that insight into the mechanisms that produce the attentional boost effect may be found in recent advances in the locus coeruleus literature and from understanding of how the neurocognitive system handles stability and change in everyday events. We consequently propose updates to an early account of the attentional boost effect, the dual-task interaction model, to better ground it in what is currently known about event cognition and the role that the LC plays in regulating brain states.

The limited reach of surprise: Evidence against effects of surprise on memory for preceding elements of an event

When reflecting on the past, some of our strongest memories are for experiences that took us by surprise. Extensive research has backed this intuition that we are more likely to remember surprising moments than mundane ones. But what about the moments leading up to the surprise? Are we more likely to remember those as well? While surprise is a well-established modulator of memory, it is unknown whether memory for the entire event will be enhanced, or only for the surprising occurrence itself. We developed a novel paradigm utilising stop-motion films, depicting of a sequence of narrative events, in which specific occurrences could be replaced with surprising ones, while keeping the rest of the film unaltered. Using this design, we tested whether surprise exerts retroactive effects on memory, and specifically whether any potential effect would be confined to elements in the same event as the surprising occurrence. In a large cohort of participants (n = 340), we found strong evidence that surprise did not retroactively modulate memory, neither when participants were tested immediately after study nor when they were tested 24 hours later. We suggest two possible accounts for these findings: (1) that the components of an event are encoded as independent episodic elements (not as a cohesive unit), or (2) that surprise segments experience, sectioning off the preceding elements as a separate event.

A neural network model of when to retrieve and encode episodic memories

Recent human behavioral and neuroimaging results suggest that people are selective in when they encode and retrieve episodic memories. To explain these findings, we trained a memory-augmented neural network to use its episodic memory to support prediction of upcoming states in an environment where past situations sometimes reoccur. We found that the network learned to retrieve selectively as a function of several factors, including its uncertainty about the upcoming state. Additionally, we found that selectively encoding episodic memories at the end of an event (but not mid-event) led to better subsequent prediction performance. In all of these cases, the benefits of selective retrieval and encoding can be explained in terms of reducing the risk of retrieving irrelevant memories. Overall, these modeling results provide a resource-rational account of why episodic retrieval and encoding should be selective and lead to several testable predictions.

Event boundaries shape temporal organization of memory by resetting temporal context

In memory, our continuous experiences are broken up into discrete events. Boundaries between events are known to influence the temporal organization of memory. However, how and through which mechanism event boundaries shape temporal order memory (TOM) remains unknown. Across four experiments, we show that event boundaries exert a dual role: improving TOM for items within an event and impairing TOM for items across events. Decreasing event length in a list enhances TOM, but only for items at earlier local event positions, an effect we term the local primacy effect. A computational model, in which items are associated to a temporal context signal that drifts over time but resets at boundaries captures all behavioural results. Our findings provide a unified algorithmic mechanism for understanding how and why event boundaries affect TOM, reconciling a long-standing paradox of why both contextual similarity and dissimilarity promote TOM.

Our memory is temporally organized, but our internal clock can be distorted. The authors demonstrate how environmental changes (termed event boundaries) affect memory for event order, and provide a computational model to explain these effects.

Switching task sets creates event boundaries in memory

People segregate continuously unfolding experiences into discrete events in memory. This process, known as event segmentation, results in better memory for the temporal order of experiences within an event and expands subjective temporal distance for items encoded across event boundaries. Previous research has suggested that the creation of event boundaries is driven by (typically unpredicted) changes in external stimulation, though many prior studies have confounded a change in bottom-up input with a concurrent change in task goal. This raises the question of whether event segmentation can be triggered by the endogenous cognitive control processes involved in switching task sets, independent of changes in bottom-up stimulation. We investigated this question by embedding task set changes during encoding of a series of trial-unique images, and comparing subsequent temporal order and distance memory for item pairs encoded across a change in task set with item pairs encoded within the same task set. Across five experiments, we demonstrate that both cued and voluntary task set changes are sufficient to create event boundaries, while ruling out potential confounding effects of shifts in stimulus set, response set, task cues, and task difficulty. Thus, internal control processes are a key determinant of segmenting episodic memories, and task set updating can trigger event segmentation independent of any externally induced, perceptual or task-based prediction error.

Reward at encoding but not retrieval modulates memory for detailed events

Much of the evidence suggesting that rewards improve memory performance has focused on how explicit rewards facilitate encoding of simplistic stimuli. To expand beyond this focus, the current study tested how explicit rewards presented at encoding as well as retrieval facilitate memory for information contained within complex events. In a single experimental session, participants (N = 88) encoded videos depicting naturalistic events (e.g., getting dressed) and then completed a recognition test probing their memory for different detail types (i.e., event, perceptual, or contextual) from the video stimuli. We manipulated the explicit reward associated with each video, such that accurate memory responses for half the videos were associated with high monetary incentives and half were associated with low monetary incentives. This reward manipulation was presented at either encoding or retrieval during a recognition memory test. The reward manipulation only affected memory when presented at encoding and this effect did not depend on the type of detail probed. Drift Diffusion Modelling further revealed that presenting reward information at encoding engendered greater encoding fidelity-indexed by an increase in drift rate-but did not change response caution at the time of retrieval-indexed by response threshold. Together, our results suggest that presenting reward information when encoding but not retrieving complex events has a general facilitatory effect, likely via attentional processing, on the ability to later remember precise details from the event.

Prediction Errors for Aversive Events Shape Long-Term Memory Formation through a Distinct Neural Mechanism

Prediction errors (PEs) have been known for decades to guide associative learning, but their role in episodic memory formation has been discovered only recently. To identify the neural mechanisms underlying the impact of aversive PEs on long-term memory formation, we used functional magnetic resonance imaging, while participants saw a series of unique stimuli and estimated the probability that an aversive shock would follow. Our behavioral data showed that negative PEs (i.e., omission of an expected outcome) were associated with superior recognition of the predictive stimuli, whereas positive PEs (i.e., presentation of an unexpected outcome) impaired subsequent memory. While medial temporal lobe (MTL) activity during stimulus encoding was overall associated with enhanced memory, memory-enhancing effects of negative PEs were linked to even decreased MTL activation. Additional large-scale network analyses showed PE-related increases in crosstalk between the "salience network" and a frontoparietal network commonly implicated in memory formation for expectancy-congruent events. These effects could not be explained by mere changes in physiological arousal or the prediction itself. Our results suggest that the superior memory for events associated with negative aversive PEs is driven by a potentially distinct neural mechanism that might serve to set these memories apart from those with expected outcomes.

Mnemonic prediction errors promote detailed memories

When our experience violates our predictions, it is adaptive to update our knowledge to promote a more accurate representation of the world and facilitate future predictions. Theoretical models propose that these mnemonic prediction errors should be encoded into a distinct memory trace to prevent interference with previous, conflicting memories. We investigated this proposal by repeatedly exposing participants to pairs of sequentially presented objects (A → B), thus evoking expectations. Then, we violated participants' expectations by replacing the second object in the pairs with a novel object (A → C). The following item memory test required participants to discriminate between identical old items and similar lures, thus testing detailed and distinctive item memory representations. In two experiments, mnemonic prediction errors enhanced item memory: Participants correctly identified more old items as old when those items violated expectations during learning, compared with items that did not violate expectations. This memory enhancement for C items was only observed when participants later showed intact memory for the related A → B pairs, suggesting that strong predictions are required to facilitate memory for violations. Following up on this, a third experiment reduced prediction strength prior to violation and subsequently eliminated the memory advantage of violations. Interestingly, mnemonic prediction errors did not increase gist-based mistakes of identifying old items as similar lures or identifying similar lures as old. Enhanced item memory in the absence of gist-based mistakes suggests that violations enhanced memory for items' details, which could be mediated via distinct memory traces. Together, these results advance our knowledge of how mnemonic prediction errors promote memory formation.

Mixing memory and desire: How memory reactivation supports deliberative decision-making

Memories affect nearly every aspect of our mental life. They allow us to both resolve uncertainty in the present and to construct plans for the future. Recently, renewed interest in the role memory plays in adaptive behavior has led to new theoretical advances and empirical observations. We review key findings, with particular emphasis on how the retrieval of many kinds of memories affects deliberative action selection. These results are interpreted in a sequential inference framework, in which reinstatements from memory serve as "samples" of potential action outcomes. The resulting model suggests a central role for the dynamics of memory reactivation in determining the influence of different kinds of memory in decisions. We propose that representation-specific dynamics can implement a bottom-up "product of experts" rule that integrates multiple sets of action-outcome predictions weighted based on their uncertainty. We close by reviewing related findings and identifying areas for further research. This article is categorized under: Psychology > Reasoning and Decision Making Neuroscience > Cognition Neuroscience > Computation.

Adaptive learning is structure learning in time

People use information flexibly. They often combine multiple sources of relevant information over time in order to inform decisions with little or no interference from intervening irrelevant sources. They adjust the degree to which they use new information over time rationally in accordance with environmental statistics and their own uncertainty. They can even use information gained in one situation to solve a problem in a very different one. Learning flexibly rests on the ability to infer the context at a given time, and therefore knowing which pieces of information to combine and which to separate. We review the psychological and neural mechanisms behind adaptive learning and structure learning to outline how people pool together relevant information, demarcate contexts, prevent interference between information collected in different contexts, and transfer information from one context to another. By examining all of these processes through the lens of optimal inference we bridge concepts from multiple fields to provide a unified multi-system view of how the brain exploits structure in time to optimize learning.

Replay in minds and machines

Experience-related brain activity patterns reactivate during sleep, wakeful rest, and brief pauses from active behavior. In parallel, machine learning research has found that experience replay can lead to substantial performance improvements in artificial agents. Together, these lines of research suggest replay has a variety of computational benefits for decision-making and learning. Here, we provide an overview of putative computational functions of replay as suggested by machine learning and neuroscientific research. We show that replay can lead to faster learning, less forgetting, reorganization or augmentation of experiences, and support planning and generalization. In addition, we highlight the benefits of reactivating abstracted internal representations rather than veridical memories, and discuss how replay could provide a mechanism to build internal representations that improve learning and decision-making.

Reward prediction errors drive declarative learning irrespective of agency

Recent years have witnessed a steady increase in the number of studies investigating the role of reward prediction errors (RPEs) in declarative learning. Specifically, in several experimental paradigms, RPEs drive declarative learning, with larger and more positive RPEs enhancing declarative learning. However, it is unknown whether this RPE must derive from the participant's own response, or whether instead, any RPE is sufficient to obtain the learning effect. To test this, we generated RPEs in the same experimental paradigm where we combined an agency and a nonagency condition. We observed no interaction between RPE and agency, suggesting that any RPE (irrespective of its source) can drive declarative learning. This result holds implications for declarative learning theory.

Human Representation Learning

The central theme of this review is the dynamic interaction between information selection and learning. We pose a fundamental question about this interaction: How do we learn what features of our experiences are worth learning about? In humans, this process depends on attention and memory, two cognitive functions that together constrain representations of the world to features that are relevant for goal attainment. Recent evidence suggests that the representations shaped by attention and memory are themselves inferred from experience with each task. We review this evidence and place it in the context of work that has explicitly characterized representation learning as statistical inference. We discuss how inference can be scaled to real-world decisions by approximating beliefs based on a small number of experiences. Finally, we highlight some implications of this inference process for human decision-making in social environments.

Signed and unsigned reward prediction errors dynamically enhance learning and memory

Memory helps guide behavior, but which experiences from the past are prioritized? Classic models of learning posit that events associated with unpredictable outcomes as well as, paradoxically, predictable outcomes, deploy more attention and learning for those events. Here, we test reinforcement learning and subsequent memory for those events, and treat signed and unsigned reward prediction errors (RPEs), experienced at the reward-predictive cue or reward outcome, as drivers of these two seemingly contradictory signals. By fitting reinforcement learning models to behavior, we find that both RPEs contribute to learning by modulating a dynamically changing learning rate. We further characterize the effects of these RPE signals on memory and show that both signed and unsigned RPEs enhance memory, in line with midbrain dopamine and locus-coeruleus modulation of hippocampal plasticity, thereby reconciling separate findings in the literature.

A predictive account of how novelty influences declarative memory

A rich body of studies in the human and non-human literature has examined the question how novelty influences memory. For a variety of different stimuli, ranging from simple objects and words to vastly complex scenarios, the literature reports that novelty improves memory in some cases, but impairs memory in other cases. In recent attempts to reconcile these conflicting findings, novelty has been divided into different subtypes, such as relative versus absolute novelty, or stimulus versus contextual novelty. Nevertheless, a single overarching theory of novelty and memory has been difficult to attain, probably due to the complexities in the interactions among stimuli, environmental factors (e.g., spatial and temporal context) and level of prior knowledge (but see Duszkiewicz et al., 2019; Kafkas & Montaldi, 2018b; Schomaker & Meeter, 2015). Here we describe how a predictive coding framework might be able to shed new light on different types of novelty and how they affect declarative memory in humans. More precisely, we consider how prior expectations modulate the influence of novelty on encoding episodes into memory, e.g., in terms of surprise, and how novelty/surprise affect memory for surrounding information. By reviewing a range of behavioural findings and their possible underlying neurobiological mechanisms, we highlight where a predictive coding framework succeeds and where it appears to struggle.

Behavioral, Physiological, and Neural Signatures of Surprise during Naturalistic Sports Viewing

Surprise signals a discrepancy between past and current beliefs. It is theorized to be linked to affective experiences, the creation of particularly resilient memories, and segmentation of the flow of experience into discrete perceived events. However, the ability to precisely measure naturalistic surprise has remained elusive. We used advanced basketball analytics to derive a quantitative measure of surprise and characterized its behavioral, physiological, and neural correlates in human subjects observing basketball games. We found that surprise was associated with segmentation of ongoing experiences, as reflected by subjectively perceived event boundaries and shifts in neocortical patterns underlying belief states. Interestingly, these effects differed by whether surprising moments contradicted or bolstered current predominant beliefs. Surprise also positively correlated with pupil dilation, activation in subcortical regions associated with dopamine, game enjoyment, and long-term memory. These investigations support key predictions from event segmentation theory and extend theoretical conceptualizations of surprise to real-world contexts.