Generative replay underlies compositional inference in the hippocampal-prefrontal circuit

Awake replay: off the clock but on the job

Hippocampal replay is widely thought to support two key cognitive functions: online decision-making and offline memory consolidation. In this review, we take a closer look at the hypothesized link between awake replay and online decision-making in rodents, and find only marginal evidence in support of this role. By contrast, the consolidation view is bolstered by new computational ideas and recent data, suggesting that (i) replay performs offline fictive learning for later goal-oriented behavior; and (ii) replay tags memories prior to sleep, prioritizing them for consolidation. Based on these recent advances, we favor an updated and refined role for awake replay - that is, supporting prioritized offline learning and tagging outside the hippocampus - rather than a direct, online role in guiding behavior.

Complementary roles for hippocampus and anterior cingulate in composing continuous choice

Naturalistic, goal directed behavior often requires continuous actions directed at dynamically changing goals. In this context, the closest analogue to choice is a strategic reweighting of multiple goal-specific control policies in response to shifting environmental pressures. To understand the algorithmic and neural bases of choice in continuous contexts, we examined behavior and brain activity in humans performing a continuous prey-pursuit task. Using a newly developed control-theoretic decomposition of behavior, we find pursuit strategies are well described by a meta-controller dictating a mixture of lower-level controllers, each linked to specific pursuit goals. Examining hippocampus and anterior cingulate cortex (ACC) population dynamics during goal switches revealed distinct roles for the two regions in parameterizing continuous controller mixing and meta-control. Hippocampal ensemble dynamics encoded the controller blending dynamics, suggesting it implements a mixing of goal-specific control policies. In contrast, ACC ensemble activity exhibited value-dependent ramping activity before goal switches, linking it to a meta-control process that accumulates evidence for switching goals. Our results suggest that hippocampus and ACC play complementary roles corresponding to a generalizable mixture controller and meta-controller that dictates value dependent changes in controller mixing.

Machine learning-based algorithm of drug-resistant prediction in newly diagnosed patients with temporal lobe epilepsy

OBJECTIVES

To develop a predicted algorithm for drug-resistant epilepsy (DRE) in newly diagnosed temporal lobe epilepsy (TLE) patients.

METHODS

A total of 139 newly diagnosed TLE patients were prospectively enrolled, and long-term video EEG monitoring was recorded. Clinical evaluations, including seizure frequency and antiseizure medications (ASMs) usage, were collected and prospectively followed up for 24 months. Interictal EEG data were used for feature extraction, identifying 216 EEG network features. Traditional machine learning and ensemble learning techniques were employed to predict DRE outcomes.

RESULTS

Over two years, TLE patients with DRE exhibited significant EEG differences, particularly in frontotemporal θ-band networks, characterized by increased connectivity metrics such as phase lag index (P = 0.000), etc. The predictive algorithm based on EEG features achieved accuracies between 59.2 %-84.6 % (AUC: 0.60-0.87). When compared to the whole brain, EEG features of the frontotemporal network showed improved classification performance in Naïve Bayes (P = 0.032), Tree Bagger (P = 0.021), and Subspace Discriminant (P = 0.022) models. The ensemble learning technique (Tree Bagger) delivered the best prediction results, achieving 91.5 % accuracy, 97 % sensitivity, 81 % specificity, and AUC of 0.92.

CONCLUSIONS

Increased frontotemporal EEG connectivity was observed in TLE patients with 2-year DRE. A predictive model based on routine EEG provides an accessible method for forecasting ASMs efficacy.

SIGNIFICANCE

This study highlights the clinical utility of EEG-based algorithms in identifying DRE early, aiding personalized treatment strategies and improving patient outcomes.

Exploring replay

Animals face uncertainty about their environments due to initial ignorance or subsequent changes. They therefore need to explore. However, the algorithmic structure of exploratory choices in the brain still remains largely elusive. Artificial agents face the same problem, and a venerable idea in reinforcement learning is that they can plan appropriate exploratory choices offline, during the equivalent of quiet wakefulness or sleep. Although offline processing in humans and other animals, in the form of hippocampal replay and preplay, has recently been the subject of highly informative modelling, existing methods only apply to known environments. Thus, they cannot predict exploratory replay choices during learning and/or behaviour in the face of uncertainty. Here, we extend an influential theory of hippocampal replay and examine its potential role in approximately optimal exploration, deriving testable predictions for the patterns of exploratory replay choices in a paradigmatic spatial navigation task. Our modelling provides a normative interpretation of the available experimental data suggestive of exploratory replay. Furthermore, we highlight the importance of sequence replay, and license a range of new experimental paradigms that should further our understanding of offline processing.

Cognitive maps and schizophrenia

Structured internal representations ('cognitive maps') shape cognition, from imagining the future and counterfactual past, to transferring knowledge to new settings. Our understanding of how such representations are formed and maintained in biological and artificial neural networks has grown enormously. The cognitive mapping hypothesis of schizophrenia extends this enquiry to psychiatry, proposing that diverse symptoms - from delusions to conceptual disorganization - stem from abnormalities in how the brain forms structured representations. These abnormalities may arise from a confluence of neurophysiological perturbations (excitation-inhibition imbalance, resulting in attractor instability and impaired representational capacity) and/or environmental factors such as early life psychosocial stressors (which impinge on representation learning). This proposal thus links knowledge of neural circuit abnormalities, environmental risk factors, and symptoms.

The time course and organization of hippocampal replay

The mechanisms by which the brain replays neural activity sequences remain unknown. Recording from large ensembles of hippocampal place cells in freely behaving rats, we observed that replay content is strictly organized over multiple timescales and governed by self-avoidance. After movement cessation, replays avoided the animal's previous path for 3 seconds. Chains of replays avoided self-repetition over a shorter timescale. We used a continuous attractor model of neural activity to demonstrate that neuronal fatigue both generates replay sequences and produces self-avoidance over the observed timescales. In addition, replay of past experience became predominant later into the stopping period, in a manner requiring cortical input. These results indicate a mechanism for replay generation that unexpectedly constrains which sequences can be produced across time.

A tale of two algorithms: Structured slots explain prefrontal sequence memory and are unified with hippocampal cognitive maps

Remembering events is crucial to intelligent behavior. Flexible memory retrieval requires a cognitive map and is supported by two key brain systems: hippocampal episodic memory (EM) and prefrontal working memory (WM). Although an understanding of EM is emerging, little is understood of WM beyond simple memory retrieval. We develop a mathematical theory relating the algorithms and representations of EM and WM by unveiling a duality between storing memories in synapses versus neural activity. This results in a formalism of prefrontal WM as structured, controllable neural subspaces (activity slots) representing dynamic cognitive maps without synaptic plasticity. Using neural networks, we elucidate differences, similarities, and trade-offs between the hippocampal and prefrontal algorithms. Lastly, we show that prefrontal representations in tasks from list learning to cue-dependent recall are unified as controllable activity slots. Our results unify frontal and temporal representations of memory and offer a new understanding for dynamic prefrontal representations of WM.

Decomposing dynamical subprocesses for compositional generalization

A striking feature of human cognition is an exceptional ability to rapidly adapt to novel situations. It is proposed this relies on abstracting and generalizing past experiences. While previous research has explored how humans detect and generalize single sequential processes, we have a limited understanding of how humans adapt to more naturalistic scenarios, for example, complex, multisubprocess environments. Here, we propose a candidate computational mechanism that posits compositional generalization of knowledge about subprocess dynamics. In two samples (N = 238 and N = 137), we combined a novel sequence learning task and computational modeling to ask whether humans extract and generalize subprocesses compositionally to solve new problems. In prior learning, participants experienced sequences of compound images formed from two graphs' product spaces (group 1: G1 and G2, group 2: G3 and G4). In transfer learning, both groups encountered compound images from the product of G1 and G3, composed entirely of new images. We show that subprocess knowledge transferred between task phases, such that in a new task environment each group had enhanced accuracy in predicting subprocess dynamics they had experienced during prior learning. Computational models utilizing predictive representations, based solely on the temporal contiguity of experienced task states, without an ability to transfer knowledge, failed to explain these data. Instead, behavior was consistent with a predictive representation model that maps task states between prior and transfer learning. These results help advance a mechanistic understanding of how humans discover and abstract subprocesses composing their experiences and compositionally reuse prior knowledge as a scaffolding for new experiences.

Human hippocampal and entorhinal neurons encode the temporal structure of experience

Extracting the underlying temporal structure of experience is a fundamental aspect of learning and memory that allows us to predict what is likely to happen next. Current knowledge about the neural underpinnings of this cognitive process in humans stems from functional neuroimaging research1-5. As these methods lack direct access to the neuronal level, it remains unknown how this process is computed by neurons in the human brain. Here we record from single neurons in individuals who have been implanted with intracranial electrodes for clinical reasons, and show that human hippocampal and entorhinal neurons gradually modify their activity to encode the temporal structure of a complex image presentation sequence. This representation was formed rapidly, without providing specific instructions to the participants, and persisted when the prescribed experience was no longer present. Furthermore, the structure recovered from the population activity of hippocampal-entorhinal neurons closely resembled the structural graph defining the sequence, but at the same time, also reflected the probability of upcoming stimuli. Finally, learning of the sequence graph was related to spontaneous, time-compressed replay of individual neurons' activity corresponding to previously experienced graph trajectories. These findings demonstrate that neurons in the hippocampus and entorhinal cortex integrate the 'what' and 'when' information to extract durable and predictive representations of the temporal structure of human experience.

A tradeoff between efficiency and robustness in the hippocampal-neocortical memory network during human and rodent sleep

Sleep constitutes a brain state of disengagement from the external world that supports memory consolidation and restores cognitive resources. The precise mechanisms how sleep and its varied stages support information processing remain largely unknown. Synaptic scaling models imply that daytime learning accumulates neural information, which is then consolidated and downregulated during sleep. Currently, there is a lack of in-vivo data from humans and rodents that elucidate if, and how, sleep renormalizes information processing capacities. From an information-theoretical perspective, a consolidation process should entail a reduction in neural pattern variability over the course of a night. Here, in a cross-species intracranial study, we identify a tradeoff in the neural population code during sleep where information coding efficiency is higher in the neocortex than in hippocampal archicortex in humans than in rodents as well as during wakefulness compared to sleep. Critically, non-REM sleep selectively reduces information coding efficiency through pattern repetition in the neocortex in both species, indicating a transition to a more robust information coding regime. Conversely, the coding regime in the hippocampus remained consistent from wakefulness to non-REM sleep. These findings suggest that new information could be imprinted to the long-term mnemonic storage in the neocortex through pattern repetition during sleep. Lastly, our results show that task engagement increased coding efficiency, while medically-induced unconsciousness disrupted the population code. In sum, these findings suggest that neural pattern variability could constitute a fundamental principle underlying cognitive engagement and memory formation, while pattern repetition reflects robust coding, possibly underlying the consolidation process.

Distractor effects in decision making are related to the individual's style of integrating choice attributes

Humans make irrational decisions in the presence of irrelevant distractor options. There is little consensus on whether decision making is facilitated or impaired by the presence of a highly rewarding distractor, or whether the distractor effect operates at the level of options' component attributes rather than at the level of their overall value. To reconcile different claims, we argue that it is important to consider the diversity of people's styles of decision making and whether choice attributes are combined in an additive or multiplicative way. Employing a multi-laboratory dataset investigating the same experimental paradigm, we demonstrated that people used a mix of both approaches and the extent to which approach was used varied across individuals. Critically, we identified that this variability was correlated with the distractor effect during decision making. Individuals who tended to use a multiplicative approach to compute value, showed a positive distractor effect. In contrast, a negative distractor effect (divisive normalisation) was prominent in individuals tending towards an additive approach. Findings suggest that the distractor effect is related to how value is constructed, which in turn may be influenced by task and subject specificities. This concurs with recent behavioural and neuroscience findings that multiple distractor effects co-exist.

Hippocampal representations of alternative possibilities are flexibly generated to meet cognitive demands

The cognitive ability to go beyond the present to consider alternative possibilities, including potential futures and counterfactual pasts, can support adaptive decision making. Complex and changing real-world environments, however, have many possible alternatives. Whether and how the brain can select among them to represent alternatives that meet current cognitive needs remains unknown. We therefore examined neural representations of alternative spatial locations in the rat hippocampus during navigation in a complex patch foraging environment with changing reward probabilities. We found representations of multiple alternatives along paths ahead and behind the animal, including in distant alternative patches. Critically, these representations were modulated in distinct patterns across successive trials: alternative paths were represented proportionate to their evolving relative value and predicted subsequent decisions, whereas distant alternatives were prevalent during value updating. These results demonstrate that the brain modulates the generation of alternative possibilities in patterns that meet changing cognitive needs for adaptive behavior.

Group Coordination Catalyzes Individual and Cultural Intelligence

A large program of research has aimed to ground large-scale cultural phenomena in processes taking place within individual minds. For example, investigating whether individual agents equipped with the right social learning strategies can enable cumulative cultural evolution given long enough time horizons. However, this approach often omits the critical group-level processes that mediate between individual agents and multi-generational societies. Here, we argue that interacting groups are a necessary and explanatory level of analysis, linking individual and collective intelligence through two characteristic feedback loops. In the first loop, more sophisticated individual-level social learning mechanisms based on Theory of Mind facilitate group-level complementarity, allowing distributed knowledge to be compositionally recombined in groups; these group-level innovations, in turn, ease the cognitive load on individuals. In the second loop, societal-level processes of cumulative culture provide groups with new cognitive technologies, including shared language and conceptual abstractions, which set in motion new group-level processes to further coordinate, recombine, and innovate. Taken together, these cycles establish group-level interaction as a dual engine of intelligence, catalyzing both individual cognition and cumulative culture.

Replay-triggered brain-wide activation in humans

The consolidation of discrete experiences into a coherent narrative shapes the cognitive map, providing structured mental representations of our experiences. In this process, past memories are reactivated and replayed in sequence, fostering hippocampal-cortical dialogue. However, brain-wide engagement coinciding with sequential reactivation (or replay) of memories remains largely unexplored. In this study, employing simultaneous EEG-fMRI, we capture both the spatial and temporal dynamics of memory replay. We find that during mental simulation, past memories are replayed in fast sequences as detected via EEG. These transient replay events are associated with heightened fMRI activity in the hippocampus and medial prefrontal cortex. Replay occurrence strengthens functional connectivity between the hippocampus and the default mode network, a set of brain regions key to representing the cognitive map. On the other hand, when subjects are at rest following learning, memory reactivation of task-related items is stronger than that of pre-learning rest, and is also associated with heightened hippocampal activation and augmented hippocampal connectivity to the entorhinal cortex. Together, our findings highlight a distributed, brain-wide engagement associated with transient memory reactivation and its sequential replay.

A recurrent network model of planning explains hippocampal replay and human behavior

When faced with a novel situation, people often spend substantial periods of time contemplating possible futures. For such planning to be rational, the benefits to behavior must compensate for the time spent thinking. Here, we capture these features of behavior by developing a neural network model where planning itself is controlled by the prefrontal cortex. This model consists of a meta-reinforcement learning agent augmented with the ability to plan by sampling imagined action sequences from its own policy, which we call 'rollouts'. In a spatial navigation task, the agent learns to plan when it is beneficial, which provides a normative explanation for empirical variability in human thinking times. Additionally, the patterns of policy rollouts used by the artificial agent closely resemble patterns of rodent hippocampal replays. Our work provides a theory of how the brain could implement planning through prefrontal-hippocampal interactions, where hippocampal replays are triggered by-and adaptively affect-prefrontal dynamics.

Prioritizing replay when future goals are unknown

Although hippocampal place cells replay nonlocal trajectories, the computational function of these events remains controversial. One hypothesis, formalized in a prominent reinforcement learning account, holds that replay plans routes to current goals. However, recent puzzling data appear to contradict this perspective by showing that replayed destinations lag current goals. These results may support an alternative hypothesis that replay updates route information to build a "cognitive map." Yet no similar theory exists to formalize this view, and it is unclear how such a map is represented or what role replay plays in computing it. We address these gaps by introducing a theory of replay that learns a map of routes to candidate goals, before reward is available or when its location may change. Our work extends the planning account to capture a general map-building function for replay, reconciling it with data, and revealing an unexpected relationship between the seemingly distinct hypotheses.

Active inference as a theory of sentient behavior

This review paper offers an overview of the history and future of active inference-a unifying perspective on action and perception. Active inference is based upon the idea that sentient behavior depends upon our brains' implicit use of internal models to predict, infer, and direct action. Our focus is upon the conceptual roots and development of this theory of (basic) sentience and does not follow a rigid chronological narrative. We trace the evolution from Helmholtzian ideas on unconscious inference, through to a contemporary understanding of action and perception. In doing so, we touch upon related perspectives, the neural underpinnings of active inference, and the opportunities for future development. Key steps in this development include the formulation of predictive coding models and related theories of neuronal message passing, the use of sequential models for planning and policy optimization, and the importance of hierarchical (temporally) deep internal (i.e., generative or world) models. Active inference has been used to account for aspects of anatomy and neurophysiology, to offer theories of psychopathology in terms of aberrant precision control, and to unify extant psychological theories. We anticipate further development in all these areas and note the exciting early work applying active inference beyond neuroscience. This suggests a future not just in biology, but in robotics, machine learning, and artificial intelligence.