A Probability Distribution over Latent Causes, in the Orbitofrontal Cortex

The Ubiquity of Time in Latent-cause Inference

Humans have an outstanding ability to generalize from past experiences, which requires parsing continuously experienced events into discrete, coherent units, and relating them to similar past experiences. Time is a key element in this process; however, how temporal information is used in generalization remains unclear. Latent-cause inference provides a Bayesian framework for clustering experiences, by building a world model in which related experiences are generated by a shared cause. Here, we examine how temporal information is used in latent-cause inference, using a novel task in which participants see "microbe" stimuli and explicitly report the latent cause ("strain") they infer for each microbe. We show that humans incorporate time in their inference of latent causes, such that recently inferred latent causes are more likely to be inferred again. In particular, a "persistent" model, in which the latent cause inferred for one observation has a fixed probability of continuing to cause the next observation, explains the data significantly better than two other time-sensitive models, although extensive individual differences exist. We show that our task and this model have good psychometric properties, highlighting their potential use for quantifying individual differences in computational psychiatry or in neuroimaging studies.

Belief inference for hierarchical hidden states in spatial navigation

Uncertainty abounds in the real world, and in environments with multiple layers of unobservable hidden states, decision-making requires resolving uncertainties based on mutual inference. Focusing on a spatial navigation problem, we develop a Tiger maze task that involved simultaneously inferring the local hidden state and the global hidden state from probabilistically uncertain observation. We adopt a Bayesian computational approach by proposing a hierarchical inference model. Applying this to human task behaviour, alongside functional magnetic resonance brain imaging, allows us to separate the neural correlates associated with reinforcement and reassessment of belief in hidden states. The imaging results also suggest that different layers of uncertainty differentially involve the basal ganglia and dorsomedial prefrontal cortex, and that the regions responsible are organised along the rostral axis of these areas according to the type of inference and the level of abstraction of the hidden state, i.e. higher-order state inference involves more anterior parts.

Integrating and fragmenting memories under stress and alcohol

Stress can powerfully influence the way we form memories, particularly the extent to which they are integrated or situated within an underlying spatiotemporal and broader knowledge architecture. These different representations in turn have significant consequences for the way we use these memories to guide later behavior. Puzzlingly, although stress has historically been argued to promote fragmentation, leading to disjoint memory representations, more recent work suggests that stress can also facilitate memory binding and integration. Understanding the circumstances under which stress fosters integration will be key to resolving this discrepancy and unpacking the mechanisms by which stress can shape later behavior. Here, we examine memory integration at multiple levels: linking together the content of an individual experience, threading associations between related but distinct events, and binding an experience into a pre-existing schema or sense of causal structure. We discuss neural and cognitive mechanisms underlying each form of integration as well as findings regarding how stress, aversive learning, and negative affect can modulate each. In this analysis, we uncover that stress can indeed promote each level of integration. We also show how memory integration may apply to understanding effects of alcohol, highlighting extant clinical and preclinical findings and opportunities for further investigation. Finally, we consider the implications of integration and fragmentation for later memory-guided behavior, and the importance of understanding which type of memory representation is potentiated in order to design appropriate interventions.

Bayesian inference is facilitated by modular neural networks with different time scales

Various animals, including humans, have been suggested to perform Bayesian inferences to handle noisy, time-varying external information. In performing Bayesian inference by the brain, the prior distribution must be acquired and represented by sampling noisy external inputs. However, the mechanism by which neural activities represent such distributions has not yet been elucidated. Our findings reveal that networks with modular structures, composed of fast and slow modules, are adept at representing this prior distribution, enabling more accurate Bayesian inferences. Specifically, the modular network that consists of a main module connected with input and output layers and a sub-module with slower neural activity connected only with the main module outperformed networks with uniform time scales. Prior information was represented specifically by the slow sub-module, which could integrate observed signals over an appropriate period and represent input means and variances. Accordingly, the neural network could effectively predict the time-varying inputs. Furthermore, by training the time scales of neurons starting from networks with uniform time scales and without modular structure, the above slow-fast modular network structure and the division of roles in which prior knowledge is selectively represented in the slow sub-modules spontaneously emerged. These results explain how the prior distribution for Bayesian inference is represented in the brain, provide insight into the relevance of modular structure with time scale hierarchy to information processing, and elucidate the significance of brain areas with slower time scales.

Latent-state and model-based learning in PTSD

Post-traumatic stress disorder (PTSD) is characterized by altered emotional and behavioral responding following a traumatic event. In this article, we review the concepts of latent-state and model-based learning (i.e., learning and inferring abstract task representations) and discuss their relevance for clinical and neuroscience models of PTSD. Recent data demonstrate evidence for brain and behavioral biases in these learning processes in PTSD. These new data potentially recast excessive fear towards trauma cues as a problem in learning and updating abstract task representations, as opposed to traditional conceptualizations focused on stimulus-specific learning. Biases in latent-state and model-based learning may also be a common mechanism targeted in common therapies for PTSD. We highlight key knowledge gaps that need to be addressed to further elaborate how latent-state learning and its associated neurocircuitry mechanisms function in PTSD and how to optimize treatments to target these processes.

Decoding context memories for threat in large-scale neural networks

Humans are often tasked with determining the degree to which a given situation poses threat. Salient cues present during prior events help bring online memories for context, which plays an informative role in this process. However, it is relatively unknown whether and how individuals use features of the environment to retrieve context memories for threat, enabling accurate inferences about the current level of danger/threat (i.e. retrieve appropriate memory) when there is a degree of ambiguity surrounding the present context. We leveraged computational neuroscience approaches (i.e. independent component analysis and multivariate pattern analyses) to decode large-scale neural network activity patterns engaged during learning and inferring threat context during a novel functional magnetic resonance imaging task. Here, we report that individuals accurately infer threat contexts under ambiguous conditions through neural reinstatement of large-scale network activity patterns (specifically striatum, salience, and frontoparietal networks) that track the signal value of environmental cues, which, in turn, allows reinstatement of a mental representation, primarily within a ventral visual network, of the previously learned threat context. These results provide novel insight into distinct, but overlapping, neural mechanisms by which individuals may utilize prior learning to effectively make decisions about ambiguous threat-related contexts as they navigate the environment.

The Computational and Neural Bases of Context-Dependent Learning

Flexible behavior requires the creation, updating, and expression of memories to depend on context. While the neural underpinnings of each of these processes have been intensively studied, recent advances in computational modeling revealed a key challenge in context-dependent learning that had been largely ignored previously: Under naturalistic conditions, context is typically uncertain, necessitating contextual inference. We review a theoretical approach to formalizing context-dependent learning in the face of contextual uncertainty and the core computations it requires. We show how this approach begins to organize a large body of disparate experimental observations, from multiple levels of brain organization (including circuits, systems, and behavior) and multiple brain regions (most prominently the prefrontal cortex, the hippocampus, and motor cortices), into a coherent framework. We argue that contextual inference may also be key to understanding continual learning in the brain. This theory-driven perspective places contextual inference as a core component of learning.

Generalization of cognitive maps across space and time

Prominent theories posit that associative memory structures, known as cognitive maps, support flexible generalization of knowledge across cognitive domains. Here, we evince a representational account of cognitive map flexibility by quantifying how spatial knowledge formed one day was used predictively in a temporal sequence task 24 hours later, biasing both behavior and neural response. Participants learned novel object locations in distinct virtual environments. After learning, hippocampus and ventromedial prefrontal cortex (vmPFC) represented a cognitive map, wherein neural patterns became more similar for same-environment objects and more discriminable for different-environment objects. Twenty-four hours later, participants rated their preference for objects from spatial learning; objects were presented in sequential triplets from either the same or different environments. We found that preference response times were slower when participants transitioned between same- and different-environment triplets. Furthermore, hippocampal spatial map coherence tracked behavioral slowing at the implicit sequence transitions. At transitions, predictive reinstatement of virtual environments decreased in anterior parahippocampal cortex. In the absence of such predictive reinstatement after sequence transitions, hippocampus and vmPFC responses increased, accompanied by hippocampal-vmPFC functional decoupling that predicted individuals' behavioral slowing after a transition. Collectively, these findings reveal how expectations derived from spatial experience generalize to support temporal prediction.

Task state representations in vmPFC mediate relevant and irrelevant value signals and their behavioral influence

The ventromedial prefrontal-cortex (vmPFC) is known to contain expected value signals that inform our choices. But expected values even for the same stimulus can differ by task. In this study, we asked how the brain flexibly switches between such value representations in a task-dependent manner. Thirty-five participants alternated between tasks in which either stimulus color or motion predicted rewards. We show that multivariate vmPFC signals contain a rich representation that includes the current task state or context (motion/color), the associated expected value, and crucially, the irrelevant value of the alternative context. We also find that irrelevant value representations in vmPFC compete with relevant value signals, interact with task-state representations and relate to behavioral signs of value competition. Our results shed light on vmPFC's role in decision making, bridging between its role in mapping observations onto the task states of a mental map, and computing expected values for multiple states.

Fast rule switching and slow rule updating in a perceptual categorization task

To adapt to a changing world, we must be able to switch between rules already learned and, at other times, learn rules anew. Often we must do both at the same time, switching between known rules while also constantly re-estimating them. Here, we show these two processes, rule switching and rule learning, rely on distinct but intertwined computations, namely fast inference and slower incremental learning. To this end, we studied how monkeys switched between three rules. Each rule was compositional, requiring the animal to discriminate one of two features of a stimulus and then respond with an associated eye movement along one of two different response axes. By modeling behavior, we found the animals learned the axis of response using fast inference (rule switching) while continuously re-estimating the stimulus-response associations within an axis (rule learning). Our results shed light on the computational interactions between rule switching and rule learning, and make testable neural predictions for these interactions.

Emotion dynamics as hierarchical Bayesian inference in time

What fundamental property of our environment would be most valuable and optimal in characterizing the emotional dynamics we experience in daily life? Empirical work has shown that an accurate estimation of uncertainty is necessary for our optimal perception, learning, and decision-making. However, the role of this uncertainty in governing our affective dynamics remains unexplored. Using Bayesian encoding, decoding and computational modeling, on a large-scale neuroimaging and behavioral data on a passive movie-watching task, we showed that emotions naturally arise due to ongoing uncertainty estimations about future outcomes in a hierarchical neural architecture. Several prefrontal subregions hierarchically encoded a lower-dimensional signal that highly correlated with the evolving uncertainty. Crucially, the lateral orbitofrontal cortex (lOFC) tracked the temporal fluctuations of this uncertainty and was predictive of the participants' predisposition to anxiety. Furthermore, we observed a distinct functional double-dissociation within OFC with increased connectivity between medial OFC and DMN, while with that of lOFC and FPN in response to the evolving affect. Finally, we uncovered a temporally predictive code updating an individual's beliefs spontaneously with fluctuating outcome uncertainty in the lOFC. A biologically relevant and computationally crucial parameter in the theories of brain function, we propose uncertainty to be central to the definition of complex emotions.

Value representations in the rodent orbitofrontal cortex drive learning, not choice

Humans and animals make predictions about the rewards they expect to receive in different situations. In formal models of behavior, these predictions are known as value representations, and they play two very different roles. Firstly, they drive choice: the expected values of available options are compared to one another, and the best option is selected. Secondly, they support learning: expected values are compared to rewards actually received, and future expectations are updated accordingly. Whether these different functions are mediated by different neural representations remains an open question. Here, we employ a recently developed multi-step task for rats that computationally separates learning from choosing. We investigate the role of value representations in the rodent orbitofrontal cortex, a key structure for value-based cognition. Electrophysiological recordings and optogenetic perturbations indicate that these representations do not directly drive choice. Instead, they signal expected reward information to a learning process elsewhere in the brain that updates choice mechanisms.

Neural mechanisms of credit assignment for inferred relationships in a structured world

Animals abstract compact representations of a task's structure, which supports accelerated learning and flexible behavior. Whether and how such abstracted representations may be used to assign credit for inferred, but unobserved, relationships in structured environments are unknown. We develop a hierarchical reversal-learning task and Bayesian learning model to assess the computational and neural mechanisms underlying how humans infer specific choice-outcome associations via structured knowledge. We find that the medial prefrontal cortex (mPFC) efficiently represents hierarchically related choice-outcome associations governed by the same latent cause, using a generalized code to assign credit for both experienced and inferred outcomes. Furthermore, the mPFC and lateral orbitofrontal cortex track the current "position" within a latent association space that generalizes over stimuli. Collectively, these findings demonstrate the importance of both tracking the current position in an abstracted task space and efficient, generalizable representations in the prefrontal cortex for supporting flexible learning and inference in structured environments.

Neural representations of others' traits predict social decisions

SignificancePeople's decisions about how to treat others are known to be influenced by societally shared expectations about the typical traits of people from particular social groups (stereotypes). We combined a social psychological framework, an economic game, and multivariate functional MRI analysis to investigate whether and how trait inferences are instantiated neurally in the service of behavior toward members of different social groups. Multidimensional representations of trait content were found in brain regions associated with social cognition and in a region associated with inference-based decision-making: the lateral orbitofrontal cortex (OFC). Only OFC representations predicted individual participants' behavior, suggesting that although stereotypes are also represented in social cognition regions, they exert influence on behavior via decision-making mechanisms centered in the OFC.

Simulating the computational mechanisms of cognitive and behavioral psychotherapeutic interventions: insights from active inference

Cognitive-behavioral therapy (CBT) leverages interactions between thoughts, feelings, and behaviors. To deepen understanding of these interactions, we present a computational (active inference) model of CBT that allows formal simulations of interactions between cognitive interventions (i.e., cognitive restructuring) and behavioral interventions (i.e., exposure) in producing adaptive behavior change (i.e., reducing maladaptive avoidance behavior). Using spider phobia as a concrete example of maladaptive avoidance more generally, we show simulations indicating that when conscious beliefs about safety/danger have strong interactions with affective/behavioral outcomes, behavioral change during exposure therapy is mediated by changes in these beliefs, preventing generalization. In contrast, when these interactions are weakened, and cognitive restructuring only induces belief uncertainty (as opposed to strong safety beliefs), behavior change leads to generalized learning (i.e., "over-writing" the implicit beliefs about action-outcome mappings that directly produce avoidance). The individual is therefore equipped to face any new context, safe or dangerous, remaining in a content state without the need for avoidance behavior-increasing resilience from a CBT perspective. These results show how the same changes in behavior during CBT can be due to distinct underlying mechanisms; they predict lower rates of relapse when cognitive interventions focus on inducing uncertainty and on reducing the effects of automatic negative thoughts on behavior.

Reduced decision bias and more rational decision making following ventromedial prefrontal cortex damage

Human decisions are susceptible to biases, but establishing causal roles of brain areas has proved to be difficult. Here we studied decision biases in 17 people with unilateral medial prefrontal cortex damage and a rare patient with bilateral ventromedial prefrontal cortex (vmPFC) lesions. Participants learned to choose which of two options was most likely to win, and then bet money on the outcome. Thus, good performance required not only selecting the best option, but also the amount to bet. Healthy people were biased by their previous bet, as well as by the unchosen option's value. Unilateral medial prefrontal lesions reduced these biases, leading to more rational decisions. Bilateral vmPFC lesions resulted in more strategic betting, again with less bias from the previous trial, paradoxically improving performance overall. Together, the results suggest that vmPFC normally imposes contextual biases, which in healthy people may actually be suboptimal in some situations.

The orbitofrontal cartographer

Theories of orbitofrontal cortex (OFC) function have evolved substantially over the last few decades. There is now a general consensus that the OFC is important for predicting aspects of future events and for using these predictions to guide behavior. Yet the precise content of these predictions and the degree to which OFC contributes to agency contingent upon them has become contentious, with several plausible theories advocating different answers to these questions. In this review we will focus on three of these ideas-the economic value, credit assignment, and cognitive map hypotheses-describing both their successes and failures. We will propose that these failures hint at a more nuanced and perhaps unique role for the OFC, particularly the lateral subdivision, in supporting the proposed functions when an underlying model or map of the causal structures in the environment must be constructed or updated. (PsycInfo Database Record (c) 2021 APA, all rights reserved).

Neural representation of abstract task structure during generalization

Cognitive models in psychology and neuroscience widely assume that the human brain maintains an abstract representation of tasks. This assumption is fundamental to theories explaining how we learn quickly, think creatively, and act flexibly. However, neural evidence for a verifiably generative abstract task representation has been lacking. Here, we report an experimental paradigm that requires forming such a representation to act adaptively in novel conditions without feedback. Using functional magnetic resonance imaging, we observed that abstract task structure was represented within left mid-lateral prefrontal cortex, bilateral precuneus, and inferior parietal cortex. These results provide support for the neural instantiation of the long-supposed abstract task representation in a setting where we can verify its influence. Such a representation can afford massive expansions of behavioral flexibility without additional experience, a vital characteristic of human cognition.

Neural encoding of task-dependent errors during adaptive learning

Effective learning requires using errors in a task-dependent manner, for example adjusting to errors that result from unpredicted environmental changes but ignoring errors that result from environmental stochasticity. Where and how the brain represents errors in a task-dependent manner and uses them to guide behavior are not well understood. We imaged the brains of human participants performing a predictive-inference task with two conditions that had different sources of errors. Their performance was sensitive to this difference, including more choice switches after fundamental changes versus stochastic fluctuations in reward contingencies. Using multi-voxel pattern classification, we identified task-dependent representations of error magnitude and past errors in posterior parietal cortex. These representations were distinct from representations of the resulting behavioral adjustments in dorsomedial frontal, anterior cingulate, and orbitofrontal cortex. The results provide new insights into how the human brain represents errors in a task-dependent manner and guides subsequent adaptive behavior.

Revaluing the Role of vmPFC in the Acquisition of Pavlovian Threat Conditioning in Humans

The role of the ventromedial prefrontal cortex (vmPFC) in human pavlovian threat conditioning has been relegated largely to the extinction or reversal of previously acquired stimulus-outcome associations. However, recent neuroimaging evidence questions this view by also showing activity in the vmPFC during threat acquisition. Here we investigate the casual role of vmPFC in the acquisition of pavlovian threat conditioning by assessing skin conductance response (SCR) and declarative memory of stimulus-outcome contingencies during a differential pavlovian threat-conditioning paradigm in eight patients with a bilateral vmPFC lesion, 10 with a lesion outside PFC and 10 healthy participants (each group included both females and males). Results showed that patients with vmPFC lesion failed to produce a conditioned SCR during threat acquisition, despite no evidence of compromised SCR to unconditioned stimulus or compromised declarative memory for stimulus-outcome contingencies. These results suggest that the vmPFC plays a causal role in the acquisition of new learning and not just in the extinction or reversal of previously acquired learning, as previously thought. Given the role of the vmPFC in schema-related processing and latent structure learning, the vmPFC may be required to construct a detailed representation of the task, which is needed to produce a sustained conditioned physiological response in anticipation of the unconditioned stimulus during threat acquisition.SIGNIFICANCE STATEMENT Pavlovian threat conditioning is an adaptive mechanism through which organisms learn to avoid potential threats, thus increasing their chances of survival. Understanding what brain regions contribute to such a process is crucial to understand the mechanisms underlying adaptive as well as maladaptive learning, and has the potential to inform the treatment of anxiety disorders. Importantly, the role of the ventromedial prefrontal cortex (vmPFC) in the acquisition of pavlovian threat conditioning has been relegated largely to the inhibition of previously acquired learning. Here, we show that the vmPFC actually plays a causal role in the acquisition of pavlovian threat conditioning.

Neural Systems for Memory-based Value Judgment and Decision-making

Real-life choices often require that we draw inferences about the value of options based on structured, schematic knowledge about their utility for our current goals. Other times, value information may be retrieved directly from a specific prior experience with an option. In an fMRI experiment, we investigated the neural systems involved in retrieving and assessing information from different memory sources to support value-based choice. Participants completed a task in which items could be conferred positive or negative value based on schematic associations (i.e., schema value) or learned directly from experience via deterministic feedback (i.e., experienced value). We found that ventromedial pFC (vmPFC) activity correlated with the influence of both experience- and schema-based values on participants' decisions. Connectivity between the vmPFC and middle temporal cortex also tracked the inferred value of items based on schematic associations on the first presentation of ingredients, before any feedback. In contrast, the striatum responded to participants' willingness to bet on ingredients as a function of the unsigned strength of their memory for those options' values. These results argue that the striatum and vmPFC play distinct roles in memory-based value judgment and decision-making. Specifically, the vmPFC assesses the value of options based on information inferred from schematic knowledge and retrieved from prior direct experience, whereas the striatum controls a decision to act on options based on memory strength.

Map Making: Constructing, Combining, and Inferring on Abstract Cognitive Maps

Cognitive maps enable efficient inferences from limited experience that can guide novel decisions. We tested whether the hippocampus (HC), entorhinal cortex (EC), and ventromedial prefrontal cortex (vmPFC)/medial orbitofrontal cortex (mOFC) organize abstract and discrete relational information into a cognitive map to guide novel inferences. Subjects learned the status of people in two unseen 2D social hierarchies, with each dimension learned on a separate day. Although one dimension was behaviorally relevant, multivariate activity patterns in HC, EC, and vmPFC/mOFC were linearly related to the Euclidean distance between people in the mentally reconstructed 2D space. Hubs created unique comparisons between the hierarchies, enabling inferences between novel pairs. We found that both behavior and neural activity in EC and vmPFC/mOFC reflected the Euclidean distance to the retrieved hub, which was reinstated in HC. These findings reveal how abstract and discrete relational structures are represented, are combined, and enable novel inferences in the human brain.

Towards a computational psychiatry of juvenile obsessive-compulsive disorder

Obsessive-Compulsive Disorder (OCD) most often emerges during adolescence, but we know little about the aberrant neural and cognitive developmental mechanisms that underlie its emergence during this critical developmental period. To move towards a computational psychiatry of juvenile OCD, we review studies on the computational, neuropsychological and neural alterations in juvenile OCD and link these findings to the adult OCD and cognitive neuroscience literature. We find consistent difficulties in tasks entailing complex decision making and set shifting, but limited evidence in other areas that are altered in adult OCD, such as habit and confidence formation. Based on these findings, we establish a neurocomputational framework that illustrates how cognition can go awry and lead to symptoms of juvenile OCD. We link these possible aberrant neural processes to neuroimaging findings in juvenile OCD and show that juvenile OCD is mainly characterised by disruptions of complex reasoning systems.

Base rate neglect and neural computations for subjective weight in decision under uncertainty

Base rate neglect, an important bias in estimating probability of uncertain events, describes humans' tendency to underweight base rate (prior) relative to individuating information (likelihood). However, the neural mechanisms that give rise to this bias remain elusive. In this study, subjects chose between uncertain prospects where estimating reward probability was essential. We found that when the variability of prior and likelihood information about reward probability were systematically manipulated, prior variability significantly affected the degree to which subjects underweight the base rate of reward probability. Activity in the orbitofrontal cortex, medial prefrontal cortex, and putamen represented the relative subjective weight that reflected such bias. Further, sensitivity to likelihood relative to prior variability in the putamen correlated with individuals' overall tendency to underweight base rate. These findings suggest that in combining prior and likelihood, relative sensitivity to information variability and subjective-weight computations critically contribute to the individual heterogeneity in base rate neglect.

Multiscale Computation and Dynamic Attention in Biological and Artificial Intelligence

Biological and artificial intelligence (AI) are often defined by their capacity to achieve a hierarchy of short-term and long-term goals that require incorporating information over time and space at both local and global scales. More advanced forms of this capacity involve the adaptive modulation of integration across scales, which resolve computational inefficiency and explore-exploit dilemmas at the same time. Research in neuroscience and AI have both made progress towards understanding architectures that achieve this. Insight into biological computations come from phenomena such as decision inertia, habit formation, information search, risky choices and foraging. Across these domains, the brain is equipped with mechanisms (such as the dorsal anterior cingulate and dorsolateral prefrontal cortex) that can represent and modulate across scales, both with top-down control processes and by local to global consolidation as information progresses from sensory to prefrontal areas. Paralleling these biological architectures, progress in AI is marked by innovations in dynamic multiscale modulation, moving from recurrent and convolutional neural networks-with fixed scalings-to attention, transformers, dynamic convolutions, and consciousness priors-which modulate scale to input and increase scale breadth. The use and development of these multiscale innovations in robotic agents, game AI, and natural language processing (NLP) are pushing the boundaries of AI achievements. By juxtaposing biological and artificial intelligence, the present work underscores the critical importance of multiscale processing to general intelligence, as well as highlighting innovations and differences between the future of biological and artificial intelligence.

Ventromedial prefrontal cortex compression during concept learning

Prefrontal cortex (PFC) is thought to support the ability to focus on goal-relevant information by filtering out irrelevant information, a process akin to dimensionality reduction. Here, we test this dimensionality reduction hypothesis by relating a data-driven approach to characterizing the complexity of neural representation with a theoretically-supported computational model of learning. We find evidence of goal-directed dimensionality reduction within human ventromedial PFC during learning. Importantly, by using computational predictions of each participant’s attentional strategies during learning, we find that that the degree of neural compression predicts an individual’s ability to selectively attend to concept-specific information. These findings suggest a domain-general mechanism of learning through compression in ventromedial PFC.

Efficient learning is akin to goal-directed dimensionality reduction, in which relevant information is highlighted and irrelevant input is ignored. Here, the authors show that ventromedial prefrontal cortex uniquely supports such learning by compressing neural codes to represent goal-specific information.

Learning task-state representations

Arguably, the most difficult part of learning is deciding what to learn about. Should I associate the positive outcome of safely completing a street-crossing with the situation 'the car approaching the crosswalk was red' or with 'the approaching car was slowing down'? In this Perspective, we summarize our recent research into the computational and neural underpinnings of 'representation learning'-how humans (and other animals) construct task representations that allow efficient learning and decision-making. We first discuss the problem of learning what to ignore when confronted with too much information, so that experience can properly generalize across situations. We then turn to the problem of augmenting perceptual information with inferred latent causes that embody unobservable task-relevant information, such as contextual knowledge. Finally, we discuss recent findings regarding the neural substrates of task representations that suggest the orbitofrontal cortex represents 'task states', deploying them for decision-making and learning elsewhere in the brain.

A flexible and generalizable model of online latent-state learning

Many models of classical conditioning fail to describe important phenomena, notably the rapid return of fear after extinction. To address this shortfall, evidence converged on the idea that learning agents rely on latent-state inferences, i.e. an ability to index disparate associations from cues to rewards (or penalties) and infer which index (i.e. latent state) is presently active. Our goal was to develop a model of latent-state inferences that uses latent states to predict rewards from cues efficiently and that can describe behavior in a diverse set of experiments. The resulting model combines a Rescorla-Wagner rule, for which updates to associations are proportional to prediction error, with an approximate Bayesian rule, for which beliefs in latent states are proportional to prior beliefs and an approximate likelihood based on current associations. In simulation, we demonstrate the model's ability to reproduce learning effects both famously explained and not explained by the Rescorla-Wagner model, including rapid return of fear after extinction, the Hall-Pearce effect, partial reinforcement extinction effect, backwards blocking, and memory modification. Lastly, we derive our model as an online algorithm to maximum likelihood estimation, demonstrating it is an efficient approach to outcome prediction. Establishing such a framework is a key step towards quantifying normative and pathological ranges of latent-state inferences in various contexts.

Learning to use past evidence in a sophisticated world model

Humans and other animals are able to discover underlying statistical structure in their environments and exploit it to achieve efficient and effective performance. However, such structure is often difficult to learn and use because it is obscure, involving long-range temporal dependencies. Here, we analysed behavioural data from an extended experiment with rats, showing that the subjects learned the underlying statistical structure, albeit suffering at times from immediate inferential imperfections as to their current state within it. We accounted for their behaviour using a Hidden Markov Model, in which recent observations are integrated with evidence from the past. We found that over the course of training, subjects came to track their progress through the task more accurately, a change that our model largely attributed to improved integration of past evidence. This learning reflected the structure of the task, decreasing reliance on recent observations, which were potentially misleading.

Humans and other animals possess the remarkable ability to find and exploit patterns and structures in their experience of a complex and varied world. However, such structures are often temporally extended and latent or hidden, being only partially correlated with immediate observations of the world. This makes it essential to integrate current and historical information, and creates a challenging statistical and computational problem. Here, we examine the behaviour of rats facing a version of this challenge posed by a brain-stimulation reward task. We find that subjects learned the general structure of the task, but struggled when immediate observations were misleading. We captured this behaviour with a model in which subjects integrated evidence from recent observations together with evidence from the past. The subjects’ performance improved markedly over successive sessions, allowing them to overcome misleading observations. According to our model, this was made possible by more effective usage of past evidence to better determine the true state of the world.

Pure correlates of exploration and exploitation in the human brain

Balancing exploration and exploitation is a fundamental problem in reinforcement learning. Previous neuroimaging studies of the exploration-exploitation dilemma could not completely disentangle these two processes, making it difficult to unambiguously identify their neural signatures. We overcome this problem using a task in which subjects can either observe (pure exploration) or bet (pure exploitation). Insula and dorsal anterior cingulate cortex showed significantly greater activity on observe trials compared to bet trials, suggesting that these regions play a role in driving exploration. A model-based analysis of task performance suggested that subjects chose to observe until a critical evidence threshold was reached. We observed a neural signature of this evidence accumulation process in the ventromedial prefrontal cortex. These findings support theories positing an important role for anterior cingulate cortex in exploration, while also providing a new perspective on the roles of insula and ventromedial prefrontal cortex.

Control of entropy in neural models of environmental state

Humans and animals construct internal models of their environment in order to select appropriate courses of action. The representation of uncertainty about the current state of the environment is a key feature of these models that controls the rate of learning as well as directly affecting choice behaviour. To maintain flexibility, given that uncertainty naturally decreases over time, most theoretical inference models include a dedicated mechanism to drive up model uncertainty. Here we probe the long-standing hypothesis that noradrenaline is involved in determining the uncertainty, or entropy, and thus flexibility, of neural models. Pupil diameter, which indexes neuromodulatory state including noradrenaline release, predicted increases (but not decreases) in entropy in a neural state model encoded in human medial orbitofrontal cortex, as measured using multivariate functional MRI. Activity in anterior cingulate cortex predicted pupil diameter. These results provide evidence for top-down, neuromodulatory control of entropy in neural state models.

Dissociable Forms of Uncertainty-Driven Representational Change Across the Human Brain

Environmental change can lead decision makers to shift rapidly among different behavioral regimes. These behavioral shifts can be accompanied by rapid changes in the firing pattern of neural networks. However, it is unknown what the populations of neurons that participate in such "network reset" phenomena are representing. Here, we investigated the following: (1) whether and where rapid changes in multivariate activity patterns are observable with fMRI during periods of rapid behavioral change and (2) what types of representations give rise to these phenomena. We did so by examining fluctuations in multivoxel patterns of BOLD activity from male and female human subjects making sequential inferences about the state of a partially observable and discontinuously changing variable. We found that, within the context of this sequential inference task, the multivariate patterns of activity in a number of cortical regions contain representations that change more rapidly during periods of uncertainty following a change in behavioral context. In motor cortex, this phenomenon was indicative of discontinuous change in behavioral outputs, whereas in visual regions, the same basic phenomenon was evoked by tracking of salient environmental changes. In most other cortical regions, including dorsolateral prefrontal and anterior cingulate cortex, the phenomenon was most consistent with directly encoding the degree of uncertainty. However, in a few other regions, including orbitofrontal cortex, the phenomenon was best explained by representations of a shifting context that evolve more rapidly during periods of rapid learning. These representations may provide a dynamic substrate for learning that facilitates rapid disengagement from learned responses during periods of change.SIGNIFICANCE STATEMENT Brain activity patterns tend to change more rapidly during periods of uncertainty and behavioral adjustment, yet the computational role of such rapid transitions is poorly understood. Here, we identify brain regions with fMRI BOLD activity patterns that change more rapidly during periods of behavioral adjustment and use computational modeling to attribute the phenomenon to specific causes. We demonstrate that the phenomenon emerges in different brain regions for different computational reasons, the most common being the representation of uncertainty itself, but that, in a selective subset of regions including orbitofrontal cortex, the phenomenon was best explained as a shifting latent state signal that may serve to control the degree to which recent temporal context affects ongoing expectations.

Understanding psychiatric disorder by capturing ecologically relevant features of learning and decision-making

Recent research in cognitive neuroscience has begun to uncover the processes underlying increasingly complex voluntary behaviours, including learning and decision-making. Partly this success has been possible by progressing from simple experimental tasks to paradigms that incorporate more ecological features. More specifically, the premise is that to understand cognitions and brain functions relevant for real life, we need to introduce some of the ecological challenges that we have evolved to solve. This often entails an increase in task complexity, which can be managed by using computational models to help parse complex behaviours into specific component mechanisms. Here we propose that using computational models with tasks that capture ecologically relevant learning and decision-making processes may provide a critical advantage for capturing the mechanisms underlying symptoms of disorders in psychiatry. As a result, it may help develop mechanistic approaches towards diagnosis and treatment. We begin this review by mapping out the basic concepts and models of learning and decision-making. We then move on to consider specific challenges that emerge in realistic environments and describe how they can be captured by tasks. These include changes of context, uncertainty, reflexive/emotional biases, cost-benefit decision-making, and balancing exploration and exploitation. Where appropriate we highlight future or current links to psychiatry. We particularly draw examples from research on clinical depression, a disorder that greatly compromises motivated behaviours in real-life, but where simpler paradigms have yielded mixed results. Finally, we highlight several paradigms that could be used to help provide new insights into the mechanisms of psychiatric disorders.

Dynamic integration of conceptual information during learning

The development and application of concepts is a critical component of cognition. Although concepts can be formed on the basis of simple perceptual or semantic features, conceptual representations can also capitalize on similarities across feature relationships. By representing these types of higher-order relationships, concepts can simplify the learning problem and facilitate decisions. Despite this, little is known about the neural mechanisms that support the construction and deployment of these kinds of higher-order concepts during learning. To address this question, we combined a carefully designed associative learning task with computational model-based functional magnetic resonance imaging (fMRI). Participants were scanned as they learned and made decisions about sixteen pairs of cues and associated outcomes. Associations were structured such that individual cues shared feature relationships, operationalized as shared patterns of cue pair-outcome associations. In order to capture the large number of possible conceptual representational structures that participants might employ and to evaluate how conceptual representations are used during learning, we leveraged a well-specified Bayesian computational model of category learning [1]. Behavioral and model-based results revealed that participants who displayed a tendency to link experiences in memory benefitted from faster learning rates, suggesting that the use of the conceptual structure in the task facilitated decisions about cue pair-outcome associations. Model-based fMRI analyses revealed that trial-by-trial integration of cue information into higher-order conceptual representations was supported by an anterior temporal (AT) network of regions previously implicated in representing complex conjunctions of features and meaning-based information.

Inactivation of Medial Frontal Cortex Changes Risk Preference

Humans and other animals need to make decisions under varying degrees of uncertainty. These decisions are strongly influenced by an individual's risk preference; however, the neuronal circuitry by which risk preference shapes choice is still unclear [1]. Supplementary eye field (SEF), an oculomotor area within primate medial frontal cortex, is thought to be an essential part of the neuronal circuit underlying oculomotor decision making, including decisions under risk [2-5]. Consistent with this view, risk-related action value and monitoring signals have been observed in SEF [6-8]. However, such activity has also been observed in other frontal areas, including orbitofrontal [9-11], cingulate [12-14], and dorsal-lateral frontal cortex [15]. It is thus unknown whether the activity in SEF causally contributes to risky decisions, or whether it is merely a reflection of neural processes in other cortical regions. Here, we tested a causal role of SEF in risky oculomotor choices. We found that SEF inactivation strongly reduced the frequency of risky choices. This reduction was largely due to a reduced attraction to reward uncertainty and high reward gain, but not due to changes in the subjective estimation of reward probability or average expected reward. Moreover, SEF inactivation also led to increased sensitivity to differences between expected and actual reward during free choice. Nevertheless, it did not affect adjustments of decisions based on reward history.

State-change decisions and dorsomedial prefrontal cortex: the importance of time

Different kinds of decision making can be categorized by their differential effect on the agent’s current and future states as well as the computational challenges they pose. Here, we draw a distinction between within-state and state-change decision-making, and propose that a dedicated decision mechanism exists in dorsomedial prefrontal cortex (dmPFC) that is specialized for state-change decisions. We set out a formal framework in which state change decisions may be made on the basis of the integrated momentary reward rate, over the intended time to be spent in a state. A key feature of this framework is that reward rate is expressed as a function of continuous time. We argue that dmPFC is suited for this type of decision making partly due to its ability to track the passage of time. This proposed function of dmPFC is placed in contrast to other evaluative systems such as the orbitofrontal cortex, which is important for careful deliberation within a specific model-space or option-space and within a decision strategy.

Neural Computations Underlying Causal Structure Learning

Behavioral evidence suggests that beliefs about causal structure constrain associative learning, determining which stimuli can enter into association, as well as the functional form of that association. Bayesian learning theory provides one mechanism by which structural beliefs can be acquired from experience, but the neural basis of this mechanism is poorly understood. We studied this question with a combination of behavioral, computational, and neuroimaging techniques. Male and female human subjects learned to predict an outcome based on cue and context stimuli while being scanned using fMRI. Using a model-based analysis of the fMRI data, we show that structure learning signals are encoded in posterior parietal cortex, lateral prefrontal cortex, and the frontal pole. These structure learning signals are distinct from associative learning signals. Moreover, representational similarity analysis and information mapping revealed that the multivariate patterns of activity in posterior parietal cortex and anterior insula encode the full posterior distribution over causal structures. Variability in the encoding of the posterior across subjects predicted variability in their subsequent behavioral performance. These results provide evidence for a neural architecture in which structure learning guides the formation of associations.SIGNIFICANCE STATEMENT Animals are able to infer the hidden structure behind causal relations between stimuli in the environment, allowing them to generalize this knowledge to stimuli they have never experienced before. A recently published computational model based on this idea provided a parsimonious account of a wide range of phenomena reported in the animal learning literature, suggesting a dedicated neural mechanism for learning this hidden structure. Here, we validate this model by measuring brain activity during a task that involves both structure learning and associative learning. We show that a distinct network of regions supports structure learning and that the neural signal corresponding to beliefs about structure predicts future behavioral performance.

Elevated Aggression and Reduced White Matter Integrity in Mild Traumatic Brain Injury: A DTI Study

Mild traumatic brain injury (mTBI) remains the most commonly reported head injury in the United States, and is associated with a wide range of post-concussive symptoms including physical, cognitive and affective impairments. Elevated aggression has been documented in mTBI; however, the neural mechanisms associated with aggression at the chronic stage of recovery remain poorly understood. In the present study, we investigated the association between white matter integrity and aggression in mTBI using diffusion tensor imaging (DTI). Twenty-six age-matched adults participated in the study, including 16 healthy controls (HCs) and 10 individuals in the chronic stage of recovery (either 6-months or 12 months post-mTBI). Psychological measures of aggression included the Buss-Perry Aggression Questionnaire and the Personality Assessment Inventory (PAI). Axonal pathways implicated in affective processing were studied, including the corpus callosum, anterior thalamic radiation, cingulum and uncinate fasciculus, and measures of white matter integrity included fractional anisotropy (FA), mean diffusivity (MD), radial diffusivity (RD) and axial diffusivity (AD). We found that adults with mTBI in the chronic stage of recovery had higher levels aggression. Individuals with mTBI also had greater RD in the corpus callosum compared to HCs, indicating reduced fiber integrity. Furthermore, we observed a significant association between reduced white matter integrity in the corpus callosum and greater aggression. Our findings provide additional evidence for underlying neuroanatomical mechanisms of aggression, although future research will be necessary to characterize the specific relationship between aggression and the white matter pathways we identified.

Memory integration constructs maps of space, time, and concepts

Recent evidence demonstrates that new events are learned in the context of their relationships to existing memories. Within the hippocampus and medial prefrontal cortex, related memories are represented by integrated codes that connect events experienced at different times and places. Integrated codes form the basis of spatial, temporal, and conceptual maps of experience. These maps represent information that goes beyond direct experience and support generalization behaviors that require knowledge be used in new ways. The degree to which an individual memory is integrated into a coherent map is determined by its spatial, temporal, and conceptual proximity to existing knowledge. Integration is observed over a wide range of scales, suggesting that memories contain information about both broad and fine-grained contexts.

Does mental context drift or shift?

Theories of episodic memory have generally proposed that individual memory traces are linked together by a representation of context that drifts slowly over time. Recent data challenge the notion that contextual drift is always slow and passive. In particular, changes in one's external environment or internal model induce discontinuities in memory that are reflected in sudden changes in neural activity, suggesting that context can shift abruptly. Furthermore, context change effects are sensitive to top-down goals, suggesting that contextual drift may be an active process. These findings call for revising models of the role of context in memory, in order to account for abrupt contextual shifts and the controllable nature of context change.

Computational approaches to fMRI analysis

Analysis methods in cognitive neuroscience have not always matched the richness of fMRI data. Early methods focused on estimating neural activity within individual voxels or regions, averaged over trials or blocks and modeled separately in each participant. This approach mostly neglected the distributed nature of neural representations over voxels, the continuous dynamics of neural activity during tasks, the statistical benefits of performing joint inference over multiple participants and the value of using predictive models to constrain analysis. Several recent exploratory and theory-driven methods have begun to pursue these opportunities. These methods highlight the importance of computational techniques in fMRI analysis, especially machine learning, algorithmic optimization and parallel computing. Adoption of these techniques is enabling a new generation of experiments and analyses that could transform our understanding of some of the most complex-and distinctly human-signals in the brain: acts of cognition such as thoughts, intentions and memories.

Complementary contributions of basolateral amygdala and orbitofrontal cortex to value learning under uncertainty

We make choices based on the values of expected outcomes, informed by previous experience in similar settings. When the outcomes of our decisions consistently violate expectations, new learning is needed to maximize rewards. Yet not every surprising event indicates a meaningful change in the environment. Even when conditions are stable overall, outcomes of a single experience can still be unpredictable due to small fluctuations (i.e., expected uncertainty) in reward or costs. In the present work, we investigate causal contributions of the basolateral amygdala (BLA) and orbitofrontal cortex (OFC) in rats to learning under expected outcome uncertainty in a novel delay-based task that incorporates both predictable fluctuations and directional shifts in outcome values. We demonstrate that OFC is required to accurately represent the distribution of wait times to stabilize choice preferences despite trial-by-trial fluctuations in outcomes, whereas BLA is necessary for the facilitation of learning in response to surprising events.

DOI:http://dx.doi.org/10.7554/eLife.27483.001

Nobody likes waiting – we opt for online shopping to avoid standing in lines, grow impatient in traffic, and often prefer restaurants that serve food quickly. When making decisions, humans and other animals try to maximize the benefits by weighing up the costs and rewards associated with a situation. Many regions in the brain help us choose the best options based on quality and size of rewards, and required waiting times. Even before we make decisions, the activity in these brain regions predicts what we will choose.

Sometimes, however, unexpected changes can lead to longer waiting times and our preferences suddenly become less desirable. The brain can detect such changes by comparing the outcomes we anticipate to those we experience. When the outcomes are surprising, specific areas in the brain such as the amygdala and the orbitofrontal cortex help us learn to make better choices. However, as surprising events can occur purely by chance, we need to be able to ignore irrelevant surprises and only learn from meaningful ones. Until now, it was not clear whether the amygdala and orbitofrontal cortex play specific roles in successfully learning under such conditions.

Stolyarova and Izquierdo trained rats to select between two images and rewarded them with sugar pellets after different delays. If rats chose one of these images they received the rewards after a predictable delay that was about 10 seconds, while choosing the other one produced variable delays – sometimes the time intervals were either very short or very long. Then, the waiting times for one of the alternatives changed unexpectedly. Rats with healthy brains quickly learned to choose the option with the shorter waiting time.

Stolyarova and Izquierdo repeated the experiments with rats that had damage in a part of the amygdala. These rats learned more slowly, particularly when the variable option changed for the better. Rats with damage to the orbitofrontal cortex failed to learn at all. Stolyarova and Izquierdo then examined the rats’ behavior during delays. Rats with damage to the orbitofrontal cortex could not distinguish between meaningful and irrelevant surprises and always looked for the food pellet (i.e. anticipated a reward) at the average delay interval.

These findings highlight two brain regions that help us distinguish meaningful surprises from irrelevant ones. A next step will be to examine how the amygdala and orbitofrontal cortex interact during learning and see if changes to the activity of these brain regions may affect responses. Advanced methods to non-invasively manipulate brain activity in humans may help people who find it hard to cope with changes; or individuals suffering from substance use disorders, who often struggle to give up drugs that provide them immediate and predictable rewards.

DOI:http://dx.doi.org/10.7554/eLife.27483.002

Lateral orbitofrontal cortex anticipates choices and integrates prior with current information

Adaptive behavior requires integrating prior with current information to anticipate upcoming events. Brain structures related to this computation should bring relevant signals from the recent past into the present. Here we report that rats can integrate the most recent prior information with sensory information, thereby improving behavior on a perceptual decision-making task with outcome-dependent past trial history. We find that anticipatory signals in the orbitofrontal cortex about upcoming choice increase over time and are even present before stimulus onset. These neuronal signals also represent the stimulus and relevant second-order combinations of past state variables. The encoding of choice, stimulus and second-order past state variables resides, up to movement onset, in overlapping populations. The neuronal representation of choice before stimulus onset and its build-up once the stimulus is presented suggest that orbitofrontal cortex plays a role in transforming immediate prior and stimulus information into choices using a compact state-space representation.

The orbitofrontal cortex encodes outcomes, expected rewards and values, but it is unclear how this region uses this information to inform action selection. Here, the authors show that lateral orbitofrontal cortex anticipates upcoming choices and combines recent prior information with current sensory information.