Distributional reinforcement learning in prefrontal cortex.

The prefrontal cortex is crucial for learning and decision-making. Classic reinforcement learning (RL) theories center on learning the expectation of potential rewarding outcomes and explain a wealth of neural data in the prefrontal cortex. Distributional RL, on the other hand, learns the full distribution of rewarding outcomes and better explains dopamine responses. In the present study, we show that distributional RL also better explains macaque anterior cingulate cortex neuronal responses, suggesting that it is a common mechanism for reward-guided learning.

The human brain reactivates context-specific past information at event boundaries of naturalistic experiences.

Although we perceive the world in a continuous manner, our experience is partitioned into discrete events. However, to make sense of these events, they must be stitched together into an overarching narrative-a model of unfolding events. It has been proposed that such a stitching process happens in offline neural reactivations when rodents build models of spatial environments. Here we show that, while understanding a natural narrative, humans reactivate neural representations of past events. Similar to offline replay, these reactivations occur in the hippocampus and default mode network, where reactivations are selective to relevant past events. However, these reactivations occur, not during prolonged offline periods, but at the boundaries between ongoing narrative events. These results, replicated across two datasets, suggest reactivations as a candidate mechanism for binding temporally distant information into a coherent understanding of ongoing experience.

A cognitive map for value-guided choice in ventromedial prefrontal cortex.

The prefrontal cortex is crucial for economic decision-making and representing the value of options. However, how such representations facilitate flexible decisions remains unknown. We reframe economic decision-making in prefrontal cortex in line with representations of structure within the medial temporal lobe because such cognitive map representations are known to facilitate flexible behaviour. Specifically, we framed choice between different options as a navigation process in value space. Here we show that choices in a 2D value space defined by reward magnitude and probability were represented with a grid-like code, analogous to that found in spatial navigation. The grid-like code was present in ventromedial prefrontal cortex (vmPFC) local field potential theta frequency and the result replicated in an independent dataset. Neurons in vmPFC similarly contained a grid-like code, in addition to encoding the linear value of the chosen option. Importantly, both signals were modulated by theta frequency - occurring at theta troughs but on separate theta cycles. Furthermore, we found sharp-wave ripples - a key neural signature of planning and flexible behaviour - in vmPFC, which were modulated by accuracy and reward. These results demonstrate that multiple cognitive map-like computations are deployed in vmPFC during economic decision-making, suggesting a new framework for the implementation of choice in prefrontal cortex.

How to build a cognitive map.

Learning and interpreting the structure of the environment is an innate feature of biological systems, and is integral to guiding flexible behaviors for evolutionary viability. The concept of a cognitive map has emerged as one of the leading metaphors for these capacities, and unraveling the learning and neural representation of such a map has become a central focus of neuroscience. In recent years, many models have been developed to explain cellular responses in the hippocampus and other brain areas. Because it can be difficult to see how these models differ, how they relate and what each model can contribute, this Review aims to organize these models into a clear ontology. This ontology reveals parallels between existing empirical results, and implies new approaches to understand hippocampal-cortical interactions and beyond.

Complementary task representations in hippocampus and prefrontal cortex for generalizing the structure of problems.

Humans and other animals effortlessly generalize prior knowledge to solve novel problems, by abstracting common structure and mapping it onto new sensorimotor specifics. To investigate how the brain achieves this, in this study, we trained mice on a series of reversal learning problems that shared the same structure but had different physical implementations. Performance improved across problems, indicating transfer of knowledge. Neurons in medial prefrontal cortex (mPFC) maintained similar representations across problems despite their different sensorimotor correlates, whereas hippocampal (dCA1) representations were more strongly influenced by the specifics of each problem. This was true for both representations of the events that comprised each trial and those that integrated choices and outcomes over multiple trials to guide an animal's decisions. These data suggest that prefrontal cortex and hippocampus play complementary roles in generalization of knowledge: PFC abstracts the common structure among related problems, and hippocampus maps this structure onto the specifics of the current situation.

Decoding cognition from spontaneous neural activity.

In human neuroscience, studies of cognition are rarely grounded in non-task-evoked, 'spontaneous' neural activity. Indeed, studies of spontaneous activity tend to focus predominantly on intrinsic neural patterns (for example, resting-state networks). Taking a 'representation-rich' approach bridges the gap between cognition and resting-state communities: this approach relies on decoding task-related representations from spontaneous neural activity, allowing quantification of the representational content and rich dynamics of such activity. For example, if we know the neural representation of an episodic memory, we can decode its subsequent replay during rest. We argue that such an approach advances cognitive research beyond a focus on immediate task demand and provides insight into the functional relevance of the intrinsic neural pattern (for example, the default mode network). This in turn enables a greater integration between human and animal neuroscience, facilitating experimental testing of theoretical accounts of intrinsic activity, and opening new avenues of research in psychiatry.

Reinforcement learning: Dopamine ramps with fuzzy value estimates.

A new study in reinforcement learning theory shows that extending the temporal difference algorithm to unbiased learning under state uncertainty explains the observed ramping behaviour of dopamine neurons.

Experience replay is associated with efficient nonlocal learning.

To make effective decisions, people need to consider the relationship between actions and outcomes. These are often separated by time and space. The neural mechanisms by which disjoint actions and outcomes are linked remain unknown. One promising hypothesis involves neural replay of nonlocal experience. Using a task that segregates direct from indirect value learning, combined with magnetoencephalography, we examined the role of neural replay in human nonlocal learning. After receipt of a reward, we found significant backward replay of nonlocal experience, with a 160-millisecond state-to-state time lag, which was linked to efficient learning of action values. Backward replay and behavioral evidence of nonlocal learning were more pronounced for experiences of greater benefit for future behavior. These findings support nonlocal replay as a neural mechanism for solving complex credit assignment problems during learning.

The Tolman-Eichenbaum Machine: Unifying Space and Relational Memory through Generalization in the Hippocampal Formation.

The hippocampal-entorhinal system is important for spatial and relational memory tasks. We formally link these domains, provide a mechanistic understanding of the hippocampal role in generalization, and offer unifying principles underlying many entorhinal and hippocampal cell types. We propose medial entorhinal cells form a basis describing structural knowledge, and hippocampal cells link this basis with sensory representations. Adopting these principles, we introduce the Tolman-Eichenbaum machine (TEM). After learning, TEM entorhinal cells display diverse properties resembling apparently bespoke spatial responses, such as grid, band, border, and object-vector cells. TEM hippocampal cells include place and landmark cells that remap between environments. Crucially, TEM also aligns with empirically recorded representations in complex non-spatial tasks. TEM also generates predictions that hippocampal remapping is not random as previously believed; rather, structural knowledge is preserved across environments. We confirm this structural transfer over remapping in simultaneously recorded place and grid cells.

Transferring structural knowledge across cognitive maps in humans and models.

Relations between task elements often follow hidden underlying structural forms such as periodicities or hierarchies, whose inferences fosters performance. However, transferring structural knowledge to novel environments requires flexible representations that are generalizable over particularities of the current environment, such as its stimuli and size. We suggest that humans represent structural forms as abstract basis sets and that in novel tasks, the structural form is inferred and the relevant basis set is transferred. Using a computational model, we show that such representation allows inference of the underlying structural form, important task states, effective behavioural policies and the existence of unobserved state-trajectories. In two experiments, participants learned three abstract graphs during two successive days. We tested how structural knowledge acquired on Day-1 affected Day-2 performance. In line with our model, participants who had a correct structural prior were able to infer the existence of unobserved state-trajectories and appropriate behavioural policies.

Neuronal Computation Underlying Inferential Reasoning in Humans and Mice.

Every day we make decisions critical for adaptation and survival. We repeat actions with known consequences. But we also draw on loosely related events to infer and imagine the outcome of entirely novel choices. These inferential decisions are thought to engage a number of brain regions; however, the underlying neuronal computation remains unknown. Here, we use a multi-day cross-species approach in humans and mice to report the functional anatomy and neuronal computation underlying inferential decisions. We show that during successful inference, the mammalian brain uses a hippocampal prospective code to forecast temporally structured learned associations. Moreover, during resting behavior, coactivation of hippocampal cells in sharp-wave/ripples represent inferred relationships that include reward, thereby "joining-the-dots" between events that have not been observed together but lead to profitable outcomes. Computing mnemonic links in this manner may provide an important mechanism to build a cognitive map that stretches beyond direct experience, thus supporting flexible behavior.

Episodic memory retrieval success is associated with rapid replay of episode content.

Retrieval of everyday experiences is fundamental for informing our future decisions. The fine-grained neurophysiological mechanisms that support such memory retrieval are largely unknown. We studied participants who first experienced, without repetition, unique multicomponent 40-80-s episodes. One day later, they engaged in cued retrieval of these episodes while undergoing magnetoencephalography. By decoding individual episode elements, we found that trial-by-trial successful retrieval was supported by the sequential replay of episode elements, with a temporal compression factor of >60. The direction of replay supporting retrieval, either backward or forward, depended on whether the task goal was to retrieve elements of an episode that followed or preceded, respectively, a retrieval cue. This sequential replay was weaker in very-high-performing participants, in whom instead we found evidence for simultaneous clustered reactivation. Our results demonstrate that memory-mediated decisions are supported by a rapid replay mechanism that can flexibly shift in direction in response to task goals.

Reinforcement Learning: Full Glass or Empty - Depends Who You Ask.

An extension of the prediction error theory of dopamine, imported from artificial intelligence, represents the full distribution over future rewards rather than only the average and better explains dopamine responses.

Human Replay Spontaneously Reorganizes Experience.

Knowledge abstracted from previous experiences can be transferred to aid new learning. Here, we asked whether such abstract knowledge immediately guides the replay of new experiences. We first trained participants on a rule defining an ordering of objects and then presented a novel set of objects in a scrambled order. Across two studies, we observed that representations of these novel objects were reactivated during a subsequent rest. As in rodents, human "replay" events occurred in sequences accelerated in time, compared to actual experience, and reversed their direction after a reward. Notably, replay did not simply recapitulate visual experience, but followed instead a sequence implied by learned abstract knowledge. Furthermore, each replay contained more than sensory representations of the relevant objects. A sensory code of object representations was preceded 50 ms by a code factorized into sequence position and sequence identity. We argue that this factorized representation facilitates the generalization of a previously learned structure to new objects.

The Hippocampus and Neocortical Inhibitory Engrams Protect against Memory Interference.

Our experiences often overlap with each other, yet we are able to selectively recall individual memories to guide decisions and future actions. The neural mechanisms that support such precise memory recall remain unclear. Here, using ultra-high field 7T MRI we reveal two distinct mechanisms that protect memories from interference. The first mechanism involves the hippocampus, where the blood-oxygen-level-dependent (BOLD) signal predicts behavioral measures of memory interference, and representations of context-dependent memories are pattern separated according to their relational overlap. The second mechanism involves neocortical inhibition. When we reduce the concentration of neocortical GABA using trans-cranial direct current stimulation (tDCS), neocortical memory interference increases in proportion to the reduction in GABA, which in turn predicts behavioral performance. These findings suggest that memory interference is mediated by both the hippocampus and neocortex, where the hippocampus separates overlapping but context-dependent memories using relational information, and neocortical inhibition prevents unwanted co-activation between overlapping memories.

What Is a Cognitive Map? Organizing Knowledge for Flexible Behavior.

It is proposed that a cognitive map encoding the relationships between entities in the world supports flexible behavior, but the majority of the neural evidence for such a system comes from studies of spatial navigation. Recent work describing neuronal parallels between spatial and non-spatial behaviors has rekindled the notion of a systematic organization of knowledge across multiple domains. We review experimental evidence and theoretical frameworks that point to principles unifying these apparently disparate functions. These principles describe how to learn and use abstract, generalizable knowledge and suggest that map-like representations observed in a spatial context may be an instance of general coding mechanisms capable of organizing knowledge of all kinds. We highlight how artificial agents endowed with such principles exhibit flexible behavior and learn map-like representations observed in the brain. Finally, we speculate on how these principles may offer insight into the extreme generalizations, abstractions, and inferences that characterize human cognition.

Triple dissociation of attention and decision computations across prefrontal cortex.

Naturalistic decision-making typically involves sequential deployment of attention to choice alternatives to gather information before a decision is made. Attention filters how information enters decision circuits, thus implying that attentional control may shape how decision computations unfold. We recorded neuronal activity from three subregions of the prefrontal cortex (PFC) while monkeys performed an attention-guided decision-making task. From the first saccade to decision-relevant information, a triple dissociation of decision- and attention-related computations emerged in parallel across PFC subregions. During subsequent saccades, orbitofrontal cortex activity reflected the value comparison between currently and previously attended information. In contrast, the anterior cingulate cortex carried several signals reflecting belief updating in light of newly attended information, the integration of evidence to a decision bound and an emerging plan for what action to choose. Our findings show how anatomically dissociable PFC representations evolve during attention-guided information search, supporting computations critical for value-guided choice.

Simultaneous representation of a spectrum of dynamically changing value estimates during decision making.

Decisions are based on value expectations derived from experience. We show that dorsal anterior cingulate cortex and three other brain regions hold multiple representations of choice value based on different timescales of experience organized in terms of systematic gradients across the cortex. Some parts of each area represent value estimates based on recent reward experience while others represent value estimates based on experience over the longer term. The value estimates within these areas interact with one another according to their temporal scaling. Some aspects of the representations change dynamically as the environment changes. The spectrum of value estimates may act as a flexible selection mechanism for combining experience-derived value information with other aspects of value to allow flexible and adaptive decisions in changing environments.

Repetition suppression: a means to index neural representations using BOLD?

Understanding how the human brain gives rise to complex cognitive processes remains one of the biggest challenges of contemporary neuroscience. While invasive recording in animal models can provide insight into neural processes that are conserved across species, our understanding of cognition more broadly relies upon investigation of the human brain itself. There is therefore an imperative to establish non-invasive tools that allow human brain activity to be measured at high spatial and temporal resolution. In recent years, various attempts have been made to refine the coarse signal available in functional magnetic resonance imaging (fMRI), providing a means to investigate neural activity at the meso-scale, i.e. at the level of neural populations. The most widely used techniques include repetition suppression and multivariate pattern analysis. Human neuroscience can now use these techniques to investigate how representations are encoded across neural populations and transformed by relevant computations. Here, we review the physiological basis, applications and limitations of fMRI repetition suppression with a brief comparison to multivariate techniques. By doing so, we show how fMRI repetition suppression holds promise as a tool to reveal complex neural mechanisms that underlie human cognitive function.This article is part of the themed issue 'Interpreting BOLD: a dialogue between cognitive and cellular neuroscience'.

The Human Connectome Project's neuroimaging approach.

Noninvasive human neuroimaging has yielded many discoveries about the brain. Numerous methodological advances have also occurred, though inertia has slowed their adoption. This paper presents an integrated approach to data acquisition, analysis and sharing that builds upon recent advances, particularly from the Human Connectome Project (HCP). The 'HCP-style' paradigm has seven core tenets: (i) collect multimodal imaging data from many subjects; (ii) acquire data at high spatial and temporal resolution; (iii) preprocess data to minimize distortions, blurring and temporal artifacts; (iv) represent data using the natural geometry of cortical and subcortical structures; (v) accurately align corresponding brain areas across subjects and studies; (vi) analyze data using neurobiologically accurate brain parcellations; and (vii) share published data via user-friendly databases. We illustrate the HCP-style paradigm using existing HCP data sets and provide guidance for future research. Widespread adoption of this paradigm should accelerate progress in understanding the brain in health and disease.