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.

Distinct neurocomputational mechanisms support informational and socially normative conformity.

A change of mind in response to social influence could be driven by informational conformity to increase accuracy, or by normative conformity to comply with social norms such as reciprocity. Disentangling the behavioural, cognitive, and neurobiological underpinnings of informational and normative conformity have proven elusive. Here, participants underwent fMRI while performing a perceptual task that involved both advice-taking and advice-giving to human and computer partners. The concurrent inclusion of 2 different social roles and 2 different social partners revealed distinct behavioural and neural markers for informational and normative conformity. Dorsal anterior cingulate cortex (dACC) BOLD response tracked informational conformity towards both human and computer but tracked normative conformity only when interacting with humans. A network of brain areas (dorsomedial prefrontal cortex (dmPFC) and temporoparietal junction (TPJ)) that tracked normative conformity increased their functional coupling with the dACC when interacting with humans. These findings enable differentiating the neural mechanisms by which different types of conformity shape social changes of mind.

Medial Frontal Cortex Activity Predicts Information Sampling in Economic Choice.

Decision-making not only requires agents to decide what to choose but also how much information to sample before committing to a choice. Previously established frameworks for economic choice argue for a deliberative process of evidence accumulation across time. These tacitly acknowledge a role of information sampling in that decisions are only made once sufficient evidence is acquired, yet few experiments have explicitly placed information sampling under the participant's control. Here, we use fMRI to investigate the neural basis of information sampling in economic choice by allowing participants (n = 30, sex not recorded) to actively sample information in a multistep decision task. We show that medial frontal cortex (MFC) activity is predictive of further information sampling before choice. Choice difficulty (inverse value difference, keeping sensory difficulty constant) was also encoded in MFC, but this effect was explained away by the inclusion of information sampling as a coregressor in the general linear model. A distributed network of regions across the prefrontal cortex encoded key features of the sampled information at the time it was presented. We propose that MFC is an important controller of the extent to which information is gathered before committing to an economic choice. This role may explain why MFC activity has been associated with evidence accumulation in previous studies in which information sampling was an implicit rather than explicit feature of the decision.SIGNIFICANCE STATEMENT The decisions we make are determined by the information we have sampled before committing to a choice. Accumulator frameworks of decision-making tacitly acknowledge the need to sample further information during the evidence accumulation process until a decision boundary is reached. However, relatively few studies explicitly place this decision to sample further information under the participant's control. In this fMRI study, we find that MFC activity is related to information sampling decisions in a multistep economic choice task. This suggests that an important role of evidence representations within MFC may be to guide adaptive sequential decisions to sample further information before committing to a final decision.

Using distance on the Riemannian manifold to compare representations in brain and in models.

Representational similarity analysis (RSA) summarizes activity patterns for a set of experimental conditions into a matrix composed of pairwise comparisons between activity patterns. Two examples of such matrices are the condition-by-condition inner product and correlation matrix. These representational matrices reside on the manifold of positive semidefinite matrices, called the Riemannian manifold. We hypothesize that representational similarities would be more accurately quantified by considering the underlying manifold of the representational matrices. Thus, we introduce the distance on the Riemannian manifold as a metric for comparing representations. Analyzing simulated and real fMRI data and considering a wide range of metrics, we show that the Riemannian distance is least susceptible to sampling bias, results in larger intra-subject reliability, and affords searchlight mapping with high sensitivity and specificity. Furthermore, we show that the Riemannian distance can be used for measuring multi-dimensional connectivity. This measure captures both univariate and multivariate connectivity and is also more sensitive to nonlinear regional interactions compared to the state-of-the-art measures. Applying our proposed metric to neural network representations of natural images, we demonstrate that it also possesses outstanding performance in quantifying similarity in models. Taken together, our results lend credence to the proposition that RSA should consider the manifold of the representational matrices to summarize response patterns in the brain and in models.

Distinct fronto-temporal substrates of distributional and taxonomic similarity among words: evidence from RSA of BOLD signals.

A class of semantic theories defines concepts in terms of statistical distributions of lexical items, basing meaning on vectors of word co-occurrence frequencies. A different approach emphasizes abstract hierarchical taxonomic relationships among concepts. However, the functional relevance of these different accounts and how they capture information-encoding of lexical meaning in the brain still remains elusive. We investigated to what extent distributional and taxonomic models explained word-elicited neural responses using cross-validated representational similarity analysis (RSA) of functional magnetic resonance imaging (fMRI) and model comparisons. Our findings show that the brain encodes both types of semantic information, but in distinct cortical regions. Posterior middle temporal regions reflected lexical-semantic similarity based on hierarchical taxonomies, in coherence with the action-relatedness of specific semantic word categories. In contrast, distributional semantics best predicted the representational patterns in left inferior frontal gyrus (LIFG, BA 47). Both representations coexisted in the angular gyrus supporting semantic binding and integration. These results reveal that neuronal networks with distinct cortical distributions across higher-order association cortex encode different representational properties of word meanings. Taxonomy may shape long-term lexical-semantic representations in memory consistently with the sensorimotor details of semantic categories, whilst distributional knowledge in the LIFG (BA 47) may enable semantic combinatorics in the context of language use. Our approach helps to elucidate the nature of semantic representations essential for understanding human language.

Comparing continual task learning in minds and machines.

Humans can learn to perform multiple tasks in succession over the lifespan ("continual" learning), whereas current machine learning systems fail. Here, we investigated the cognitive mechanisms that permit successful continual learning in humans and harnessed our behavioral findings for neural network design. Humans categorized naturalistic images of trees according to one of two orthogonal task rules that were learned by trial and error. Training regimes that focused on individual rules for prolonged periods (blocked training) improved human performance on a later test involving randomly interleaved rules, compared with control regimes that trained in an interleaved fashion. Analysis of human error patterns suggested that blocked training encouraged humans to form "factorized" representation that optimally segregated the tasks, especially for those individuals with a strong prior bias to represent the stimulus space in a well-structured way. By contrast, standard supervised deep neural networks trained on the same tasks suffered catastrophic forgetting under blocked training, due to representational interference in the deeper layers. However, augmenting deep networks with an unsupervised generative model that allowed it to first learn a good embedding of the stimulus space (similar to that observed in humans) reduced catastrophic forgetting under blocked training. Building artificial agents that first learn a model of the world may be one promising route to solving continual task performance in artificial intelligence research.

Multi-dimensional connectivity: a conceptual and mathematical review.

The estimation of functional connectivity between regions of the brain, for example based on statistical dependencies between the time series of activity in each region, has become increasingly important in neuroimaging. Typically, multiple time series (e.g. from each voxel in fMRI data) are first reduced to a single time series that summarises the activity in a region of interest, e.g. by averaging across voxels or by taking the first principal component; an approach we call one-dimensional connectivity. However, this summary approach ignores potential multi-dimensional connectivity between two regions, and a number of recent methods have been proposed to capture such complex dependencies. Here we review the most common multi-dimensional connectivity methods, from an intuitive perspective, from a formal (mathematical) point of view, and through a number of simulated and real (fMRI and MEG) data examples that illustrate the strengths and weaknesses of each method. The paper is accompanied with both functions and scripts, which implement each method and reproduce all the examples.

Representational Similarity Mapping of Distributional Semantics in Left Inferior Frontal, Middle Temporal, and Motor Cortex.

Language comprehension engages a distributed network of frontotemporal, parietal, and sensorimotor regions, but it is still unclear how meaning of words and their semantic relationships are represented and processed within these regions and to which degrees lexico-semantic representations differ between regions and semantic types. We used fMRI and representational similarity analysis to relate word-elicited multivoxel patterns to semantic similarity between action and object words. In left inferior frontal (BA 44-45-47), left posterior middle temporal and left precentral cortex, the similarity of brain response patterns reflected semantic similarity among action-related verbs, as well as across lexical classes-between action verbs and tool-related nouns and, to a degree, between action verbs and food nouns, but not between action verbs and animal nouns. Instead, posterior inferior temporal cortex exhibited a reverse response pattern, which reflected the semantic similarity among object-related nouns, but not action-related words. These results show that semantic similarity is encoded by a range of cortical areas, including multimodal association (e.g., anterior inferior frontal, posterior middle temporal) and modality-preferential (premotor) cortex and that the representational geometries in these regions are partly dependent on semantic type, with semantic similarity among action-related words crossing lexical-semantic category boundaries.

Reliability of dissimilarity measures for multi-voxel pattern analysis.

Representational similarity analysis of activation patterns has become an increasingly important tool for studying brain representations. The dissimilarity between two patterns is commonly quantified by the correlation distance or the accuracy of a linear classifier. However, there are many different ways to measure pattern dissimilarity and little is known about their relative reliability. Here, we compare the reliability of three classes of dissimilarity measure: classification accuracy, Euclidean/Mahalanobis distance, and Pearson correlation distance. Using simulations and four real functional magnetic resonance imaging (fMRI) datasets, we demonstrate that continuous dissimilarity measures are substantially more reliable than the classification accuracy. The difference in reliability can be explained by two characteristics of classifiers: discretization and susceptibility of the discriminant function to shifts of the pattern ensemble between imaging runs. Reliability can be further improved through multivariate noise normalization for all measures. Finally, unlike conventional distance measures, crossvalidated distances provide unbiased estimates of pattern dissimilarity on a ratio scale, thus providing an interpretable zero point. Overall, our results indicate that the crossvalidated Mahalanobis distance is preferable to both the classification accuracy and the correlation distance for characterizing representational geometries.

A toolbox for representational similarity analysis.

Neuronal population codes are increasingly being investigated with multivariate pattern-information analyses. A key challenge is to use measured brain-activity patterns to test computational models of brain information processing. One approach to this problem is representational similarity analysis (RSA), which characterizes a representation in a brain or computational model by the distance matrix of the response patterns elicited by a set of stimuli. The representational distance matrix encapsulates what distinctions between stimuli are emphasized and what distinctions are de-emphasized in the representation. A model is tested by comparing the representational distance matrix it predicts to that of a measured brain region. RSA also enables us to compare representations between stages of processing within a given brain or model, between brain and behavioral data, and between individuals and species. Here, we introduce a Matlab toolbox for RSA. The toolbox supports an analysis approach that is simultaneously data- and hypothesis-driven. It is designed to help integrate a wide range of computational models into the analysis of multichannel brain-activity measurements as provided by modern functional imaging and neuronal recording techniques. Tools for visualization and inference enable the user to relate sets of models to sets of brain regions and to statistically test and compare the models using nonparametric inference methods. The toolbox supports searchlight-based RSA, to continuously map a measured brain volume in search of a neuronal population code with a specific geometry. Finally, we introduce the linear-discriminant t value as a measure of representational discriminability that bridges the gap between linear decoding analyses and RSA. In order to demonstrate the capabilities of the toolbox, we apply it to both simulated and real fMRI data. The key functions are equally applicable to other modalities of brain-activity measurement. The toolbox is freely available to the community under an open-source license agreement (http://www.mrc-cbu.cam.ac.uk/methods-and-resources/toolboxes/license/).

Causal role of a neural system for separating and selecting multidimensional social cognitive information.

People are multi-faceted, typically good at some things but bad at others, and a critical aspect of social judgement is the ability to focus on those traits relevant for the task at hand. However, it remains unknown how the brain supports such context-dependent social judgement. Here, we examine how people represent multidimensional individuals, and how the brain extracts relevant information and filters out irrelevant information when comparing individuals within a specific dimension. Using human fMRI, we identify distinct neural representations in dorsomedial prefrontal cortex (dmPFC) and anterior insula (AI) supporting separation and selection of information for context-dependent social judgement. Causal evaluation using non-invasive brain stimulation shows that AI disruption alters the impact of relevant information on social comparison, whereas dmPFC disruption only affects the impact of irrelevant information. This neural circuit is distinct from the one supporting integration across, as opposed to separation of, different features of a multidimensional cognitive space.

An information-theoretic quantification of the content of communication between brain regions.

Quantifying the amount, content and direction of communication between brain regions is key to understanding brain function. Traditional methods to analyze brain activity based on the Wiener-Granger causality principle quantify the overall information propagated by neural activity between simultaneously recorded brain regions, but do not reveal the information flow about specific features of interest (such as sensory stimuli). Here, we develop a new information theoretic measure termed Feature-specific Information Transfer (FIT), quantifying how much information about a specific feature flows between two regions. FIT merges the Wiener-Granger causality principle with information-content specificity. We first derive FIT and prove analytically its key properties. We then illustrate and test them with simulations of neural activity, demonstrating that FIT identifies, within the total information flowing between regions, the information that is transmitted about specific features. We then analyze three neural datasets obtained with different recording methods, magneto- and electro-encephalography, and spiking activity, to demonstrate the ability of FIT to uncover the content and direction of information flow between brain regions beyond what can be discerned with traditional anaytical methods. FIT can improve our understanding of how brain regions communicate by uncovering previously hidden feature-specific information flow.

Caveats and Nuances of Model-Based and Model-Free Representational Connectivity Analysis.

Brain connectivity analyses have conventionally relied on statistical relationship between one-dimensional summaries of activation in different brain areas. However, summarizing activation patterns within each area to a single dimension ignores the potential statistical dependencies between their multi-dimensional activity patterns. Representational Connectivity Analyses (RCA) is a method that quantifies the relationship between multi-dimensional patterns of activity without reducing the dimensionality of the data. We consider two variants of RCA. In model-free RCA, the goal is to quantify the shared information for two brain regions. In model-based RCA, one tests whether two regions have shared information about a specific aspect of the stimuli/task, as defined by a model. However, this is a new approach and the potential caveats of model-free and model-based RCA are still understudied. We first explain how model-based RCA detects connectivity through the lens of models, and then present three scenarios where model-based and model-free RCA give discrepant results. These conflicting results complicate the interpretation of functional connectivity. We highlight the challenges in three scenarios: complex intermediate models, common patterns across regions, and transformation of representational structure across brain regions. The article is accompanied by scripts (https://osf.io/3nxfa/) that reproduce the results. In each case, we suggest potential ways to mitigate the difficulties caused by inconsistent results. The results of this study shed light on some understudied aspects of RCA, and allow researchers to use the method more effectively.

Distinct neural representations for prosocial and self-benefiting effort.

Prosocial behaviors-actions that benefit others-are central to individual and societal well-being. Although the mechanisms underlying the financial and moral costs of prosocial behaviors are increasingly understood, this work has often ignored a key influence on behavior: effort. Many prosocial acts are effortful, and people are averse to the costs of exerting them. However, how the brain encodes effort costs when actions benefit others is unknown. During fMRI, participants completed a decision-making task where they chose in each trial whether to "work" and exert force (30%-70% of maximum grip strength) or "rest" (no effort) for rewards (2-10 credits). Crucially, on separate trials, they made these decisions either to benefit another person or themselves. We used a combination of multivariate representational similarity analysis and model-based univariate analysis to reveal how the costs of prosocial and self-benefiting efforts are processed. Strikingly, we identified a unique neural signature of effort in the anterior cingulate gyrus (ACCg) for prosocial acts, both when choosing to help others and when exerting force to benefit them. This pattern was absent for self-benefiting behaviors. Moreover, stronger, specific representations of prosocial effort in the ACCg were linked to higher levels of empathy and higher subsequent exerted force to benefit others. In contrast, the ventral tegmental area and ventral insula represented value preferentially when choosing for oneself and not for prosocial acts. These findings advance our understanding of the neural mechanisms of prosocial behavior, highlighting the critical role that effort has in the brain circuits that guide helping others.

Correction: Inferring exemplar discriminability in brain representations.

[This corrects the article DOI: 10.1371/journal.pone.0232551.].

Entorhinal and ventromedial prefrontal cortices abstract and generalize the structure of reinforcement learning problems.

Knowledge of the structure of a problem, such as relationships between stimuli, enables rapid learning and flexible inference. Humans and other animals can abstract this structural knowledge and generalize it to solve new problems. For example, in spatial reasoning, shortest-path inferences are immediate in new environments. Spatial structural transfer is mediated by cells in entorhinal and (in humans) medial prefrontal cortices, which maintain their co-activation structure across different environments and behavioral states. Here, using fMRI, we show that entorhinal and ventromedial prefrontal cortex (vmPFC) representations perform a much broader role in generalizing the structure of problems. We introduce a task-remapping paradigm, where subjects solve multiple reinforcement learning (RL) problems differing in structural or sensory properties. We show that, as with space, entorhinal representations are preserved across different RL problems only if task structure is preserved. In vmPFC and ventral striatum, representations of prediction error also depend on task structure.

Allocentric representation in the human amygdala and ventral visual stream.

The hippocampus and the entorhinal cortex are considered the main brain structures for allocentric representation of the external environment. Here, we show that the amygdala and the ventral visual stream are involved in allocentric representation. Thirty-one young men explored 35 virtual environments during high-resolution functional magnetic resonance imaging (fMRI) of the medial temporal lobe (MTL) and were subsequently tested on recall of the allocentric pattern of the objects in each environment-in other words, the positions of the objects relative to each other and to the outer perimeter. We find increasingly unique brain activation patterns associated with increasing allocentric accuracy in distinct neural populations in the perirhinal cortex, parahippocampal cortex, fusiform cortex, amygdala, hippocampus, and entorhinal cortex. In contrast to the traditional view of a hierarchical MTL network with the hippocampus at the top, we demonstrate, using recently developed graph analyses, a hierarchical allocentric MTL network without a main connector hub.

Target detection by opponent coding in monkey prefrontal cortex.

The pFC plays a key role in flexible, context-specific decision making. One proposal [Machens, C. K., Romo, R., & Brody, C. D. Flexible control of mutual inhibition: A neural model of two-interval discrimination. Science, 307, 1121-1124, 2005] is that prefrontal cells may be dynamically organized into opponent coding circuits, with competitive groups of cells coding opposite behavioral decisions. Here, we show evidence for extensive, temporally evolving opponent organization in the monkey pFC during a cued target detection task. More than a half of all randomly selected cells discriminated stimulus category in this task. The largest set showed target-positive activity, with the strongest responses to the current target, intermediate activity for a nontarget that was a target on other trials, and lowest activity for nontargets never associated with the target category. Second most frequent was a reverse, antitarget pattern. In the ventrolateral frontal cortex, opponent organization was strongly established in phasic responses at stimulus onset; later, such activity was widely spread across dorsolateral and ventrolateral sites. Task-specific organization into opponent cell groups may be a general feature of prefrontal decision making.

Inferring exemplar discriminability in brain representations.

Representational distinctions within categories are important in all perceptual modalities and also in cognitive and motor representations. Recent pattern-information studies of brain activity have used condition-rich designs to sample the stimulus space more densely. To test whether brain response patterns discriminate among a set of stimuli (e.g. exemplars within a category) with good sensitivity, we can pool statistical evidence over all pairwise comparisons. Here we describe a wide range of statistical tests of exemplar discriminability and assess the validity (specificity) and power (sensitivity) of each test. The tests include previously used and novel, parametric and nonparametric tests, which treat subject as a random or fixed effect, and are based on different dissimilarity measures, different test statistics, and different inference procedures. We use simulated and real data to determine which tests are valid and which are most sensitive. A popular test statistic reflecting exemplar information is the exemplar discriminability index (EDI), which is defined as the average of the pattern dissimilarity estimates between different exemplars minus the average of the pattern dissimilarity estimates between repetitions of identical exemplars. The popular across-subject t test of the EDI (typically using correlation distance as the pattern dissimilarity measure) requires the assumption that the EDI is 0-mean normal under H0. Although this assumption is not strictly true, our simulations suggest that the test controls the false-positives rate at the nominal level, and is thus valid, in practice. However, test statistics based on average Mahalanobis distances or average linear-discriminant t values (both accounting for the multivariate error covariance among responses) are substantially more powerful for both random- and fixed-effects inference. Unlike average cross-validated distances, the EDI is sensitive to differences between the distributions associated with different exemplars (e.g. greater variability for some exemplars than for others), which complicates its interpretation. We suggest preferred procedures for safely and sensitively detecting subtle pattern differences between exemplars.

Private-public mappings in human prefrontal cortex.

A core feature of human cognition is an ability to separate private states of mind - what we think or believe - from public actions - what we say or do. This ability is central to successful social interaction - with different social contexts often requiring different mappings between private states and public actions in order to minimise conflict and facilitate communication. Here we investigated how the human brain supports private-public mappings, using an interactive task which required subjects to adapt how they communicated their confidence about a perceptual decision to the social context. Univariate and multivariate analysis of fMRI data revealed that a private-public distinction is reflected in a medial-lateral division of prefrontal cortex - with lateral frontal pole (FPl) supporting the context-dependent mapping from a private sense of confidence to a public report. The concept of private-public mappings provides a promising framework for understanding flexible social behaviour.