Inferring relevance in a changing world

A Comparison of Rapid Rule-Learning Strategies in Humans and Monkeys

Interspecies comparisons are key to deriving an understanding of the behavioral and neural correlates of human cognition from animal models. We perform a detailed comparison of the strategies of female macaque monkeys to male and female humans on a variant of the Wisconsin Card Sorting Test (WCST), a widely studied and applied task that provides a multiattribute measure of cognitive function and depends on the frontal lobe. WCST performance requires the inference of a rule change given ambiguous feedback. We found that well-trained monkeys infer new rules three times more slowly than minimally instructed humans. Input-dependent hidden Markov model-generalized linear models were fit to their choices, revealing hidden states akin to feature-based attention in both species. Decision processes resembled a win-stay, lose-shift strategy with interspecies similarities as well as key differences. Monkeys and humans both test multiple rule hypotheses over a series of rule-search trials and perform inference-like computations to exclude candidate choice options. We quantitatively show that perseveration, random exploration, and poor sensitivity to negative feedback account for the slower task-switching performance in monkeys.

Naturalizing relevance realization: why agency and cognition are fundamentally not computational

The way organismic agents come to know the world, and the way algorithms solve problems, are fundamentally different. The most sensible course of action for an organism does not simply follow from logical rules of inference. Before it can even use such rules, the organism must tackle the problem of relevance. It must turn ill-defined problems into well-defined ones, turn semantics into syntax. This ability to realize relevance is present in all organisms, from bacteria to humans. It lies at the root of organismic agency, cognition, and consciousness, arising from the particular autopoietic, anticipatory, and adaptive organization of living beings. In this article, we show that the process of relevance realization is beyond formalization. It cannot be captured completely by algorithmic approaches. This implies that organismic agency (and hence cognition as well as consciousness) are at heart not computational in nature. Instead, we show how the process of relevance is realized by an adaptive and emergent triadic dialectic (a trialectic), which manifests as a metabolic and ecological-evolutionary co-constructive dynamic. This results in a meliorative process that enables an agent to continuously keep a grip on its arena, its reality. To be alive means to make sense of one's world. This kind of embodied ecological rationality is a fundamental aspect of life, and a key characteristic that sets it apart from non-living matter.

A goal-centric outlook on learning

Goals play a central role in human cognition. However, computational theories of learning and decision-making often take goals as given. Here, we review key empirical findings showing that goals shape the representations of inputs, responses, and outcomes, such that setting a goal crucially influences the central aspects of any learning process: states, actions, and rewards. We thus argue that studying goal selection is essential to advance our understanding of learning. By following existing literature in framing goal selection within a hierarchy of decision-making problems, we synthesize important findings on the principles underlying goal value attribution and exploration strategies. Ultimately, we propose that a goal-centric perspective will help develop more complete accounts of learning in both biological and artificial agents.

Competing neural representations of choice shape evidence accumulation in humans

Making adaptive choices in dynamic environments requires flexible decision policies. Previously, we showed how shifts in outcome contingency change the evidence accumulation process that determines decision policies. Using in silico experiments to generate predictions, here we show how the cortico-basal ganglia-thalamic (CBGT) circuits can feasibly implement shifts in decision policies. When action contingencies change, dopaminergic plasticity redirects the balance of power, both within and between action representations, to divert the flow of evidence from one option to another. When competition between action representations is highest, the rate of evidence accumulation is the lowest. This prediction was validated in in vivo experiments on human participants, using fMRI, which showed that (1) evoked hemodynamic responses can reliably predict trial-wise choices and (2) competition between action representations, measured using a classifier model, tracked with changes in the rate of evidence accumulation. These results paint a holistic picture of how CBGT circuits manage and adapt the evidence accumulation process in mammals.

Lowered inter-stimulus discriminability hurts incremental contributions to learning

How does the similarity between stimuli affect our ability to learn appropriate response associations for them? In typical laboratory experiments learning is investigated under somewhat ideal circumstances, where stimuli are easily discriminable. This is not representative of most real-life learning, where overlapping "stimuli" can result in different "rewards" and may be learned simultaneously (e.g., you may learn over repeated interactions that a specific dog is friendly, but that a very similar looking one isn't). With two experiments, we test how humans learn in three stimulus conditions: one "best case" condition in which stimuli have idealized and highly discriminable visual and semantic representations, and two in which stimuli have overlapping representations, making them less discriminable. We find that, unsurprisingly, decreasing stimuli discriminability decreases performance. We develop computational models to test different hypotheses about how reinforcement learning (RL) and working memory (WM) processes are affected by different stimulus conditions. Our results replicate earlier studies demonstrating the importance of both processes to capture behavior. However, our results extend previous studies by demonstrating that RL, and not WM, is affected by stimulus distinctness: people learn slower and have higher across-stimulus value confusion at decision when stimuli are more similar to each other. These results illustrate strong effects of stimulus type on learning and demonstrate the importance of considering parallel contributions of different cognitive processes when studying behavior.

M1-selective muscarinic allosteric modulation enhances cognitive flexibility and effective salience in nonhuman primates

Acetylcholine (ACh) in cortical neural circuits mediates how selective attention is sustained in the presence of distractors and how flexible cognition adjusts to changing task demands. The cognitive domains of attention and cognitive flexibility might be differentially supported by the M1 muscarinic acetylcholine receptor (mAChR) subtype. Understanding how M1 mAChR mechanisms support these cognitive subdomains is of highest importance for advancing novel drug treatments for conditions with altered attention and reduced cognitive control including Alzheimer's disease or schizophrenia. Here, we tested this question by assessing how the subtype-selective M1 mAChR positive allosteric modulator (PAM) VU0453595 affects visual search and flexible reward learning in nonhuman primates. We found that allosteric potentiation of M1 mAChRs enhanced flexible learning performance by improving extradimensional set shifting, reducing latent inhibition from previously experienced distractors and reducing response perseveration in the absence of adverse side effects. These procognitive effects occurred in the absence of apparent changes of attentional performance during visual search. In contrast, nonselective ACh modulation using the acetylcholinesterase inhibitor (AChEI) donepezil improved attention during visual search at doses that did not alter cognitive flexibility and that already triggered gastrointestinal cholinergic side effects. These findings illustrate that M1 mAChR positive allosteric modulation enhances cognitive flexibility without affecting attentional filtering of distraction, consistent with M1 activity boosting the effective salience of relevant over irrelevant objects specifically during learning. These results suggest that M1 PAMs are versatile compounds for enhancing cognitive flexibility in disorders spanning schizophrenia and Alzheimer's diseases.

Visuospatial information foraging describes search behavior in learning latent environmental features

In the real world, making sequences of decisions to achieve goals often depends upon the ability to learn aspects of the environment that are not directly perceptible. Learning these so-called latent features requires seeking information about them. Prior efforts to study latent feature learning often used single decisions, used few features, and failed to distinguish between reward-seeking and information-seeking. To overcome this, we designed a task in which humans and monkeys made a series of choices to search for shapes hidden on a grid. On our task, the effects of reward and information outcomes from uncovering parts of shapes could be disentangled. Members of both species adeptly learned the shapes and preferred to select tiles expected to be informative earlier in trials than previously rewarding ones, searching a part of the grid until their outcomes dropped below the average information outcome-a pattern consistent with foraging behavior. In addition, how quickly humans learned the shapes was predicted by how well their choice sequences matched the foraging pattern, revealing an unexpected connection between foraging and learning. This adaptive search for information may underlie the ability in humans and monkeys to learn latent features to support goal-directed behavior in the long run.

Comparing rapid rule-learning strategies in humans and monkeys

Inter-species comparisons are key to deriving an understanding of the behavioral and neural correlates of human cognition from animal models. We perform a detailed comparison of macaque monkey and human strategies on an analogue of the Wisconsin Card Sort Test, a widely studied and applied multi-attribute measure of cognitive function, wherein performance requires the inference of a changing rule given ambiguous feedback. We found that well-trained monkeys rapidly infer rules but are three times slower than humans. Model fits to their choices revealed hidden states akin to feature-based attention in both species, and decision processes that resembled a Win-stay lose-shift strategy with key differences. Monkeys and humans test multiple rule hypotheses over a series of rule-search trials and perform inference-like computations to exclude candidates. An attention-set based learning stage categorization revealed that perseveration, random exploration and poor sensitivity to negative feedback explain the under-performance in monkeys.

Humans combine value learning and hypothesis testing strategically in multi-dimensional probabilistic reward learning

Realistic and complex decision tasks often allow for many possible solutions. How do we find the correct one? Introspection suggests a process of trying out solutions one after the other until success. However, such methodical serial testing may be too slow, especially in environments with noisy feedback. Alternatively, the underlying learning process may involve implicit reinforcement learning that learns about many possibilities in parallel. Here we designed a multi-dimensional probabilistic active-learning task tailored to study how people learn to solve such complex problems. Participants configured three-dimensional stimuli by selecting features for each dimension and received probabilistic reward feedback. We manipulated task complexity by changing how many feature dimensions were relevant to maximizing reward, as well as whether this information was provided to the participants. To investigate how participants learn the task, we examined models of serial hypothesis testing, feature-based reinforcement learning, and combinations of the two strategies. Model comparison revealed evidence for hypothesis testing that relies on reinforcement-learning when selecting what hypothesis to test. The extent to which participants engaged in hypothesis testing depended on the instructed task complexity: people tended to serially test hypotheses when instructed that there were fewer relevant dimensions, and relied more on gradual and parallel learning of feature values when the task was more complex. This demonstrates a strategic use of task information to balance the costs and benefits of the two methods of learning.

Updating Contextual Sensory Expectations for Adaptive Behavior

The brain has the extraordinary capacity to construct predictive models of the environment by internalizing statistical regularities in the sensory inputs. The resulting sensory expectations shape how we perceive and react to the world; at the neural level, this relates to decreased neural responses to expected than unexpected stimuli ("expectation suppression"). Crucially, expectations may need revision as context changes. However, existing research has often neglected this issue. Further, it is unclear whether contextual revisions apply selectively to expectations relevant to the task at hand, hence serving adaptive behavior. The present fMRI study examined how contextual visual expectations spread throughout the cortical hierarchy as we update our beliefs. We created a volatile environment: two alternating contexts contained different sequences of object images, thereby producing context-dependent expectations that needed revision when the context changed. Human participants of both sexes attended a training session before scanning to learn the contextual sequences. The fMRI experiment then tested for the emergence of contextual expectation suppression in two separate tasks, respectively, with task-relevant and task-irrelevant expectations. Effects of contextual expectation emerged progressively across the cortical hierarchy as participants attuned themselves to the context: expectation suppression appeared first in the insula, inferior frontal gyrus, and posterior parietal cortex, followed by the ventral visual stream, up to early visual cortex. This applied selectively to task-relevant expectations. Together, the present results suggest that an insular and frontoparietal executive control network may guide the flexible deployment of contextual sensory expectations for adaptive behavior in our complex and dynamic world.SIGNIFICANCE STATEMENT The world is structured by statistical regularities, which we use to predict the future. This is often accompanied by suppressed neural responses to expected compared with unexpected events ("expectation suppression"). Crucially, the world is also highly volatile and context-dependent: expected events may become unexpected when the context changes, thus raising the crucial need for belief updating. However, this issue has generally been neglected. By setting up a volatile environment, we show that expectation suppression emerges first in executive control regions, followed by relevant sensory areas, only when observers use their expectations to optimize behavior. This provides surprising yet clear evidence on how the brain controls the updating of sensory expectations for adaptive behavior in our ever-changing world.

Working memory capacity estimates moderate value learning for outcome-irrelevant features

To establish accurate action-outcome associations in the environment, individuals must refrain from assigning value to outcome-irrelevant features. However, studies have largely ignored the role of attentional control processes on action value updating. In the current study, we examined the extent to which working memory-a system that can filter and block the processing of irrelevant information in one's mind-also filters outcome-irrelevant information during value-based learning. For this aim, 174 individuals completed a well-established working memory capacity measurement and a reinforcement learning task designed to estimate outcome-irrelevant learning. We replicated previous studies showing a group-level tendency to assign value to tasks' response keys, despite clear instructions and practice suggesting they are irrelevant to the prediction of monetary outcomes. Importantly, individuals with higher working memory capacity were less likely to assign value to the outcome-irrelevant response keys, thus suggesting a significant moderation effect of working memory capacity on outcome-irrelevant learning. We discuss the role of working memory processing on value-based learning through the lens of a cognitive control failure.

Neurocognitive Artificial Neural Network Models Are Superior to Linear Models at Accounting for Dimensional Psychopathology

In recent years, there has been debate about the optimal conceptualisation of psychopathology. Structural models of psychopathology have been developed to counter issues, including comorbidity and poor diagnostic stability prevalent within the traditional nosological approach. Regardless of the conceptualisation of psychological dysfunction, deficits in neurocognitive abilities have been claimed to be an aetiological feature of psychopathology. Explorations of the association between neurocognition and psychopathology have typically taken a linear approach, overlooking the potential interactive dynamics of neurocognitive abilities. Previously, we proposed a multidimensional hypothesis, where within-person interactions between neurocognitive domains are fundamental to understanding the role of neurocognition within psychopathology. In this study, we used previously collected psychopathology data for 400 participants on psychopathological symptoms, substance use, and performance on eight neurocognitive tasks and compared the predictive accuracy of linear models to artificial neural network models. The artificial neural network models were significantly more accurate than the traditional linear models at predicting actual (a) lower-level and (b) high-level dimensional psychopathology. These results provide support for the multidimensional hypothesis: that the study of non-linear interactions and compensatory neurocognitive profiles are integral to understanding the functional associations between neurocognition and of psychopathology.

Rumination Derails Reinforcement Learning with Possible Implications for Ineffective Behavior

How does rumination affect reinforcement learning-the ubiquitous process by which we adjust behavior after error in order to behave more effectively in the future? In a within-subject design (n=49), we tested whether experimentally manipulated rumination disrupts reinforcement learning in a multidimensional learning task previously shown to rely on selective attention. Rumination impaired performance, yet unexpectedly this impairment could not be attributed to decreased attentional breadth (quantified using a "decay" parameter in a computational model). Instead, trait rumination (between subjects) was associated with higher decay rates (implying narrower attention), yet not with impaired performance. Our task-performance results accord with the possibility that state rumination promotes stress-generating behavior in part by disrupting reinforcement learning. The trait-rumination finding accords with the predictions of a prominent model of trait rumination (the attentional-scope model). More work is needed to understand the specific mechanisms by which state rumination disrupts reinforcement learning.

Decision uncertainty during hypothesis testing enhances memory accuracy for incidental information

Humans actively seek information to reduce uncertainty, providing insight on how our decisions causally affect the world. While we know that episodic memories can help support future goal-oriented behaviors, little is known about how hypothesis testing during exploration influences episodic memory. To investigate this question, we designed a hypothesis testing paradigm, in which participants figured out rules to unlock treasure chests. Using this paradigm, we characterized how hypothesis testing during exploration influenced memory for the contents of the treasure chests. We found that there was an inverted U-shaped relationship between decision uncertainty and memory, such that memory was best when decision uncertainty was moderate. An exploratory analysis also showed that surprising outcomes lead to lower memory confidence independent of accuracy. These findings support a model in which moderate decision uncertainty during hypothesis testing enhances incidental information encoding.

What Accounts for the Factors of Psychopathology? An Investigation of the Neurocognitive Correlates of Internalising, Externalising, and the p-Factor

Neurocognitive deficits have been consistently associated with a wide range of psychopathology and are proposed to not only be a consequence of the development of psychopathology but also directly involved in its aetiology. However, there is no clear understanding of what neurocognitive processes are particularly important to mental health. In this paper, we explored the association between neurocognitive abilities and the factors derived from structural models of psychopathology. Four hundred participants from a representative community sample completed measures of symptomology and substance use, as well as 8 neurocognitive tasks. We found a correlated-factors model, with internalising and externalising as the higher-order factors, and a single-factor model with only the p-factor, to be good fits for the data. Tasks that measured the speed of processing were significantly associated with internalising, externalising, and the p-factor, and accounted for significant amounts of unique variance in the factors after accounting for the common variance of the other tasks. Tasks that measured working memory, shifting, and inhibition were not significantly associated with psychopathology factors. Our findings suggest that neurocognitive abilities may not be differentially associated with psychopathology factors, but that speed of processing is a common correlate of the factors. We emphasise the importance of examining neurocognitive abilities and psychopathology on the individual level.

Monkey plays Pac-Man with compositional strategies and hierarchical decision-making

Humans can often handle daunting tasks with ease by developing a set of strategies to reduce decision-making into simpler problems. The ability to use heuristic strategies demands an advanced level of intelligence and has not been demonstrated in animals. Here, we trained macaque monkeys to play the classic video game Pac-Man. The monkeys’ decision-making may be described with a strategy-based hierarchical decision-making model with over 90% accuracy. The model reveals that the monkeys adopted the take-the-best heuristic by using one dominating strategy for their decision-making at a time and formed compound strategies by assembling the basis strategies to handle particular game situations. With the model, the computationally complex but fully quantifiable Pac-Man behavior paradigm provides a new approach to understanding animals’ advanced cognition.

Selective particle attention: Rapidly and flexibly selecting features for deep reinforcement learning

Deep Reinforcement Learning (RL) is often criticised for being data inefficient and inflexible to changes in task structure. Part of the reason for these issues is that Deep RL typically learns end-to-end using backpropagation, which results in task-specific representations. One approach for circumventing these problems is to apply Deep RL to existing representations that have been learned in a more task-agnostic fashion. However, this only partially solves the problem as the Deep RL algorithm learns a function of all pre-existing representations and is therefore still susceptible to data inefficiency and a lack of flexibility. Biological agents appear to solve this problem by forming internal representations over many tasks and only selecting a subset of these features for decision-making based on the task at hand; a process commonly referred to as selective attention. We take inspiration from selective attention in biological agents and propose a novel algorithm called Selective Particle Attention (SPA), which selects subsets of existing representations for Deep RL. Crucially, these subsets are not learned through backpropagation, which is slow and prone to overfitting, but instead via a particle filter that rapidly and flexibly identifies key subsets of features using only reward feedback. We evaluate SPA on two tasks that involve raw pixel input and dynamic changes to the task structure, and show that it greatly increases the efficiency and flexibility of downstream Deep RL algorithms.

Latent motives guide structure learning during adaptive social choice

Predicting the behaviour of others is an essential part of social cognition. Despite its ubiquity, social prediction poses a poorly understood generalization problem: we cannot assume that others will repeat past behaviour in new settings or that their future actions are entirely unrelated to the past. We demonstrate that humans solve this challenge using a structure learning mechanism that uncovers other people's latent, unobservable motives, such as greed and risk aversion. In four studies, participants (N = 501) predicted other players' decisions across four economic games, each with different social tensions (for example, Prisoner's Dilemma and Stag Hunt). Participants achieved accurate social prediction by learning the stable motivational structure underlying a player's changing actions across games. This motive-based abstraction enabled participants to attend to information diagnostic of the player's next move and disregard irrelevant contextual cues. Participants who successfully learned another's motives were more strategic in a subsequent competitive interaction with that player in entirely new contexts, reflecting that social structure learning supports adaptive social behaviour.

Advances in modeling learning and decision-making in neuroscience

An organism's survival depends on its ability to learn about its environment and to make adaptive decisions in the service of achieving the best possible outcomes in that environment. To study the neural circuits that support these functions, researchers have increasingly relied on models that formalize the computations required to carry them out. Here, we review the recent history of computational modeling of learning and decision-making, and how these models have been used to advance understanding of prefrontal cortex function. We discuss how such models have advanced from their origins in basic algorithms of updating and action selection to increasingly account for complexities in the cognitive processes required for learning and decision-making, and the representations over which they operate. We further discuss how a deeper understanding of the real-world complexities in these computations has shed light on the fundamental constraints on optimal behavior, and on the complex interactions between corticostriatal pathways to determine such behavior. The continuing and rapid development of these models holds great promise for understanding the mechanisms by which animals adapt to their environments, and what leads to maladaptive forms of learning and decision-making within clinical populations.

Dynamic decision policy reconfiguration under outcome uncertainty

In uncertain or unstable environments, sometimes the best decision is to change your mind. To shed light on this flexibility, we evaluated how the underlying decision policy adapts when the most rewarding action changes. Human participants performed a dynamic two-armed bandit task that manipulated the certainty in relative reward (conflict) and the reliability of action-outcomes (volatility). Continuous estimates of conflict and volatility contributed to shifts in exploratory states by changing both the rate of evidence accumulation (drift rate) and the amount of evidence needed to make a decision (boundary height), respectively. At the trialwise level, following a switch in the optimal choice, the drift rate plummets and the boundary height weakly spikes, leading to a slow exploratory state. We find that the drift rate drives most of this response, with an unreliable contribution of boundary height across experiments. Surprisingly, we find no evidence that pupillary responses associated with decision policy changes. We conclude that humans show a stereotypical shift in their decision policies in response to environmental changes.

How Working Memory and Reinforcement Learning Are Intertwined: A Cognitive, Neural, and Computational Perspective

Reinforcement learning and working memory are two core processes of human cognition and are often considered cognitively, neuroscientifically, and algorithmically distinct. Here, we show that the brain networks that support them actually overlap significantly and that they are less distinct cognitive processes than often assumed. We review literature demonstrating the benefits of considering each process to explain properties of the other and highlight recent work investigating their more complex interactions. We discuss how future research in both computational and cognitive sciences can benefit from one another, suggesting that a key missing piece for artificial agents to learn to behave with more human-like efficiency is taking working memory's role in learning seriously. This review highlights the risks of neglecting the interplay between different processes when studying human behavior (in particular when considering individual differences). We emphasize the importance of investigating these dynamics to build a comprehensive understanding of human cognition.

Computational mechanisms of distributed value representations and mixed learning strategies

Learning appropriate representations of the reward environment is challenging in the real world where there are many options, each with multiple attributes or features. Despite existence of alternative solutions for this challenge, neural mechanisms underlying emergence and adoption of value representations and learning strategies remain unknown. To address this, we measure learning and choice during a multi-dimensional probabilistic learning task in humans and trained recurrent neural networks (RNNs) to capture our experimental observations. We find that human participants estimate stimulus-outcome associations by learning and combining estimates of reward probabilities associated with the informative feature followed by those of informative conjunctions. Through analyzing representations, connectivity, and lesioning of the RNNs, we demonstrate this mixed learning strategy relies on a distributed neural code and opponency between excitatory and inhibitory neurons through value-dependent disinhibition. Together, our results suggest computational and neural mechanisms underlying emergence of complex learning strategies in naturalistic settings.

Learning at Variable Attentional Load Requires Cooperation of Working Memory, Meta-learning, and Attention-augmented Reinforcement Learning

Flexible learning of changing reward contingencies can be realized with different strategies. A fast learning strategy involves using working memory of recently rewarded objects to guide choices. A slower learning strategy uses prediction errors to gradually update value expectations to improve choices. How the fast and slow strategies work together in scenarios with real-world stimulus complexity is not well known. Here, we aim to disentangle their relative contributions in rhesus monkeys while they learned the relevance of object features at variable attentional load. We found that learning behavior across six monkeys is consistently best predicted with a model combining (i) fast working memory and (ii) slower reinforcement learning from differently weighted positive and negative prediction errors as well as (iii) selective suppression of nonchosen feature values and (iv) a meta-learning mechanism that enhances exploration rates based on a memory trace of recent errors. The optimal model parameter settings suggest that these mechanisms cooperate differently at low and high attentional loads. Whereas working memory was essential for efficient learning at lower attentional loads, enhanced weighting of negative prediction errors and meta-learning were essential for efficient learning at higher attentional loads. Together, these findings pinpoint a canonical set of learning mechanisms and suggest how they may cooperate when subjects flexibly adjust to environments with variable real-world attentional demands.

Cognitive maps of social features enable flexible inference in social networks

In order to navigate a complex web of relationships, an individual must learn and represent the connections between people in a social network. However, the sheer size and complexity of the social world makes it impossible to acquire firsthand knowledge of all relations within a network, suggesting that people must make inferences about unobserved relationships to fill in the gaps. Across three studies (n = 328), we show that people can encode information about social features (e.g., hobbies, clubs) and subsequently deploy this knowledge to infer the existence of unobserved friendships in the network. Using computational models, we test various feature-based mechanisms that could support such inferences. We find that people's ability to successfully generalize depends on two representational strategies: a simple but inflexible similarity heuristic that leverages homophily, and a complex but flexible cognitive map that encodes the statistical relationships between social features and friendships. Together, our studies reveal that people can build cognitive maps encoding arbitrary patterns of latent relations in many abstract feature spaces, allowing social networks to be represented in a flexible format. Moreover, these findings shed light on open questions across disciplines about how people learn and represent social networks and may have implications for generating more human-like link prediction in machine learning algorithms.

An empirical evaluation of active inference in multi-armed bandits

A key feature of sequential decision making under uncertainty is a need to balance between exploiting-choosing the best action according to the current knowledge, and exploring-obtaining information about values of other actions. The multi-armed bandit problem, a classical task that captures this trade-off, served as a vehicle in machine learning for developing bandit algorithms that proved to be useful in numerous industrial applications. The active inference framework, an approach to sequential decision making recently developed in neuroscience for understanding human and animal behaviour, is distinguished by its sophisticated strategy for resolving the exploration-exploitation trade-off. This makes active inference an exciting alternative to already established bandit algorithms. Here we derive an efficient and scalable approximate active inference algorithm and compare it to two state-of-the-art bandit algorithms: Bayesian upper confidence bound and optimistic Thompson sampling. This comparison is done on two types of bandit problems: a stationary and a dynamic switching bandit. Our empirical evaluation shows that the active inference algorithm does not produce efficient long-term behaviour in stationary bandits. However, in the more challenging switching bandit problem active inference performs substantially better than the two state-of-the-art bandit algorithms. The results open exciting venues for further research in theoretical and applied machine learning, as well as lend additional credibility to active inference as a general framework for studying human and animal behaviour.

Computational framework for investigating predictive processing in auditory perception

BACKGROUND

The brain tracks sound sources as they evolve in time, collecting contextual information to predict future sensory inputs. Previous work in predictive coding typically focuses on the perception of predictable stimuli, leaving the implementation of these same neural processes in more complex, real-world environments containing randomness and uncertainty up for debate.

NEW METHOD

To facilitate investigation into the perception of less tightly-controlled listening scenarios, we present a computational model as a tool to ask targeted questions about the underlying predictive processes that connect complex sensory inputs to listener behavior and neural responses. In the modeling framework, observed sound features (e.g. pitch) are tracked sequentially using Bayesian inference. Sufficient statistics are inferred from past observations at multiple time scales and used to make predictions about future observation while tracking the statistical structure of the sensory input.

RESULTS

Facets of the model are discussed in terms of their application to perceptual research, and examples taken from real-world audio demonstrate the model's flexibility to capture a variety of statistical structures along various perceptual dimensions.

COMPARISON WITH EXISTING METHODS

Previous models are often targeted toward interpreting a particular experimental paradigm (e.g., oddball paradigm), perceptual dimension (e.g., pitch processing), or task (e.g., speech segregation), thus limiting their ability to generalize to other domains. The presented model is designed as a flexible and practical tool for broad application.

CONCLUSION

The model is presented as a general framework for generating new hypotheses and guiding investigation into the neural processes underlying predictive coding of complex scenes.

A Study on Human Learning Ability during Classification of Motion and Colour Visual Cues and Their Combination

The paper presents a study on the human learning process during the classification of stimuli, defined by motion and color visual cues and their combination. Because the classification dimension and the features that define each category are uncertain, we model the learning curves using Bayesian inference and more precisely the Normalized Conjunctive Consensus rule, and also on the base of the more efficient probabilistic Proportional Conflict Redistribution rule No 5 (pPCR5) defined within Dezert-Smarandache Theory (DSmT) of plausible and paradoxical reasoning. Our goal is to study how these rules succeed to model consistently both: human individual and group behaviour during the learning of the associations between the stimuli and the responses in categorization tasks varying by the amount of relevant stimulus information. The effect of age on this process is also evaluated.

Human Representation Learning

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

Learning a reach trajectory based on binary reward feedback

Binary reward feedback on movement success is sufficient for learning some simple sensorimotor mappings in a reaching task, but not for some other tasks in which multiple kinematic factors contribute to performance. The critical condition for learning in more complex tasks remains unclear. Here, we investigate whether reward-based motor learning is possible in a multi-dimensional trajectory matching task and whether simplifying the task by providing feedback on one factor at a time ('factorized feedback') can improve learning. In two experiments, participants performed a trajectory matching task in which learning was measured as a reduction in the error. In Experiment 1, participants matched a straight trajectory slanted in depth. We factorized the task by providing feedback on the slant error, the length error, or on their composite. In Experiment 2, participants matched a curved trajectory, also slanted in depth. In this experiment, we factorized the feedback by providing feedback on the slant error, the curvature error, or on the integral difference between the matched and target trajectory. In Experiment 1, there was anecdotal evidence that participants learnt the multidimensional task. Factorization did not improve learning. In Experiment 2, there was anecdotal evidence the multidimensional task could not be learnt. We conclude that, within a complexity range, multiple kinematic factors can be learnt in parallel.

Computational evidence for hierarchically structured reinforcement learning in humans

Humans have the fascinating ability to achieve goals in a complex and constantly changing world, still surpassing modern machine-learning algorithms in terms of flexibility and learning speed. It is generally accepted that a crucial factor for this ability is the use of abstract, hierarchical representations, which employ structure in the environment to guide learning and decision making. Nevertheless, how we create and use these hierarchical representations is poorly understood. This study presents evidence that human behavior can be characterized as hierarchical reinforcement learning (RL). We designed an experiment to test specific predictions of hierarchical RL using a series of subtasks in the realm of context-based learning and observed several behavioral markers of hierarchical RL, such as asymmetric switch costs between changes in higher-level versus lower-level features, faster learning in higher-valued compared to lower-valued contexts, and preference for higher-valued compared to lower-valued contexts. We replicated these results across three independent samples. We simulated three models-a classic RL, a hierarchical RL, and a hierarchical Bayesian model-and compared their behavior to human results. While the flat RL model captured some aspects of participants' sensitivity to outcome values, and the hierarchical Bayesian model captured some markers of transfer, only hierarchical RL accounted for all patterns observed in human behavior. This work shows that hierarchical RL, a biologically inspired and computationally simple algorithm, can capture human behavior in complex, hierarchical environments and opens the avenue for future research in this field.

Learning arbitrary stimulus-reward associations for naturalistic stimuli involves transition from learning about features to learning about objects

Most cognitive processes are studied using abstract or synthetic stimuli with specific features to fully control what is presented to subjects. However, recent studies have revealed enhancements of cognitive capacities (such as working memory) when processing naturalistic versus abstract stimuli. Using abstract stimuli constructed from distinct visual features (e.g., color and shape), we have recently shown that human subjects can learn multidimensional stimulus-reward associations via initially estimating reward value of individual features (feature-based learning) before gradually switching to learning about reward value of individual stimuli (object-based learning). Here, we examined whether similar strategies are adopted during learning about naturalistic stimuli that are clearly perceived as objects (instead of a combination of features) and contain both task-relevant and irrelevant features. We found that similar to learning about abstract stimuli, subjects initially adopted feature-based learning more strongly before transitioning to object-based learning. However, there were three key differences between learning about naturalistic and abstract stimuli. First, compared with abstract stimuli, the initial learning strategy was less feature-based for naturalistic stimuli. Second, subjects transitioned to object-based learning faster for naturalistic stimuli. Third, unexpectedly, subjects were more likely to adopt feature-based learning for naturalistic stimuli, both at the steady state and overall. These results suggest that despite the stronger tendency to perceive naturalistic stimuli as objects, which leads to greater likelihood of using object-based learning as the initial strategy and a faster transition to object-based learning, the influence of individual features on learning is stronger for these stimuli such that ultimately the object-based strategy is adopted less. Overall, our findings suggest that feature-based learning is a general initial strategy for learning about reward value of all types of multi-dimensional stimuli.

Under construction: ventral and lateral frontal lobe contributions to value-based decision-making and learning

Even apparently simple choices, like selecting a dessert in a pastry shop, involve options characterized by multiple motivationally relevant attributes. Neuroeconomic research suggests that the human brain may track the subjective value of such options, allowing disparate reward-predictive information to be compared in a common currency. However, the brain mechanisms involved in identifying value-predictive features and combining these to assess the value of each decision option remain unclear. Here, we review recent evidence from studies of multi-attribute decision-making in people with focal frontal lobe damage and in healthy people undergoing functional magnetic resonance imaging. This work suggests that ventromedial and lateral prefrontal cortex and orbitofrontal cortex are important for forming value judgments under conditions of complexity. We discuss studies supporting the involvement of these regions in selecting among and evaluating option attributes during value judgment and decision-making and when learning from reward feedback. These findings are consistent with roles for these regions in guiding value construction. They argue for a more nuanced understanding of how ventral and lateral prefrontal cortex contribute to discovering and recognizing value, processes that are required under the complex conditions typical of many everyday decisions.

Doubting what you already know: Uncertainty regarding state transitions is associated with obsessive compulsive symptoms

Obsessive compulsive (OC) symptoms involve excessive information gathering (e.g., checking, reassurance-seeking), and uncertainty about possible, often catastrophic, future events. Here we propose that these phenomena are the result of excessive uncertainty regarding state transitions (transition uncertainty): a computational impairment in Bayesian inference leading to a reduced ability to use the past to predict the present and future, and to oversensitivity to feedback (i.e. prediction errors). Using a computational model of Bayesian learning under uncertainty in a reversal learning task, we investigate the relationship between OC symptoms and transition uncertainty. Individuals high and low in OC symptoms performed a task in which they had to detect shifts (i.e. transitions) in cue-outcome contingencies. Modeling subjects' choices was used to estimate each individual participant's transition uncertainty and associated responses to feedback. We examined both an optimal observer model and an approximate Bayesian model in which participants were assumed to attend (and learn about) only one of several cues on each trial. Results suggested the participants were more likely to distribute attention across cues, in accordance with the optimal observer model. As hypothesized, participants with higher OC symptoms exhibited increased transition uncertainty, as well as a pattern of behavior potentially indicative of a difficulty in relying on learned contingencies, with no evidence for perseverative behavior. Increased transition uncertainty compromised these individuals' ability to predict ensuing feedback, rendering them more surprised by expected outcomes. However, no evidence for excessive belief updating was found. These results highlight a potential computational basis for OC symptoms and obsessive compulsive disorder (OCD). The fact the OC symptoms predicted a decreased reliance on the past rather than perseveration challenges preconceptions of OCD as a disorder of inflexibility. Our results have implications for the understanding of the neurocognitive processes leading to excessive uncertainty and distrust of past experiences in OCD.

Spontaneous Task Structure Formation Results in a Cost to Incidental Memory of Task Stimuli

Humans are characterized by their ability to leverage rules for classifying and linking stimuli to context-appropriate actions. Previous studies have shown that when humans learn stimulus-response associations for two-dimensional stimuli, they implicitly form and generalize hierarchical rule structures (task-sets). However, the cognitive processes underlying structure formation are poorly understood. Across four experiments, we manipulated how trial-unique images mapped onto responses to bias spontaneous task-set formation and investigated structure learning through the lens of incidental stimulus encoding. Participants performed a learning task designed to either promote task-set formation (by “motor-clustering” possible stimulus-action rules), or to discourage it (by using arbitrary category-response mappings). We adjudicated between two hypotheses: Structure learning may promote attention to task stimuli, thus resulting in better subsequent memory. Alternatively, building task-sets might impose cognitive demands (for instance, on working memory) that divert attention away from stimulus encoding. While the clustering manipulation affected task-set formation, there were also substantial individual differences. Importantly, structure learning incurred a cost: spontaneous task-set formation was associated with diminished stimulus encoding. Thus, spontaneous hierarchical task-set formation appears to involve cognitive demands that divert attention away from encoding of task stimuli during structure learning.

Intact Reinforcement Learning But Impaired Attentional Control During Multidimensional Probabilistic Learning in Older Adults

To efficiently learn optimal behavior in complex environments, humans rely on an interplay of learning and attention. Healthy aging has been shown to independently affect both of these functions. Here, we investigate how reinforcement learning and selective attention interact during learning from trial and error across age groups. We acquired behavioral and fMRI data from older and younger adults (male and female) performing two probabilistic learning tasks with varying attention demands. Although learning in the unidimensional task did not differ across age groups, older adults performed worse than younger adults in the multidimensional task, which required high levels of selective attention. Computational modeling showed that choices of older adults are better predicted by reinforcement learning than Bayesian inference, and that older adults rely more on reinforcement learning-based predictions than younger adults. Conversely, a higher proportion of younger adults' choices was predicted by a computationally demanding Bayesian approach. In line with the behavioral findings, we observed no group differences in reinforcement-learning related fMRI activation. Specifically, prediction-error activation in the nucleus accumbens was similar across age groups, and numerically higher in older adults. However, activation in the default mode was less suppressed in older adults in for higher attentional task demands, and the level of suppression correlated with behavioral performance. Our results indicate that healthy aging does not significantly impair simple reinforcement learning. However, in complex environments, older adults rely more heavily on suboptimal reinforcement-learning strategies supported by the ventral striatum, whereas younger adults use attention processes supported by cortical networks.SIGNIFICANCE STATEMENT Changes in the way that healthy human aging affects how we learn to optimally behave are not well understood; it has been suggested that age-related declines in dopaminergic function may impair older adult's ability to learn from reinforcement. In the present fMRI experiment, we show that learning and nucleus accumbens activation in a simple unidimensional reinforcement-learning task was not significantly affected by age. However, in a more complex multidimensional task, older adults showed worse performance and relied more on reinforcement-learning strategies than younger adults, while failing to disengage their default-mode network during learning. These results imply that older adults are only impaired in reinforcement learning if they additionally need to learn which dimensions of the environment are currently important.

Cooperative update of beliefs and state-transition functions in human reinforcement learning

It is widely known that reinforcement learning systems in the brain contribute to learning via interactions with the environment. These systems are capable of solving multidimensional problems, in which some dimensions are relevant to a reward, while others are not. To solve these problems, computational models use Bayesian learning, a strategy supported by behavioral and neural evidence in human. Bayesian learning takes into account beliefs, which represent a learner’s confidence in a particular dimension being relevant to the reward. Beliefs are given as a posterior probability of the state-transition (reward) function that maps the optimal actions to the states in each dimension. However, when it comes to implementing this learning strategy, the order in which beliefs and state-transition functions update remains unclear. The present study investigates this update order using a trial-by-trial analysis of human behavior and electroencephalography signals during a task in which learners have to identify the reward-relevant dimension. Our behavioral and neural results reveal a cooperative update—within 300 ms after the outcome feedback, the state-transition functions are updated, followed by the beliefs for each dimension.

Ten simple rules for the computational modeling of behavioral data

Computational modeling of behavior has revolutionized psychology and neuroscience. By fitting models to experimental data we can probe the algorithms underlying behavior, find neural correlates of computational variables and better understand the effects of drugs, illness and interventions. But with great power comes great responsibility. Here, we offer ten simple rules to ensure that computational modeling is used with care and yields meaningful insights. In particular, we present a beginner-friendly, pragmatic and details-oriented introduction on how to relate models to data. What, exactly, can a model tell us about the mind? To answer this, we apply our rules to the simplest modeling techniques most accessible to beginning modelers and illustrate them with examples and code available online. However, most rules apply to more advanced techniques. Our hope is that by following our guidelines, researchers will avoid many pitfalls and unleash the power of computational modeling on their own data.

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.

Pupil dilation as an implicit measure of appetitive Pavlovian learning

Appetitive Pavlovian conditioning is a learning mechanism of fundamental biological and pathophysiological significance. Nonetheless, its exploration in humans remains sparse, which is partly attributed to the lack of an established psychophysiological parameter that aptly represents conditioned responding. This study evaluated pupil diameter and other ocular response measures (gaze dwelling time, blink duration and count) as indices of conditioning. Additionally, a learning model was used to infer participants' learning progress on the basis of their pupil dilation. Twenty-nine healthy volunteers completed an appetitive differential delay conditioning paradigm with a primary reward, while the ocular response measures along with other psychophysiological (heart rate, electrodermal activity, postauricular and eyeblink reflex) and behavioral (ratings, contingency awareness) parameters were obtained to examine the relation among different measures. A significantly stronger increase in pupil diameter, longer gaze duration and shorter eyeblink duration was observed in response to the reward-predicting cue compared to the control cue. The Pearce-Hall attention model best predicted the trial-by-trial pupil diameter. This conditioned response was corroborated by a pronounced heart rate deceleration to the reward-predicting cue, while no conditioning effect was observed in the electrodermal activity or startle responses. There was no discernible correlation between the psychophysiological response measures. These results highlight the potential value of ocular response measures as sensitive indices for representing appetitive conditioning.

Uncovering the 'state': Tracing the hidden state representations that structure learning and decision-making

We review the abstract concept of a 'state' - an internal representation posited by reinforcement learning theories to be used by an agent, whether animal, human or artificial, to summarize the features of the external and internal environment that are relevant for future behavior on a particular task. Armed with this summary representation, an agent can make decisions and perform actions to interact effectively with the world. Here, we review recent findings from the neurobiological and behavioral literature to ask: 'what is a state?' with respect to the internal representations that organize learning and decision making across a range of tasks. We find that state representations include information beyond a straightforward summary of the immediate cues in the environment, providing timing or contextual information from the recent or more distant past, which allows these additional factors to influence decision making and other goal-directed behaviors in complex and perhaps unexpected ways.

State representation in mental illness

Reinforcement learning theory provides a powerful set of computational ideas for modeling human learning and decision making. Reinforcement learning algorithms rely on state representations that enable efficient behavior by focusing only on aspects relevant to the task at hand. Forming such representations often requires selective attention to the sensory environment, and recalling memories of relevant past experiences. A striking range of psychiatric disorders, including bipolar disorder and schizophrenia, involve changes in these cognitive processes. We review and discuss evidence that these changes can be cast as altered state representation, with the goal of providing a useful transdiagnostic dimension along which mental disorders can be understood and compared.

Synaptic plasticity/dysplasticity, process memory and item memory in rodent models of mental dysfunction

Activity-dependent changes in the effective connection strength of synapses are a fundamental feature of a nervous system. This so-called synaptic plasticity is thought to underlie storage of information in memory and has been hypothesized to be crucial for the effects of cognitive behavioral therapy. Synaptic plasticity stores information in a neural network, creating a trace of neural activity from past experience. The plasticity can also change the behavior of the network so the network can differentially transform/compute information in future activations. We discuss these two related but separable functions of synaptic plasticity; one we call "item memory" as it represents and stores items of information in memory, the other we call "process memory" as it encodes and stores functions such as computations to modify network information processing capabilities. We review evidence of item and process memory operations in behavior and evidence that experience modifies the brain's functional networks. We discuss neurodevelopmental rodent models relevant for understanding mental illness and compare two models in which one model, neonatal ventral hippocampal lesion (NVHL) has beneficial adult outcomes after being exposed to an adolescent cognitive experience that is potentially similar to cognitive behavioral therapy. The other model, gestational day 17 methylazoxymethanol acetate (GD17-MAM), does not benefit from the same adolescent cognitive experience. We propose that process memory is altered by early cognitive experience in NVHL rats but not in GD17-MAM rats, and discuss how dysplasticity factors may contribute to the differential adult outcomes after early cognitive experience in the NVHL and MAM models.

Competitive Frontoparietal Interactions Mediate Implicit Inferences

Frequent experience with regularities in our environment allows us to use predictive information to guide our decision process. However, contingencies in our environment are not always explicitly present and sometimes need to be inferred. Heretofore, it remained unknown how predictive information guides decision-making when explicit knowledge is absent and how the brain shapes such implicit inferences. In the present experiment, 17 human participants (9 females) performed a discrimination task in which a target stimulus was preceded by a predictive cue. Critically, participants had no explicit knowledge that some of the cues signaled an upcoming target, allowing us to investigate how implicit inferences emerge and guide decision-making. Despite unawareness of the cue-target contingencies, participants were able to use implicit information to improve performance. Concurrent EEG recordings demonstrate that implicit inferences rely upon interactions between internally and externally oriented networks, whereby prefrontal regions inhibit parietal cortex under internal implicit control.SIGNIFICANCE STATEMENT Regularities in our environment can guide our behavior providing information about upcoming events. Interestingly, such predictive information does not need to be explicitly represented to effectively guide our decision process. Here, we show how the brain engages in such real-world "data mining" and how implicit inferences emerge. We used a contingency cueing task and demonstrated that implicit inferences influenced responses to subsequent targets despite a lack of awareness of cue-target contingencies. Further, we show that these implicit inferences emerge through interactions between internally and externally oriented neural networks. The current results highlight the importance of prefrontal processes in transforming external events into predictive internalized models of the world.

Holistic Reinforcement Learning: The Role of Structure and Attention

Compact representations of the environment allow humans to behave efficiently in a complex world. Reinforcement learning models capture many behavioral and neural effects but do not explain recent findings showing that structure in the environment influences learning. In parallel, Bayesian cognitive models predict how humans learn structured knowledge but do not have a clear neurobiological implementation. We propose an integration of these two model classes in which structured knowledge learned via approximate Bayesian inference acts as a source of selective attention. In turn, selective attention biases reinforcement learning towards relevant dimensions of the environment. An understanding of structure learning will help to resolve the fundamental challenge in decision science: explaining why people make the decisions they do.

Feature-specific prediction errors and surprise across macaque fronto-striatal circuits

To adjust expectations efficiently, prediction errors need to be associated with the precise features that gave rise to the unexpected outcome, but this credit assignment may be problematic if stimuli differ on multiple dimensions and it is ambiguous which feature dimension caused the outcome. Here, we report a potential solution: neurons in four recorded areas of the anterior fronto-striatal networks encode prediction errors that are specific to feature values of different dimensions of attended multidimensional stimuli. The most ubiquitous prediction error occurred for the reward-relevant dimension. Feature-specific prediction error signals a) emerge on average shortly after non-specific prediction error signals, b) arise earliest in the anterior cingulate cortex and later in dorsolateral prefrontal cortex, caudate and ventral striatum, and c) contribute to feature-based stimulus selection after learning. Thus, a widely-distributed feature-specific eligibility trace may be used to update synaptic weights for improved feature-based attention.

In order to adjust expectations efficiently, prediction errors need to be associated with the features that gave rise to the unexpected outcome. Here, the authors show that neurons in anterior fronto-striatal networks encode prediction errors that are specific to feature values of different stimulus dimensions.

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.

Deterministic response strategies in a trial-and-error learning task

Trial-and-error learning is a universal strategy for establishing which actions are beneficial or harmful in new environments. However, learning stimulus-response associations solely via trial-and-error is often suboptimal, as in many settings dependencies among stimuli and responses can be exploited to increase learning efficiency. Previous studies have shown that in settings featuring such dependencies, humans typically engage high-level cognitive processes and employ advanced learning strategies to improve their learning efficiency. Here we analyze in detail the initial learning phase of a sample of human subjects (N = 85) performing a trial-and-error learning task with deterministic feedback and hidden stimulus-response dependencies. Using computational modeling, we find that the standard Q-learning model cannot sufficiently explain human learning strategies in this setting. Instead, newly introduced deterministic response models, which are theoretically optimal and transform stimulus sequences unambiguously into response sequences, provide the best explanation for 50.6% of the subjects. Most of the remaining subjects either show a tendency towards generic optimal learning (21.2%) or at least partially exploit stimulus-response dependencies (22.3%), while a few subjects (5.9%) show no clear preference for any of the employed models. After the initial learning phase, asymptotic learning performance during the subsequent practice phase is best explained by the standard Q-learning model. Our results show that human learning strategies in the presented trial-and-error learning task go beyond merely associating stimuli and responses via incremental reinforcement. Specifically during initial learning, high-level cognitive processes support sophisticated learning strategies that increase learning efficiency while keeping memory demands and computational efforts bounded. The good asymptotic fit of the Q-learning model indicates that these cognitive processes are successively replaced by the formation of stimulus-response associations over the course of learning.

Older adults fail to form stable task representations during model-based reversal inference

Older adults struggle in dealing with changeable and uncertain environments across several cognitive domains. This has been attributed to difficulties in forming adequate task representations that help navigate uncertain environments. Here, we investigate how, in older adults, inadequate task representations impact on model-based reversal learning. We combined computational modeling and pupillometry during a novel model-based reversal learning task, which allowed us to isolate the relevance of task representations at feedback evaluation. We find that older adults overestimate the changeability of task states and consequently are less able to converge on unequivocal task representations through learning. Pupillometric measures and behavioral data show that these unreliable task representations in older adults manifest as a reduced ability to focus on feedback that is relevant for updating task representations, and as a reduced metacognitive awareness in the accuracy of their actions. Instead, the data suggested older adults' choice behavior was more consistent with a guidance by uninformative feedback properties such as outcome valence. Our study highlights that an inability to form adequate task representations may be a crucial factor underlying older adults' impaired model-based inference.

•

Older adults overestimate the changeability of task states in uncertain environments.

•

Unreliable task representations impact model updating in older adults.

•

Older adults focus on outcome valence rather than model updating during feedback.

•

Older adults show reduced metacognitive awareness of their decision-making accuracy.

•

Pupil diameter encodes model updating and not surprise in young and older adults.

Dual-Task Processing With Identical Stimulus and Response Sets: Assessing the Importance of Task Representation in Dual-Task Interference

Limitations in our ability to produce two responses at the same time – that is, dual-task interference – are typically measured by comparing performance when two stimuli are presented and two responses are made in close temporal proximity to when a single stimulus is presented and a single response is made. While straightforward, this approach leaves open multiple possible sources for observed differences. For example, on dual-task trials, it is typically necessary to identify two stimuli nearly simultaneously, whereas on typical single-task trials, only one stimulus is presented at a time. These processes are different from selecting and producing two distinct responses and complicate the interpretation of dual- and single-task performance differences. Ideally, performance when two tasks are executed should be compared to conditions in which only a single task is executed, while holding constant all other stimuli, response, and control processing. We introduce an alternative dual-task procedure designed to approach this ideal. It holds stimulus processing constant while manipulating the number of “tasks.” Participants produced unimanual or bimanual responses to pairs of stimuli. For one set of stimuli (two-task set), the mappings were organized so an image of a face and a building were mapped to particular responses (including no response) on the left or right hands. For the other set of stimuli (one-task set), the stimuli indicated the same set of responses, but there was not a one-to-one mapping between the individual stimuli and responses. Instead, each stimulus pair had to be considered together to determine the appropriate unimanual or bimanual response. While the stimulus pairs were highly similar and the responses identical across the two conditions, performance was strikingly different. For the two-task set condition, bimanual responses were made more slowly than unimanual responses, reflecting typical dual-task interference, whereas for the one-task set, unimanual responses were made more slowly than bimanual. These findings indicate that dual-task costs occur, at least in part, because of the interfering effects of task representation rather than simply the additional stimulus, response, or other processing typically required on dual-task trials.