Catecholaminergic Regulation of Learning Rate in a Dynamic Environment

Toward a computational role for locus coeruleus/norepinephrine arousal systems

Brain and behavior undergo measurable changes in their underlying state and neuromodulators are thought to contribute to these fluctuations. Why do we undergo such changes, and what function could the underlying neuromodulatory systems perform? Here we examine theoretical answers to these questions with respect to the locus coeruleus/norepinephrine system focusing on peripheral markers for arousal, such as pupil diameter, that are thought to provide a window into brain wide noradrenergic signaling. We explore a computational role for arousal systems in facilitating internal state transitions that facilitate credit assignment and promote accurate perceptions in non-stationary environments. We summarize recent work that supports this idea and highlight open questions as well as alternative views of how arousal affects cognition.

Regulation of reinforcement learning parameters captures long-term changes in rat behaviour

In uncertain environments in which resources fluctuate continuously, animals must permanently decide whether to stabilise learning and exploit what they currently believe to be their best option, or instead explore potential alternatives and learn fast from new observations. While such a trade-off has been extensively studied in pretrained animals facing non-stationary decision-making tasks, it is yet unknown how they progressively tune it while learning the task structure during pretraining. Here, we compared the ability of different computational models to account for long-term changes in the behaviour of 24 rats while they learned to choose a rewarded lever in a three-armed bandit task across 24 days of pretraining. We found that the day-by-day evolution of rat performance and win-shift tendency revealed a progressive stabilisation of the way they regulated reinforcement learning parameters. We successfully captured these behavioural adaptations using a meta-learning model in which either the learning rate or the inverse temperature was controlled by the average reward rate.

Correlated P300b and phasic pupil-dilation responses to motivationally significant stimuli

Motivationally significant events like oddball stimuli elicit both a characteristic event-related potential (ERPs) known as P300 and a set of autonomic responses including a phasic pupil dilation. Although co-occurring, P300 and pupil-dilation responses to oddball events have been repeatedly found to be uncorrelated, suggesting separate origins. We re-examined their relationship in the context of a three-stimulus version of the auditory oddball task, independently manipulating the frequency (rare vs. repeated) and motivational significance (relevance for the participant's task) of the stimuli. We used independent component analysis to derive a P300b component from EEG traces and linear modeling to separate a stimulus-related pupil-dilation response from a potentially confounding action-related response. These steps revealed that, once the complexity of ERP and pupil-dilation responses to oddball targets is accounted for, the amplitude of phasic pupil dilations and P300b are tightly and positively correlated (across participants: r = .69 p = .002), supporting their coordinated generation.

Prediction-error-dependent processing of immediate and delayed positive feedback

Learning often involves trial-and-error, i.e. repeating behaviours that lead to desired outcomes, and adjusting behaviour when outcomes do not meet our expectations and thus lead to prediction errors (PEs). PEs have been shown to be reflected in the reward positivity (RewP), an event-related potential (ERP) component between 200 and 350 ms after performance feedback which is linked to striatal processing and assessed via electroencephalography (EEG). Here we show that this is also true for delayed feedback processing, for which a critical role of the hippocampus has been suggested. We found a general reduction of the RewP for delayed feedback, but the PE was similarly reflected in the RewP and the later P300 for immediate and delayed positive feedback, while no effect was found for negative feedback. Our results suggest that, despite processing differences between immediate and delayed feedback, positive PEs drive feedback processing and learning irrespective of delay.

Adaptive learning from outcome contingencies in eating-disorder risk groups

Eating disorders are characterised by altered eating patterns alongside overvaluation of body weight or shape, and have relatively low rates of successful treatment and recovery. Notably, cognitive inflexibility has been implicated in both the development and maintenance of eating disorders, and understanding the reasons for this inflexibility might indicate avenues for treatment development. We therefore investigate one potential cause of this inflexibility: an inability to adjust learning when outcome contingencies change. We recruited (n = 82) three groups of participants: those who had recovered from anorexia nervosa (RA), those who had high levels of eating disorder symptoms but no formal diagnosis (EA), and control participants (HC). They performed a reinforcement learning task (alongside eye-tracking) in which the volatility of wins and losses was independently manipulated. We predicted that both the RA and EA groups would adjust their learning rates less than the control participants. Unexpectedly, the RA group showed elevated adjustment of learning rates for both win and loss outcomes compared to control participants. The RA group also showed increased pupil dilation to stable wins and reduced pupil dilation to stable losses. Their learning rate adjustment was associated with the difference between their pupil dilation to volatile vs. stable wins. In conclusion, we find evidence that learning rate adjustment is unexpectedly higher in those who have recovered from anorexia nervosa, indicating that the relationship between eating disorders and cognitive inflexibility may be complex. Given our findings, investigation of noradrenergic agents may be valuable in the field of eating disorders.

Dopaminergic and norepinephrinergic modulation of endogenous event-related potentials: A systematic review and meta-analysis

Event-related potentials (ERPs) represent the cortical processing of sensory, motor or cognitive functions invoked by particular events or stimuli. A current theory posits that the catecholaminergic neurotransmitters dopamine (DA) and norepinephrine (NE) modulate a number of endogenous ERPs during various cognitive processes. This manuscript aims to evaluate a leading neurotransmitter hypothesis with a systematic overview and meta-analysis of pharmacologic DA and NE manipulation of specific ERPs in healthy subjects during executive function. Specifically, the frontally-distributed P3a, N2, and Ne/ERN (or error-related negativity) are supposedly modulated primarily by DA, whereas the parietally-distributed P3b is thought to be modulated by NE. Based on preceding research, we refer to this distinction between frontally-distributed DA-sensitive and parietally-distributed NE-sensitive ERP components as the Extended Neurobiological Polich (ENP) hypothesis. Our systematic review and meta-analysis indicate that this distinction is too simplistic and many factors interact with DA and NE to influence these specific ERPs. These may include genetic factors, the specific cognitive processes engaged, or elements of study design, i.e. session or sequence effects or data-analysis strategies.

How uncertain are you? Disentangling expected and unexpected uncertainty in pupil-linked brain arousal during reversal learning

During decision making, we are continuously faced with two sources of uncertainty regarding the links between stimuli, our actions, and outcomes. On the one hand, our expectations are often probabilistic, that is, stimuli or actions yield the expected outcome only with a certain probability (expected uncertainty). On the other hand, expectations might become invalid due to sudden, unexpected changes in the environment (unexpected uncertainty). Several lines of research show that pupil-linked brain arousal is a sensitive indirect measure of brain mechanisms underlying uncertainty computations. Thus, we investigated whether it is involved in disentangling these two forms of uncertainty. To this aim, we measured pupil size during a probabilistic reversal learning task. In this task, participants had to figure out which of two response options led to reward with higher probability, whereby sometimes the identity of the more advantageous response option was switched. Expected uncertainty was manipulated by varying the reward probability of the advantageous choice option, whereas the level of unexpected uncertainty was assessed by using a Bayesian computational model estimating change probability and resulting uncertainty. We found that both aspects of unexpected uncertainty influenced pupil responses, confirming that pupil-linked brain arousal is involved in model updating after unexpected changes in the environment. Furthermore, high level of expected uncertainty impeded the detection of sudden changes in the environment, both on physiological and behavioral level. These results emphasize the role of pupil-linked brain arousal and underlying neural structures in handling situations in which the previously established contingencies are no longer valid.

Anterior cingulate and medial prefrontal cortex oscillations underlie learning alterations in trait anxiety in humans

Anxiety has been linked to altered belief formation and uncertainty estimation, impacting learning. Identifying the neural processes underlying these changes is important for understanding brain pathology. Here, we show that oscillatory activity in the medial prefrontal, anterior cingulate and orbitofrontal cortex (mPFC, ACC, OFC) explains anxiety-related learning alterations. In a magnetoencephalography experiment, two groups of human participants pre-screened with high and low trait anxiety (HTA, LTA: 39) performed a probabilistic reward-based learning task. HTA undermined learning through an overestimation of volatility, leading to faster belief updating, more stochastic decisions and pronounced lose-shift tendencies. On a neural level, we observed increased gamma activity in the ACC, dmPFC, and OFC during encoding of precision-weighted prediction errors in HTA, accompanied by suppressed ACC alpha/beta activity. Our findings support the association between altered learning and belief updating in anxiety and changes in gamma and alpha/beta activity in the ACC, dmPFC, and OFC.

Experience-driven recalibration of learning from surprising events

Different environments favor different patterns of adaptive learning. A surprising event that in one context would accelerate belief updating might, in another context, be downweighted as a meaningless outlier. Here, we investigated whether people would spontaneously regulate the influence of surprise on learning in response to event-by-event experiential feedback. Across two experiments, we examined whether participants performing a perceptual judgment task under spatial uncertainty (n = 29, n = 63) adapted their patterns of predictive gaze according to the informativeness or uninformativeness of surprising events in their current environment. Uninstructed predictive eye movements exhibited a form of metalearning in which surprise came to modulate event-by-event learning rates in opposite directions across contexts. Participants later appropriately readjusted their patterns of adaptive learning when the statistics of the environment underwent an unsignaled reversal. Although significant adjustments occurred in both directions, performance was consistently superior in environments in which surprising events reflected meaningful change, potentially reflecting a bias towards interpreting surprise as informative and/or difficulty ignoring salient outliers. Our results provide evidence for spontaneous, context-appropriate recalibration of the role of surprise in adaptive learning.

Feedback-related EEG dynamics separately reflect decision parameters, biases, and future choices

Optimal decision making in complex environments requires dynamic learning from unexpected events. To speed up learning, we should heavily weight information that indicates state-action-outcome contingency changes and ignore uninformative fluctuations in the environment. Often, however, unrelated information is hard to ignore and can potentially bias our learning. Here we used computational modelling and EEG to investigate learning behaviour in a modified probabilistic choice task that introduced two task-irrelevant factors that were uninformative for optimal task performance, but nevertheless could potentially bias learning: pay-out magnitudes were varied randomly and, occasionally, feedback presentation was enhanced by visual surprise. We found that participants' overall good learning performance was biased by distinct effects of these non-normative factors. On the neural level, these parameters are represented in a dynamic and spatiotemporally dissociable sequence of EEG activity. Later in feedback processing the different streams converged on a central to centroparietal positivity reflecting a signal that is interpreted by downstream learning processes that adjust future behaviour.

Disentangling performance-monitoring signals encoded in feedback-related EEG dynamics

The feedback-related negativity (FRN) is a well-established electrophysiological correlate of feedback-processing. However, there is still an ongoing debate whether the FRN is driven by negative or positive reward prediction errors (RPE), valence of feedback, or mere surprise. Our study disentangles independent contributions of valence, surprise, and RPE on the feedback-related neuronal signal including the FRN and P3 components using the statistical power of a sample of N = 992 healthy individuals. The participants performed a modified time-estimation task, while EEG from 64 scalp electrodes was recorded. Our results show that valence coding is present during the FRN with larger amplitudes for negative feedback. The FRN is further modulated by surprise in a valence-dependent way being more positive-going for surprising positive outcomes. The P3 was strongly driven by both global and local surprise, with larger amplitudes for unexpected feedback and local deviants. Behavioral adaptations after feedback and FRN just show small associations. Results support the theory of the FRN as a representation of a signed RPE. Additionally, our data indicates that surprising positive feedback enhances the EEG response in the time window of the P3. These results corroborate previous findings linking the P3 to the evaluation of PEs in decision making and learning tasks.

Noradrenergic deficits contribute to apathy in Parkinson's disease through the precision of expected outcomes

Apathy is a debilitating feature of many neuropsychiatric diseases, that is typically described as a reduction of goal-directed behaviour. Despite its prevalence and prognostic importance, the mechanisms underlying apathy remain controversial. Degeneration of the locus coeruleus-noradrenaline system is known to contribute to motivational deficits, including apathy. In healthy people, noradrenaline has been implicated in signalling the uncertainty of expectations about the environment. We proposed that noradrenergic deficits contribute to apathy by modulating the relative weighting of prior beliefs about action outcomes. We tested this hypothesis in the clinical context of Parkinson’s disease, given its associations with apathy and noradrenergic dysfunction. Participants with mild-to-moderate Parkinson’s disease (N = 17) completed a randomised double-blind, placebo-controlled, crossover study with 40 mg of the noradrenaline reuptake inhibitor atomoxetine. Prior weighting was inferred from psychophysical analysis of performance in an effort-based visuomotor task, and was confirmed as negatively correlated with apathy. Locus coeruleus integrity was assessed in vivo using magnetisation transfer imaging at ultra-high field 7T. The effect of atomoxetine depended on locus coeruleus integrity: participants with a more degenerate locus coeruleus showed a greater increase in prior weighting on atomoxetine versus placebo. The results indicate a contribution of the noradrenergic system to apathy and potential benefit from noradrenergic treatment of people with Parkinson’s disease, subject to stratification according to locus coeruleus integrity. More broadly, these results reconcile emerging predictive processing accounts of the role of noradrenaline in goal-directed behaviour with the clinical symptom of apathy and its potential pharmacological treatment.

Apathy is a common and harmful consequence of many neuropsychiatric diseases. Its underlying causes are not fully understood, which prevents the development of new treatments. We approach the problem in a new way, modelling human behaviour in terms of the continuously updated interaction between sensory information and brain-based predictions or ‘priors’ about the consequences of our actions. We have previously shown that apathy is related to a loss of precision of these ‘priors’. We proposed that the precision is controlled by noradrenaline (like adrenaline, but made in the brain). We tested whether the noradrenaline-enhancing drug called atomoxetine can restore the priors’ precision in apathetic people. We enrolled participants with Parkinson’s disease, which is associated with both apathy and noradrenaline loss. We used ultra-high field MRI to measure individual differences in the integrity of specialist region called the locus coeruleus–the brain’s source of noradrenaline. We found that the effect of treatment with atomoxetine on prior precision depended on locus coeruleus integrity: Participants with a degenerated locus coeruleus had a more positive change in prior precision. Our results highlight how individual differences in neuroanatomy can predict the potential benefit of noradrenaline treatments in people suffering from apathy.

Reinforcement Learning in Patients With Mood and Anxiety Disorders vs Control Individuals: A Systematic Review and Meta-analysis

IMPORTANCE

Computational psychiatry studies have investigated how reinforcement learning may be different in individuals with mood and anxiety disorders compared with control individuals, but results are inconsistent.

OBJECTIVE

To assess whether there are consistent differences in reinforcement-learning parameters between patients with depression or anxiety and control individuals.

DATA SOURCES

Web of Knowledge, PubMed, Embase, and Google Scholar searches were performed between November 15, 2019, and December 6, 2019, and repeated on December 3, 2020, and February 23, 2021, with keywords (reinforcement learning) AND (computational OR model) AND (depression OR anxiety OR mood).

STUDY SELECTION

Studies were included if they fit reinforcement-learning models to human choice data from a cognitive task with rewards or punishments, had a case-control design including participants with mood and/or anxiety disorders and healthy control individuals, and included sufficient information about all parameters in the models.

DATA EXTRACTION AND SYNTHESIS

Articles were assessed for inclusion according to MOOSE guidelines. Participant-level parameters were extracted from included articles, and a conventional meta-analysis was performed using a random-effects model. Subsequently, these parameters were used to simulate choice performance for each participant on benchmarking tasks in a simulation meta-analysis. Models were fitted, parameters were extracted using bayesian model averaging, and differences between patients and control individuals were examined. Overall effect sizes across analytic strategies were inspected.

MAIN OUTCOMES AND MEASURES

The primary outcomes were estimated reinforcement-learning parameters (learning rate, inverse temperature, reward learning rate, and punishment learning rate).

RESULTS

A total of 27 articles were included (3085 participants, 1242 of whom had depression and/or anxiety). In the conventional meta-analysis, patients showed lower inverse temperature than control individuals (standardized mean difference [SMD], -0.215; 95% CI, -0.354 to -0.077), although no parameters were common across all studies, limiting the ability to infer differences. In the simulation meta-analysis, patients showed greater punishment learning rates (SMD, 0.107; 95% CI, 0.107 to 0.108) and slightly lower reward learning rates (SMD, -0.021; 95% CI, -0.022 to -0.020) relative to control individuals. The simulation meta-analysis showed no meaningful difference in inverse temperature between patients and control individuals (SMD, 0.003; 95% CI, 0.002 to 0.004).

CONCLUSIONS AND RELEVANCE

The simulation meta-analytic approach introduced in this article for inferring meta-group differences from heterogeneous computational psychiatry studies indicated elevated punishment learning rates in patients compared with control individuals. This difference may promote and uphold negative affective bias symptoms and hence constitute a potential mechanistic treatment target for mood and anxiety disorders.

The interpretation of computational model parameters depends on the context

Reinforcement Learning (RL) models have revolutionized the cognitive and brain sciences, promising to explain behavior from simple conditioning to complex problem solving, to shed light on developmental and individual differences, and to anchor cognitive processes in specific brain mechanisms. However, the RL literature increasingly reveals contradictory results, which might cast doubt on these claims. We hypothesized that many contradictions arise from two commonly-held assumptions about computational model parameters that are actually often invalid: That parameters generalize between contexts (e.g. tasks, models) and that they capture interpretable (i.e. unique, distinctive) neurocognitive processes. To test this, we asked 291 participants aged 8–30 years to complete three learning tasks in one experimental session, and fitted RL models to each. We found that some parameters (exploration / decision noise) showed significant generalization: they followed similar developmental trajectories, and were reciprocally predictive between tasks. Still, generalization was significantly below the methodological ceiling. Furthermore, other parameters (learning rates, forgetting) did not show evidence of generalization, and sometimes even opposite developmental trajectories. Interpretability was low for all parameters. We conclude that the systematic study of context factors (e.g. reward stochasticity; task volatility) will be necessary to enhance the generalizability and interpretability of computational cognitive models.

Confidence-Controlled Hebbian Learning Efficiently Extracts Category Membership From Stimuli Encoded in View of a Categorization Task

In experiments on perceptual decision making, individuals learn a categorization task through trial-and-error protocols. We explore the capacity of a decision-making attractor network to learn a categorization task through reward-based, Hebbian-type modifications of the weights incoming from the stimulus encoding layer. For the latter, we assume a standard layer of a large number of stimulus-specific neurons. Within the general framework of Hebbian learning, we have hypothesized that the learning rate is modulated by the reward at each trial. Surprisingly, we find that when the coding layer has been optimized in view of the categorization task, such reward-modulated Hebbian learning (RMHL) fails to extract efficiently the category membership. In previous work, we showed that the attractor neural networks' nonlinear dynamics accounts for behavioral confidence in sequences of decision trials. Taking advantage of these findings, we propose that learning is controlled by confidence, as computed from the neural activity of the decision-making attractor network. Here we show that this confidence-controlled, reward-based Hebbian learning efficiently extracts categorical information from the optimized coding layer. The proposed learning rule is local and, in contrast to RMHL, does not require storing the average rewards obtained on previous trials. In addition, we find that the confidence-controlled learning rule achieves near-optimal performance. In accordance with this result, we show that the learning rule approximates a gradient descent method on a maximizing reward cost function.

Relative salience signaling within a thalamo-orbitofrontal circuit governs learning rate

Learning to predict rewards is essential for the sustained fitness of animals. Contemporary views suggest that such learning is driven by a reward prediction error (RPE)-the difference between received and predicted rewards. The magnitude of learning induced by an RPE is proportional to the product of the RPE and a learning rate. Here we demonstrate using two-photon calcium imaging and optogenetics in mice that certain functionally distinct subpopulations of ventral/medial orbitofrontal cortex (vmOFC) neurons signal learning rate control. Consistent with learning rate control, trial-by-trial fluctuations in vmOFC activity positively correlate with behavioral updating when the RPE is positive, and negatively correlates with behavioral updating when the RPE is negative. Learning rate is affected by many variables including the salience of a reward. We found that the average reward response of these neurons signals the relative salience of a reward, because it decreases after reward prediction learning or the introduction of another highly salient aversive stimulus. The relative salience signaling in vmOFC is sculpted by medial thalamic inputs. These results support emerging theoretical views that prefrontal cortex encodes and controls learning parameters.

Task Learnability Modulates Surprise but Not Valence Processing for Reinforcement Learning in Probabilistic Choice Tasks

The goal of temporal difference (TD) reinforcement learning is to maximize outcomes and improve future decision-making. It does so by utilizing a prediction error (PE), which quantifies the difference between the expected and the obtained outcome. In gambling tasks, however, decision-making cannot be improved because of the lack of learnability. On the basis of the idea that TD utilizes two independent bits of information from the PE (valence and surprise), we asked which of these aspects is affected when a task is not learnable. We contrasted behavioral data and ERPs in a learning variant and a gambling variant of a simple two-armed bandit task, in which outcome sequences were matched across tasks. Participants were explicitly informed that feedback could be used to improve performance in the learning task but not in the gambling task, and we predicted a corresponding modulation of the aspects of the PE. We used a model-based analysis of ERP data to extract the neural footprints of the valence and surprise information in the two tasks. Our results revealed that task learnability modulates reinforcement learning via the suppression of surprise processing but leaves the processing of valence unaffected. On the basis of our model and the data, we propose that task learnability can selectively suppress TD learning as well as alter behavioral adaptation based on a flexible cost-benefit arbitration.

Adaptive learning is structure learning in time

People use information flexibly. They often combine multiple sources of relevant information over time in order to inform decisions with little or no interference from intervening irrelevant sources. They adjust the degree to which they use new information over time rationally in accordance with environmental statistics and their own uncertainty. They can even use information gained in one situation to solve a problem in a very different one. Learning flexibly rests on the ability to infer the context at a given time, and therefore knowing which pieces of information to combine and which to separate. We review the psychological and neural mechanisms behind adaptive learning and structure learning to outline how people pool together relevant information, demarcate contexts, prevent interference between information collected in different contexts, and transfer information from one context to another. By examining all of these processes through the lens of optimal inference we bridge concepts from multiple fields to provide a unified multi-system view of how the brain exploits structure in time to optimize learning.

Baseline-dependent effect of dopamine's precursor L-tyrosine on working memory gating but not updating

Adaptive goal-directed behavior requires a dynamic balance between maintenance and updating within working memory (WM). This balance is controlled by an input-gating mechanism implemented by dopamine in the basal ganglia. Given that dopaminergic manipulations can modulate performance on WM-related tasks, it is important to gain mechanistic insight into whether such manipulations differentially affect updating (i.e., encoding and removal) and the closely-related gate opening/closing processes that respectively enable/prevent updating. To clarify this issue, 2.0 g of dopamine’s precursor L-tyrosine was administered to healthy young adults (N = 45) in a double-blind, placebo-controlled, within-subjects study. WM processes were empirically distinguished using the reference-back paradigm, which isolates performance related to updating, gate opening, and gate closing. L-tyrosine had a selective, baseline-dependent effect only on gate opening, which was evidenced by markedly reduced variance across subjects in gate opening performance in the L-tyrosine compared with the placebo condition, whereas the whole-sample average performance did not differ between conditions. This indicates a pattern of results whereby low-performing subjects improved, whereas high-performing subjects were impaired on L-tyrosine. Importantly, this inverted U-shaped pattern was not explained by regression to the mean. These results are consistent with an inverted-U relationship between dopamine and WM, and they indicate that updating and gating are differentially affected by a dopaminergic manipulation. This highlights the importance of distinguishing these processes when studying WM, for example, in the context of WM deficits in disorders with a dopaminergic pathophysiology.

Electrophysiological Signatures of Hierarchical Learning

Human perception and learning is thought to rely on a hierarchical generative model that is continuously updated via precision-weighted prediction errors (pwPEs). However, the neural basis of such cognitive process and how it unfolds during decision-making remain poorly understood. To investigate this question, we combined a hierarchical Bayesian model (i.e., Hierarchical Gaussian Filter [HGF]) with electroencephalography (EEG), while participants performed a probabilistic reversal learning task in alternatingly stable and volatile environments. Behaviorally, the HGF fitted significantly better than two control, nonhierarchical, models. Neurally, low-level and high-level pwPEs were independently encoded by the P300 component. Low-level pwPEs were reflected in the theta (4-8 Hz) frequency band, but high-level pwPEs were not. Furthermore, the expressions of high-level pwPEs were stronger for participants with better HGF fit. These results indicate that the brain employs hierarchical learning and encodes both low- and high-level learning signals separately and adaptively.

Pupil Dilation and the Slow Wave ERP Reflect Surprise about Choice Outcome Resulting from Intrinsic Variability in Decision Confidence

Central to human and animal cognition is the ability to learn from feedback in order to optimize future rewards. Such a learning signal might be encoded and broadcasted by the brain's arousal systems, including the noradrenergic locus coeruleus. Pupil responses and the positive slow wave component of event-related potentials reflect rapid changes in the arousal level of the brain. Here, we ask whether and how these variables may reflect surprise: the mismatch between one's expectation about being correct and the outcome of a decision, when expectations fluctuate due to internal factors (e.g., engagement). We show that during an elementary decision task in the face of uncertainty both physiological markers of phasic arousal reflect surprise. We further show that pupil responses and slow wave event-related potential are unrelated to each other and that prediction error computations depend on feedback awareness. These results further advance our understanding of the role of central arousal systems in decision-making under uncertainty.

The Computational, Pharmacological, and Physiological Determinants of Sensory Learning under Uncertainty

The ability to represent and respond to uncertainty is fundamental to human cognition and decision-making. Noradrenaline (NA) is hypothesized to play a key role in coordinating the sensory, learning, and physiological states necessary to adapt to a changing world, but direct evidence for this is lacking in humans. Here, we tested the effects of attenuating noradrenergic neurotransmission on learning under uncertainty. We probed the effects of the β-adrenergic receptor antagonist propranolol (40 mg) using a between-subjects, double-blind, placebo-controlled design. Participants performed a probabilistic associative learning task, and we employed a hierarchical learning model to formally quantify prediction errors about cue-outcome contingencies and changes in these associations over time (volatility). Both unexpectedness and noise slowed down reaction times, but propranolol augmented the interaction between these main effects such that behavior was influenced more by prior expectations when uncertainty was high. Computationally, this was driven by a reduction in learning rates, with people slower to update their beliefs in the face of new information. Attenuating the global effects of NA also eliminated the phasic effects of prediction error and volatility on pupil size, consistent with slower belief updating. Finally, estimates of environmental volatility were predicted by baseline cardiac measures in all participants. Our results demonstrate that NA underpins behavioral and computational responses to uncertainty. These findings have important implications for understanding the impact of uncertainty on human biology and cognition.

Fast fronto-parietal cortical dynamics of conflict detection and context updating in a flanker task

Recent research has found that the traditional target P3 consists of a family of P3-like positivities that can be functionally and topographically dissociated from one another. The current study examined target N2 and P3-like subcomponents indexing conflict detection and context updating at low- and high-order levels in the neural hierarchy during cognitive control. Electroencephalographic signals were recorded from 45 young adults while they completed a hybrid go/nogo flanker task, and Residue Iteration Decomposition (RIDE) was applied to functionally dissociate these peaks. Analyses showed a stimulus-locked frontal N2 revealing early detection and fast perceptual categorization of nogo, congruent and incongruent trials, resulting in frontal P3-like activity elicited by nogo trials in the latency-variable RIDE cluster, and by incongruent trials in the response-locked cluster. The congruent trials did not elicit frontal P3-like activity. These findings suggest that behavioral incongruency effects are related to intermediate and later stages of motor response re-programming.

State anxiety biases estimates of uncertainty and impairs reward learning in volatile environments

Clinical and subclinical (trait) anxiety impairs decision making and interferes with learning. Less understood are the effects of temporary anxious states on learning and decision making in healthy populations, and whether these can serve as a model for clinical anxiety. Here we test whether anxious states in healthy individuals elicit a pattern of aberrant behavioural, neural, and physiological responses comparable with those found in anxiety disorders-particularly when processing uncertainty in unstable environments. In our study, both a state anxious and a control group learned probabilistic stimulus-outcome mappings in a volatile task environment while we recorded their electrophysiological (EEG) signals. By using a hierarchical Bayesian model of inference and learning, we assessed the effect of state anxiety on Bayesian belief updating with a focus on uncertainty estimates. State anxiety was associated with an underestimation of environmental uncertainty, and informational uncertainty about the reward tendency. Anxious individuals' beliefs about reward contingencies were more precise (had smaller uncertainty) and thus more resistant to updating, ultimately leading to impaired reward-based learning. State anxiety was also associated with greater uncertainty about volatility. We interpret this pattern as evidence that state anxious individuals are less tolerant to informational uncertainty about the contingencies governing their environment and more willing to be uncertain about the level of stability of the world itself. Further, we tracked the neural representation of belief update signals in the trial-by-trial EEG amplitudes. In control participants, lower-level precision-weighted prediction errors (pwPEs) about reward tendencies were represented in the ERP signals across central and parietal electrodes peaking at 496 ms, overlapping with the late P300 in classical ERP analysis. The state anxiety group did not exhibit a significant representation of low-level pwPEs, and there were no significant differences between the groups. Smaller variance in low-level pwPE about reward tendencies in state anxiety could partially account for the null results. Expanding previous computational work on trait anxiety, our findings establish that temporary anxious states in healthy individuals impair reward-based learning in volatile environments, primarily through changes in uncertainty estimates, which play a central role in current Bayesian accounts of perceptual inference and learning.

Uncertainty-driven regulation of learning and exploration in adolescents: A computational account

Healthy adults flexibly adapt their learning strategies to ongoing changes in uncertainty, a key feature of adaptive behaviour. However, the developmental trajectory of this ability is yet unknown, as developmental studies have not incorporated trial-to-trial variation in uncertainty in their analyses or models. To address this issue, we compared adolescents’ and adults’ trial-to-trial dynamics of uncertainty, learning rate, and exploration in two tasks that assess learning in noisy but otherwise stable environments. In an estimation task—which provides direct indices of trial-specific learning rate—both age groups reduced their learning rate over time, as self-reported uncertainty decreased. Accordingly, the estimation data in both groups was better explained by a Bayesian model with dynamic learning rate (Kalman filter) than by conventional reinforcement-learning models. Furthermore, adolescents’ learning rates asymptoted at a higher level, reflecting an over-weighting of the most recent outcome, and the estimated Kalman-filter parameters suggested that this was due to an overestimation of environmental volatility. In a choice task, both age groups became more likely to choose the higher-valued option over time, but this increase in choice accuracy was smaller in the adolescents. In contrast to the estimation task, we found no evidence for a Bayesian expectation-updating process in the choice task, suggesting that estimation and choice tasks engage different learning processes. However, our modeling results of the choice task suggested that both age groups reduced their degree of exploration over time, and that the adolescents explored overall more than the adults. Finally, age-related differences in exploration parameters from fits to the choice data were mediated by participants’ volatility parameter from fits to the estimation data. Together, these results suggest that adolescents overestimate the rate of environmental change, resulting in elevated learning rates and increased exploration, which may help understand developmental changes in learning and decision-making.

To successfully learn the value of stimuli and actions, people should take into account their current (un)certainty about these values: Learning rates and exploration should be high when one’s value estimates are highly uncertain (in the beginning of learning), and decrease over time as evidence accumulates and uncertainty decreases. Recent studies have shown that healthy adults flexibly adapt their learning strategies based on ongoing changes in uncertainty, consistent with normative learning. However, the development of this ability prior to adulthood is yet unknown, as developmental learning studies have not considered trial-to-trial changes in uncertainty. Here, we show that adolescents, as compared to adults, showed a smaller decrease in both learning rate and exploration over time. Computational modeling revealed that both of these effects were due to adolescents overestimating the amount of environmental volatility, which made them more sensitive to recent relative to older evidence. The overestimation of volatility during adolescence may represent the rapidly changing environmental demands during this developmental period, and can help understand the surge in real-life risk taking and exploratory behaviours characteristic of adolescents.

Noradrenergic Regulation of Cognitive Flexibility: No Effects of Stress, Transcutaneous Vagus Nerve Stimulation, and Atomoxetine on Task-switching in Humans

Cognitive flexibility allows us to adaptively switch between different responsibilities in important domains of our daily life. Previous work has elucidated the neurochemical basis underlying the ability to switch responses to a previously nonreinforced exemplar and to switch between attentional sets. However, the role of neuromodulators in task switching, the ability to rapidly switch between two or more cognitive tasks afforded by the same stimuli, is still poorly understood. We attempted to fill this gap by manipulating norepinephrine levels using stress manipulation (Study 1a, n = 48; between-group design), transcutaneous vagus nerve stimulation at two different intensities (Study 1b, n = 48; sham-controlled between-group design), and pharmacological manipulation (Study 2, n = 24; double-blind crossover design), all of which increased salivary cortisol measures. Participants repeatedly switched between two cognitive tasks (classifying a digit as high/low [Task 1] or as odd/even [Task 2]), depending on the preceding cue. On each trial, a cue indicated the task to be performed. The cue-stimulus interval was varied to manipulate the time to prepare for the switch. Participants showed typical switch costs, which decreased with the time available for preparation. None of the manipulations modulated the size of the switch costs or the preparation effect, as supported by frequentist and Bayesian model comparisons. Task-switching performance reflects a complex mix of cognitive control and bottom-up dynamics of task-set representations. Our findings suggest that norepinephrine does not affect either of these aspects of cognitive flexibility.

Ketamine Affects Prediction Errors about Statistical Regularities: A Computational Single-Trial Analysis of the Mismatch Negativity

The auditory mismatch negativity (MMN) is significantly reduced in schizophrenia. Notably, a similar MMN reduction can be achieved with NMDA receptor (NMDAR) antagonists. Both phenomena have been interpreted as reflecting an impairment of predictive coding or, more generally, the "Bayesian brain" notion that the brain continuously updates a hierarchical model to infer the causes of its sensory inputs. Specifically, neurobiological interpretations of predictive coding view perceptual inference as an NMDAR-dependent process of minimizing hierarchical precision-weighted prediction errors (PEs), and disturbances of this putative process play a key role in hierarchical Bayesian theories of schizophrenia. Here, we provide empirical evidence for this theory, demonstrating the existence of multiple, hierarchically related PEs in a "roving MMN" paradigm. We applied a hierarchical Bayesian model to single-trial EEG data from healthy human volunteers of either sex who received the NMDAR antagonist S-ketamine in a placebo-controlled, double-blind, within-subject fashion. Using an unrestricted analysis of the entire time-sensor space, our trial-by-trial analysis indicated that low-level PEs (about stimulus transitions) are expressed early (102-207 ms poststimulus), while high-level PEs (about transition probability) are reflected by later components (152-199 and 215-277 ms) of single-trial responses. Furthermore, we find that ketamine significantly diminished the expression of high-level PE responses, implying that NMDAR antagonism disrupts the inference on abstract statistical regularities. Our findings suggest that NMDAR dysfunction impairs hierarchical Bayesian inference about the world's statistical structure. Beyond the relevance of this finding for schizophrenia, our results illustrate the potential of computational single-trial analyses for assessing potential pathophysiological mechanisms.

Brain dynamics for confidence-weighted learning

Learning in a changing, uncertain environment is a difficult problem. A popular solution is to predict future observations and then use surprising outcomes to update those predictions. However, humans also have a sense of confidence that characterizes the precision of their predictions. Bayesian models use a confidence-weighting principle to regulate learning: for a given surprise, the update is smaller when the confidence about the prediction was higher. Prior behavioral evidence indicates that human learning adheres to this confidence-weighting principle. Here, we explored the human brain dynamics sub-tending the confidence-weighting of learning using magneto-encephalography (MEG). During our volatile probability learning task, subjects’ confidence reports conformed with Bayesian inference. MEG revealed several stimulus-evoked brain responses whose amplitude reflected surprise, and some of them were further shaped by confidence: surprise amplified the stimulus-evoked response whereas confidence dampened it. Confidence about predictions also modulated several aspects of the brain state: pupil-linked arousal and beta-range (15–30 Hz) oscillations. The brain state in turn modulated specific stimulus-evoked surprise responses following the confidence-weighting principle. Our results thus indicate that there exist, in the human brain, signals reflecting surprise that are dampened by confidence in a way that is appropriate for learning according to Bayesian inference. They also suggest a mechanism for confidence-weighted learning: confidence about predictions would modulate intrinsic properties of the brain state to amplify or dampen surprise responses evoked by discrepant observations.

Learning in a changing and uncertain world is difficult. In this context, facing a discrepancy between my current belief and new observations may reflect random fluctuations (e.g. my commute train is unexpectedly late, but it happens sometimes), if so, I should ignore this discrepancy and not change erratically my belief. However, this discrepancy could also denote a profound change (e.g. the train company changed and is less reliable), in this case, I should promptly revise my current belief. Human learning is adaptive: we change how much we learn from new observations, in particular, we promote flexibility when facing profound changes. A mathematical analysis of the problem shows that we should increase flexibility when the confidence about our current belief is low, which occurs when a change is suspected. Here, I show that human learners entertain rational confidence levels during the learning of changing probabilities. This confidence modulates intrinsic properties of the brain state (oscillatory activity and neuromodulation) which in turn amplifies or reduces, depending on whether confidence is low or high, the neural responses to discrepant observations. This confidence-weighting mechanism could underpin adaptive learning.

Stimulation of the vagus nerve reduces learning in a go/no-go reinforcement learning task

When facing decisions to approach rewards or to avoid punishments, we often figuratively go with our gut, and the impact of metabolic states such as hunger on motivation are well documented. However, whether and how vagal feedback signals from the gut influence instrumental actions is unknown. Here, we investigated the effect of non-invasive transcutaneous auricular vagus nerve stimulation (taVNS) vs. sham (randomized cross-over design) on approach and avoidance behavior using an established go/no-go reinforcement learning paradigm in 39 healthy human participants (23 female) after an overnight fast. First, mixed-effects logistic regression analysis of choice accuracy showed that taVNS acutely impaired decision-making, p = .041. Computational reinforcement learning models identified the cause of this as a reduction in the learning rate through taVNS (∆α = -0.092, p_boot = .002), particularly after punishment (∆α_Pun = -0.081, p_boot = .012 vs. ∆α_Rew =-0.031, p_boot = .22). However, taVNS had no effect on go biases, Pavlovian response biases or response time. Hence, taVNS appeared to influence learning rather than action execution. These results highlight a novel role of vagal afferent input in modulating reinforcement learning by tuning the learning rate according to homeostatic needs.

Catecholaminergic modulation of meta-learning

The remarkable expedience of human learning is thought to be underpinned by meta-learning, whereby slow accumulative learning processes are rapidly adjusted to the current learning environment. To date, the neurobiological implementation of meta-learning remains unclear. A burgeoning literature argues for an important role for the catecholamines dopamine and noradrenaline in meta-learning. Here, we tested the hypothesis that enhancing catecholamine function modulates the ability to optimise a meta-learning parameter (learning rate) as a function of environmental volatility. 102 participants completed a task which required learning in stable phases, where the probability of reinforcement was constant, and volatile phases, where probabilities changed every 10-30 trials. The catecholamine transporter blocker methylphenidate enhanced participants' ability to adapt learning rate: Under methylphenidate, compared with placebo, participants exhibited higher learning rates in volatile relative to stable phases. Furthermore, this effect was significant only with respect to direct learning based on the participants' own experience, there was no significant effect on inferred-value learning where stimulus values had to be inferred. These data demonstrate a causal link between catecholaminergic modulation and the adjustment of the meta-learning parameter learning rate.

Motivational deficits in schizophrenia relate to abnormalities in cortical learning rate signals

Individuals from across the psychosis spectrum display impairments in reinforcement learning. In some individuals, these deficits may result from aberrations in reward prediction error (RPE) signaling, conveyed by dopaminergic projections to the ventral striatum (VS). However, there is mounting evidence that VS RPE signals are relatively intact in medicated people with schizophrenia (PSZ). We hypothesized that, in PSZ, reinforcement learning deficits often are not related to RPE signaling per se but rather their impact on learning and behavior (i.e., learning rate modulation), due to dysfunction in anterior cingulate and dorsomedial prefrontal cortex (dmPFC). Twenty-six PSZ and 23 healthy volunteers completed a probabilistic reinforcement learning paradigm with occasional, sudden, shifts in contingencies. Using computational modeling, we found evidence of an impairment in trial-wise learning rate modulation (α) in PSZ before and after a reinforcement contingency shift, expressed most in PSZ with more severe motivational deficits. In a subsample of 22 PSZ and 22 healthy volunteers, we found little evidence for between-group differences in VS RPE and dmPFC learning rate signals, as measured with fMRI. However, a follow-up psychophysiological interaction analysis revealed decreased dmPFC-VS connectivity concurrent with learning rate modulation, most prominently in individuals with the most severe motivational deficits. These findings point to an impairment in learning rate modulation in PSZ, leading to a reduced ability to adjust task behavior in response to unexpected outcomes. At the level of the brain, learning rate modulation deficits may be associated with decreased involvement of the dmPFC within a greater RL network.

Eye pupil signals information gain

In conditions of constant illumination, the eye pupil diameter indexes the modulation of arousal state and responds to a large breadth of cognitive processes, including mental effort, attention, surprise, decision processes, decision biases, value beliefs, uncertainty, volatility, exploitation/exploration trade-off, or learning rate. Here, I propose an information theoretic framework that has the potential to explain the ensemble of these findings as reflecting pupillary response to information processing. In short, updates of the brain’s internal model, quantified formally as the Kullback–Leibler (KL) divergence between prior and posterior beliefs, would be the common denominator to all these instances of pupillary dilation to cognition. I show that stimulus presentation leads to pupillary response that is proportional to the amount of information the stimulus carries about itself and to the quantity of information it provides about other task variables. In the context of decision making, pupil dilation in relation to uncertainty is explained by the wandering of the evidence accumulation process, leading to large summed KL divergences. Finally, pupillary response to mental effort and variations in tonic pupil size are also formalized in terms of information theory. On the basis of this framework, I compare pupillary data from past studies to simple information-theoretic simulations of task designs and show good correspondance with data across studies. The present framework has the potential to unify the large set of results reported on pupillary dilation to cognition and to provide a theory to guide future research.

Pupil-linked phasic arousal evoked by violation but not emergence of regularity within rapid sound sequences

The ability to track the statistics of our surroundings is a key computational challenge. A prominent theory proposes that the brain monitors for unexpected uncertainty - events which deviate substantially from model predictions, indicating model failure. Norepinephrine is thought to play a key role in this process by serving as an interrupt signal, initiating model-resetting. However, evidence is from paradigms where participants actively monitored stimulus statistics. To determine whether Norepinephrine routinely reports the statistical structure of our surroundings, even when not behaviourally relevant, we used rapid tone-pip sequences that contained salient pattern-changes associated with abrupt structural violations vs. emergence of regular structure. Phasic pupil dilations (PDR) were monitored to assess Norepinephrine. We reveal a remarkable specificity: When not behaviourally relevant, only abrupt structural violations evoke a PDR. The results demonstrate that Norepinephrine tracks unexpected uncertainty on rapid time scales relevant to sensory signals.

Statistical context dictates the relationship between feedback-related EEG signals and learning

Learning should be adjusted according to the surprise associated with observed outcomes but calibrated according to statistical context. For example, when occasional changepoints are expected, surprising outcomes should be weighted heavily to speed learning. In contrast, when uninformative outliers are expected to occur occasionally, surprising outcomes should be less influential. Here we dissociate surprising outcomes from the degree to which they demand learning using a predictive inference task and computational modeling. We show that the P300, a stimulus-locked electrophysiological response previously associated with adjustments in learning behavior, does so conditionally on the source of surprise. Larger P300 signals predicted greater learning in a changing context, but less learning in a context where surprise was indicative of a one-off outlier (oddball). Our results suggest that the P300 provides a surprise signal that is interpreted by downstream learning processes differentially according to statistical context in order to appropriately calibrate learning across complex environments.

With an eye on uncertainty: Modelling pupillary responses to environmental volatility

Living creatures must accurately infer the nature of their environments. They do this despite being confronted by stochastic and context sensitive contingencies—and so must constantly update their beliefs regarding their uncertainty about what might come next. In this work, we examine how we deal with uncertainty that evolves over time. This prospective uncertainty (or imprecision) is referred to as volatility and has previously been linked to noradrenergic signals that originate in the locus coeruleus. Using pupillary dilatation as a measure of central noradrenergic signalling, we tested the hypothesis that changes in pupil diameter reflect inferences humans make about environmental volatility. To do so, we collected pupillometry data from participants presented with a stream of numbers. We generated these numbers from a process with varying degrees of volatility. By measuring pupillary dilatation in response to these stimuli—and simulating the inferences made by an ideal Bayesian observer of the same stimuli—we demonstrate that humans update their beliefs about environmental contingencies in a Bayes optimal way. We show this by comparing general linear (convolution) models that formalised competing hypotheses about the causes of pupillary changes. We found greater evidence for models that included Bayes optimal estimates of volatility than those without. We additionally explore the interaction between different causes of pupil dilation and suggest a quantitative approach to characterising a person’s prior beliefs about volatility.

Humans are constantly confronted with surprising events. To navigate such a world, we must understand the chances of an unexpected event occurring at any given point in time. We do this by creating a model of the world around us, in which we allow for these unexpected events to occur by holding beliefs about how volatile our environment is. In this work we explore the way in which we update our beliefs, demonstrating that this updating relies on the number of unexpected events in relation to the expected number. We do this by examining the pupil diameter, since—in controlled environments—changes in pupil diameter reflect our response to unexpected observations. Finally, we show that our methodology is appropriate for assessing the individual participant’s prior expectations about the amount of uncertainty in their environment.

Monetary feedback modulates performance and electrophysiological indices of belief updating in reward learning

Belief updating entails the incorporation of new information about the environment into internal models of the world. Bayesian inference is the statistically optimal strategy for performing belief updating in the presence of uncertainty. An important open question is whether the use of cognitive strategies that implement Bayesian inference is dependent upon motivational state and, if so, how this is reflected in electrophysiological signatures of belief updating in the brain. Here, we recorded the EEG of participants performing a simple reward learning task with both monetary and nonmonetary instructive feedback conditions. Our aim was to distinguish the influence of the rewarding properties of feedback on belief updating from the information content of the feedback itself. A Bayesian updating model allowed us to quantify different aspects of belief updating across trials, including the size of belief updates and the uncertainty of beliefs. Faster learning rates were observed in the monetary feedback condition compared to the instructive feedback condition, while belief updates were generally larger, and belief uncertainty smaller, with monetary compared to instructive feedback. Larger amplitudes in the monetary feedback condition were found for three ERP components: the P3a, the feedback-related negativity, and the late positive potential. These findings suggest that motivational state influences inference strategies in reward learning, and this is reflected in the electrophysiological correlates of belief updating.

Dopamine blockade impairs the exploration-exploitation trade-off in rats

In a volatile environment where rewards are uncertain, successful performance requires a delicate balance between exploitation of the best option and exploration of alternative choices. It has theoretically been proposed that dopamine contributes to the control of this exploration-exploitation trade-off, specifically that the higher the level of tonic dopamine, the more exploitation is favored. We demonstrate here that there is a formal relationship between the rescaling of dopamine positive reward prediction errors and the exploration-exploitation trade-off in simple non-stationary multi-armed bandit tasks. We further show in rats performing such a task that systemically antagonizing dopamine receptors greatly increases the number of random choices without affecting learning capacities. Simulations and comparison of a set of different computational models (an extended Q-learning model, a directed exploration model, and a meta-learning model) fitted on each individual confirm that, independently of the model, decreasing dopaminergic activity does not affect learning rate but is equivalent to an increase in random exploration rate. This study shows that dopamine could adapt the exploration-exploitation trade-off in decision-making when facing changing environmental contingencies.

Dorsal anterior cingulate-brainstem ensemble as a reinforcement meta-learner

Optimal decision-making is based on integrating information from several dimensions of decisional space (e.g., reward expectation, cost estimation, effort exertion). Despite considerable empirical and theoretical efforts, the computational and neural bases of such multidimensional integration have remained largely elusive. Here we propose that the current theoretical stalemate may be broken by considering the computational properties of a cortical-subcortical circuit involving the dorsal anterior cingulate cortex (dACC) and the brainstem neuromodulatory nuclei: ventral tegmental area (VTA) and locus coeruleus (LC). From this perspective, the dACC optimizes decisions about stimuli and actions, and using the same computational machinery, it also modulates cortical functions (meta-learning), via neuromodulatory control (VTA and LC). We implemented this theory in a novel neuro-computational model–the Reinforcement Meta Learner (RML). We outline how the RML captures critical empirical findings from an unprecedented range of theoretical domains, and parsimoniously integrates various previous proposals on dACC functioning.

A major challenge for all organisms is selecting optimal behaviour to obtain resources while minimizing energetic and other expenses. Evolution provided mammals with exceptional decision-making capabilities to face this challenge. Even though neuroscientists have identified a heterogeneous and distributed set of brain structures to be involved, a comprehensive theory about the biological and computational basis of such decision-making is yet to be formulated. We propose that the interaction between the medial prefrontal cortex (a part of the frontal lobes) and the subcortical nuclei releasing catecholaminergic neuromodulators will be key to such a theory. We argue that this interaction allows both the selection of optimal behaviour and, more importantly, the optimal modulation of the very brain circuits that drive such behavioral selection (i.e., meta-learning). We implemented this theory in a novel neuro-computational model, the Reinforcement Meta-Learner (RML). By means of computer simulations we showed that the RML provides a biological and computational account for a set of neuroscientific data with unprecedented scope, thereby suggesting a critical mechanism of decision-making in the mammalian brain.

Noradrenergic and Cholinergic Modulation of Belief Updating

To make optimal predictions in a dynamic environment, the impact of new observations on existing beliefs-that is, the learning rate-should be guided by ongoing estimates of change and uncertainty. Theoretical work has proposed specific computational roles for various neuromodulatory systems in the control of learning rate, but empirical evidence is still sparse. The aim of the current research was to examine the role of the noradrenergic and cholinergic systems in learning rate regulation. First, we replicated our recent findings that the centroparietal P3 component of the EEG-an index of phasic catecholamine release in the cortex-predicts trial-to-trial variability in learning rate and mediates the effects of surprise and belief uncertainty on learning rate (Study 1, n = 17). Second, we found that pharmacological suppression of either norepinephrine or acetylcholine activity produced baseline-dependent effects on learning rate following nonobvious changes in an outcome-generating process (Study 1). Third, we identified two genes, coding for α2A receptor sensitivity (ADRA2A) and norepinephrine reuptake (NET), as promising targets for future research on the genetic basis of individual differences in learning rate (Study 2, n = 137). Our findings suggest a role for the noradrenergic and cholinergic systems in belief updating and underline the importance of studying interactions between different neuromodulatory systems.

Visual Mismatch and Predictive Coding: A Computational Single-Trial ERP Study

Predictive coding (PC) posits that the brain uses a generative model to infer the environmental causes of its sensory data and uses precision-weighted prediction errors (pwPEs) to continuously update this model. While supported by much circumstantial evidence, experimental tests grounded in formal trial-by-trial predictions are rare. One partial exception is event-related potential (ERP) studies of the auditory mismatch negativity (MMN), where computational models have found signatures of pwPEs and related model-updating processes. Here, we tested this hypothesis in the visual domain, examining possible links between visual mismatch responses and pwPEs. We used a novel visual "roving standard" paradigm to elicit mismatch responses in humans (of both sexes) by unexpected changes in either color or emotional expression of faces. Using a hierarchical Bayesian model, we simulated pwPE trajectories of a Bayes-optimal observer and used these to conduct a comprehensive trial-by-trial analysis across the time × sensor space. We found significant modulation of brain activity by both color and emotion pwPEs. The scalp distribution and timing of these single-trial pwPE responses were in agreement with visual mismatch responses obtained by traditional averaging and subtraction (deviant-minus-standard) approaches. Finally, we compared the Bayesian model to a more classical change model of MMN. Model comparison revealed that trial-wise pwPEs explained the observed mismatch responses better than categorical change detection. Our results suggest that visual mismatch responses reflect trial-wise pwPEs, as postulated by PC. These findings go beyond classical ERP analyses of visual mismatch and illustrate the utility of computational analyses for studying automatic perceptual processes.SIGNIFICANCE STATEMENT Human perception is thought to rely on a predictive model of the environment that is updated via precision-weighted prediction errors (pwPEs) when events violate expectations. This "predictive coding" view is supported by studies of the auditory mismatch negativity brain potential. However, it is less well known whether visual perception of mismatch relies on similar processes. Here we combined computational modeling and electroencephalography to test whether visual mismatch responses reflected trial-by-trial pwPEs. Applying a Bayesian model to series of face stimuli that violated expectations about color or emotional expression, we found significant modulation of brain activity by both color and emotion pwPEs. A categorical change detection model performed less convincingly. Our findings support the predictive coding interpretation of visual mismatch responses.

Differential Noradrenergic Modulation of Monetary Reward and Visual Erotic Stimulus Processing

We recently investigated the effects of the noradrenergic antidepressant reboxetine and the antipsychotic amisulpride compared to placebo on neural correlates of primary reinforcers by visual erotic stimulation in healthy subjects. Whereas, amisulpride left subjective sexual functions and corresponding neural activations unimpaired, attenuated neural activations were observed under reboxetine within the nucleus accumbens (Nacc) along with diminished behavioral sexual functioning. However, a global dampening of the reward system under reboxetine seemed not intuitive considering the complementary role of the noradrenergic to the dopamine system in reward-related learning mediated by prediction error processing. We therefore investigated the sample of 17 healthy males in a mean age of 23.8 years again by functional magnetic resonance imaging (fMRI), to explore the noradrenergic effects on neural reward prediction error signaling. Participants took reboxetine (4 mg/d), amisulpride (200 mg/d), and placebo each for 7 days within a randomized, double-blind, within-subject cross-over design. During fMRI, we used an established monetary incentive task to assess neural reward expectation and prediction error signals within the bilateral Nacc using an independent anatomical mask for a region of interest (ROI) analysis. Activations within the same ROI were also assessed for the erotic picture paradigm. We confirmed our previous results from the whole brain analysis for the selected ROI by significant (p < 0.05 FWE-corrected) attenuated activations within the Nacc during visual sexual stimulation under reboxetine compared to placebo. However, activations in the Nacc concerning prediction error processing and monetary reward expectation were unimpaired under reboxetine compared to placebo, along with unimpaired reaction times in the reward task. For both tasks, neural activations and behavioral processing were not altered by amisulpride compared to placebo. The observed attenuated neural activations within the Nacc during visual erotic stimulation along with unimpaired neural prediction error and monetary reward expectation processing provide evidence for a differential modulation of the neural reward system by the noradrenergic agent reboxetine depending on the presence of primary reinforcers such as erotic stimuli in contrast to secondary such as monetary rewards.

Affective bias as a rational response to the statistics of rewards and punishments

Affective bias, the tendency to differentially prioritise the processing of negative relative to positive events, is commonly observed in clinical and non-clinical populations. However, why such biases develop is not known. Using a computational framework, we investigated whether affective biases may reflect individuals' estimates of the information content of negative relative to positive events. During a reinforcement learning task, the information content of positive and negative outcomes was manipulated independently by varying the volatility of their occurrence. Human participants altered the learning rates used for the outcomes selectively, preferentially learning from the most informative. This behaviour was associated with activity of the central norepinephrine system, estimated using pupilometry, for loss outcomes. Humans maintain independent estimates of the information content of distinct positive and negative outcomes which may bias their processing of affective events. Normalising affective biases using computationally inspired interventions may represent a novel approach to treatment development.