A primer on the use of computational modelling to investigate affective states, affective disorders and animal welfare in non-human animals.

Objective measures of animal emotion-like and mood-like states are essential for preclinical studies of affective disorders and for assessing the welfare of laboratory and other animals. However, the development and validation of measures of these affective states poses a challenge partly because the relationships between affect and its behavioural, physiological and cognitive signatures are complex. Here, we suggest that the crisp characterisations offered by computational modelling of the underlying, but unobservable, processes that mediate these signatures should provide better insights. Although this computational psychiatry approach has been widely used in human research in both health and disease, translational computational psychiatry studies remain few and far between. We explain how building computational models with data from animal studies could play a pivotal role in furthering our understanding of the aetiology of affective disorders, associated affective states and the likely underlying cognitive processes involved. We end by outlining the basic steps involved in a simple computational analysis.

Definition of early age at onset in bipolar disorder according to distinctive neurodevelopmental pathways: insights from the FACE-BD study.

BACKGROUND: Converging evidence suggests that a subgroup of bipolar disorder (BD) with an early age at onset (AAO) may develop from aberrant neurodevelopment. However, the definition of early AAO remains unprecise. We thus tested which age cut-off for early AAO best corresponds to distinguishable neurodevelopmental pathways. METHODS: We analyzed data from the FondaMental Advanced Center of Expertise-Bipolar Disorder cohort, a naturalistic sample of 4421 patients. First, a supervised learning framework was applied in binary classification experiments using neurodevelopmental history to predict early AAO, defined either with Gaussian mixture models (GMM) clustering or with each of the different cut-offs in the range 14 to 25 years. Second, an unsupervised learning approach was used to find clusters based on neurodevelopmental factors and to examine the overlap between such data-driven groups and definitions of early AAO used for supervised learning. RESULTS: A young cut-off, i.e. 14 up to 16 years, induced higher separability [mean nested cross-validation test AUROC = 0.7327 (± 0.0169) for ⩽16 years]. Predictive performance deteriorated increasing the cut-off or setting early AAO with GMM. Similarly, defining early AAO below 17 years was associated with a higher degree of overlap with data-driven clusters (Normalized Mutual Information = 0.41 for ⩽17 years) relatively to other definitions. CONCLUSIONS: Early AAO best captures distinctive neurodevelopmental patterns when defined as ⩽17 years. GMM-based definition of early AAO falls short of mapping to highly distinguishable neurodevelopmental pathways. These results should be used to improve patients' stratification in future studies of BD pathophysiology and biomarkers.

Prediction of depression symptoms in individual subjects with face and eye movement tracking.

BACKGROUND: Depression is a challenge to diagnose reliably and the current gold standard for trials of DSM-5 has been in agreement between two or more medical specialists. Research studies aiming to objectively predict depression have typically used brain scanning. Less expensive methods from cognitive neuroscience may allow quicker and more reliable diagnoses, and contribute to reducing the costs of managing the condition. In the current study we aimed to develop a novel inexpensive system for detecting elevated symptoms of depression based on tracking face and eye movements during the performance of cognitive tasks. METHODS: In total, 75 participants performed two novel cognitive tasks with verbal affective distraction elements while their face and eye movements were recorded using inexpensive cameras. Data from 48 participants (mean age 25.5 years, standard deviation of 6.1 years, 25 with elevated symptoms of depression) passed quality control and were included in a case-control classification analysis with machine learning. RESULTS: Classification accuracy using cross-validation (within-study replication) reached 79% (sensitivity 76%, specificity 82%), when face and eye movement measures were combined. Symptomatic participants were characterised by less intense mouth and eyelid movements during different stages of the two tasks, and by differences in frequencies and durations of fixations on affectively salient distraction words. CONCLUSIONS: Elevated symptoms of depression can be detected with face and eye movement tracking during the cognitive performance, with a close to clinically-relevant accuracy (~80%). Future studies should validate these results in larger samples and in clinical populations.

No increased circular inference in adults with high levels of autistic traits or autism.

Autism spectrum disorders have been proposed to arise from impairments in the probabilistic integration of prior knowledge with sensory inputs. Circular inference is one such possible impairment, in which excitation-to-inhibition imbalances in the cerebral cortex cause the reverberation and amplification of prior beliefs and sensory information. Recent empirical work has associated circular inference with the clinical dimensions of schizophrenia. Inhibition impairments have also been observed in autism, suggesting that signal reverberation might be present in that condition as well. In this study, we collected data from 21 participants with self-reported diagnoses of autism spectrum disorders and 155 participants with a broad range of autistic traits in an online probabilistic decision-making task (the fisher task). We used previously established Bayesian models to investigate possible associations between autistic traits or autism and circular inference. There was no correlation between prior or likelihood reverberation and autistic traits across the whole sample. Similarly, no differences in any of the circular inference model parameters were found between autistic participants and those with no diagnosis. Furthermore, participants incorporated information from both priors and likelihoods in their decisions, with no relationship between their weights and psychiatric traits, contrary to what common theories for both autism and schizophrenia would suggest. These findings suggest that there is no increased signal reverberation in autism, despite the known presence of excitation-to-inhibition imbalances. They can be used to further contrast and refine the Bayesian theories of schizophrenia and autism, revealing a divergence in the computational mechanisms underlying the two conditions.

Blunted medial prefrontal cortico-limbic reward-related effective connectivity and depression.

Major depressive disorder is a leading cause of disability and significant mortality, yet mechanistic understanding remains limited. Over the past decade evidence has accumulated from case-control studies that depressive illness is associated with blunted reward activation in the basal ganglia and other regions such as the medial prefrontal cortex. However it is unclear whether this finding can be replicated in a large number of subjects. The functional anatomy of the medial prefrontal cortex and basal ganglia has been extensively studied and the former has excitatory glutamatergic projections to the latter. Reduced effect of glutamatergic projections from the prefrontal cortex to the nucleus accumbens has been argued to underlie motivational disorders such as depression, and many prominent theories of major depressive disorder propose a role for abnormal cortico-limbic connectivity. However, it is unclear whether there is abnormal reward-linked effective connectivity between the medial prefrontal cortex and basal ganglia related to depression. While resting state connectivity abnormalities have been frequently reported in depression, it has not been possible to directly link these findings to reward-learning studies. Here, we tested two main hypotheses. First, mood symptoms are associated with blunted striatal reward prediction error signals in a large community-based sample of recovered and currently ill patients, similar to reports from a number of studies. Second, event-related directed medial prefrontal cortex to basal ganglia effective connectivity is abnormally increased or decreased related to the severity of mood symptoms. Using a Research Domain Criteria approach, data were acquired from a large community-based sample of subjects who participated in a probabilistic reward learning task during event-related functional MRI. Computational modelling of behaviour, model-free and model-based functional MRI, and effective connectivity dynamic causal modelling analyses were used to test hypotheses. Increased depressive symptom severity was related to decreased reward signals in areas which included the nucleus accumbens in 475 participants. Decreased reward-related effective connectivity from the medial prefrontal cortex to striatum was associated with increased depressive symptom severity in 165 participants. Decreased striatal activity may have been due to decreased cortical to striatal connectivity consistent with glutamatergic and cortical-limbic related theories of depression and resulted in reduced direct pathway basal ganglia output. Further study of basal ganglia pathophysiology is required to better understand these abnormalities in patients with depressive symptoms and syndromes.

Acquisition of visual priors and induced hallucinations in chronic schizophrenia.

Prominent theories suggest that symptoms of schizophrenia stem from learning deficiencies resulting in distorted internal models of the world. To test these theories further, we used a visual statistical learning task known to induce rapid implicit learning of the stimulus statistics. In this task, participants are presented with a field of coherently moving dots and are asked to report the presented direction of the dots (estimation task), and whether they saw any dots or not (detection task). Two of the directions were more frequently presented than the others. In controls, the implicit acquisition of the stimuli statistics influences their perception in two ways: (i) motion directions are perceived as being more similar to the most frequently presented directions than they really are (estimation biases); and (ii) in the absence of stimuli, participants sometimes report perceiving the most frequently presented directions (a form of hallucinations). Such behaviour is consistent with probabilistic inference, i.e. combining learnt perceptual priors with sensory evidence. We investigated whether patients with chronic, stable, treated schizophrenia (n = 20) differ from controls (n = 23) in the acquisition of the perceptual priors and/or their influence on perception. We found that although patients were slower than controls, they showed comparable acquisition of perceptual priors, approximating the stimulus statistics. This suggests that patients have no statistical learning deficits in our task. This may reflect our patients' relative wellbeing on antipsychotic medication. Intriguingly, however, patients experienced significantly fewer (P = 0.016) hallucinations of the most frequently presented directions than controls when the stimulus was absent or when it was very weak (prior-based lapse estimations). This suggests that prior expectations had less influence on patients' perception than on controls when stimuli were absent or below perceptual threshold.

Conditioned task-set competition: Neural mechanisms of emotional interference in depression.

Depression has been associated with increased response times at the incongruent-, neutral-, and negative-word trials of the classical and emotional Stroop tasks (Epp et al., Clinical Psychology Review, 32, 316-328, 2012). Response-time slowdown effects at incongruent- and negative-word trials of the Stroop tasks were reported to correlate with depressive severity, indicating strong relevance of the effects to the symptomatology. This study proposes a novel integrative computational model of neural mechanisms of both the classical and emotional Stroop effects, drawing on the previous prominent theoretical explanations of performance at the classical Stroop task (Cohen, Dunbar, & McClelland, Psychological Review, 97, 332-361, 1990; Herd, Banich, & O'Reilly, Journal of Cognitive Neuroscience, 18, 22-32, 2006), and in addition suggesting that negative emotional words represent conditioned stimuli for future negative outcomes. The model is shown to explain the classical Stroop effect and the slow (between-trial) emotional Stroop effect with biologically plausible mechanisms, providing an advantage over the previous theoretical accounts (Matthews & Harley, Cognition & Emotion, 10, 561-600, 1996; Wyble, Sharma, & Bowman, Cognition & Emotion, 22, 1019-1051, 2008). Simulation results suggested a candidate mechanism responsible for the pattern of depressive performance at the classical and the emotional Stroop tasks. Hyperactivity of the amygdala, together with increased inhibitory influence of the amygdala over dopaminergic neurotransmission, could be at the origin of the performance deficits.

Optimism as a prior belief about the probability of future reward.

Optimists hold positive a priori beliefs about the future. In Bayesian statistical theory, a priori beliefs can be overcome by experience. However, optimistic beliefs can at times appear surprisingly resistant to evidence, suggesting that optimism might also influence how new information is selected and learned. Here, we use a novel Pavlovian conditioning task, embedded in a normative framework, to directly assess how trait optimism, as classically measured using self-report questionnaires, influences choices between visual targets, by learning about their association with reward progresses. We find that trait optimism relates to an a priori belief about the likelihood of rewards, but not losses, in our task. Critically, this positive belief behaves like a probabilistic prior, i.e. its influence reduces with increasing experience. Contrary to findings in the literature related to unrealistic optimism and self-beliefs, it does not appear to influence the iterative learning process directly.

Expectations developed over multiple timescales facilitate visual search performance.

Our perception of the world is strongly influenced by our expectations, and a question of key importance is how the visual system develops and updates its expectations through interaction with the environment. We used a visual search task to investigate how expectations of different timescales (from the last few trials to hours to long-term statistics of natural scenes) interact to alter perception. We presented human observers with low-contrast white dots at 12 possible locations equally spaced on a circle, and we asked them to simultaneously identify the presence and location of the dots while manipulating their expectations by presenting stimuli at some locations more frequently than others. Our findings suggest that there are strong acuity differences between absolute target locations (e.g., horizontal vs. vertical) and preexisting long-term biases influencing observers' detection and localization performance, respectively. On top of these, subjects quickly learned about the stimulus distribution, which improved their detection performance but caused increased false alarms at the most frequently presented stimulus locations. Recent exposure to a stimulus resulted in significantly improved detection performance and significantly more false alarms but only at locations at which it was more probable that a stimulus would be presented. Our results can be modeled and understood within a Bayesian framework in terms of a near-optimal integration of sensory evidence with rapidly learned statistical priors, which are skewed toward the very recent history of trials and may help understanding the time scale of developing expectations at the neural level.

Confidence-based integrated reweighting model of task-difficulty explains location-based specificity in perceptual learning.

Perceptual learning is classically thought to be highly specific to the trained stimuli's retinal locations. However, recent research using a novel double-training paradigm has found dramatic transfer of perceptual learning to untrained locations. These results challenged existing models of perceptual learning and provoked intense debate in the field. Recently, Hung and Seitz (2014) showed that previously reported results could be reconciled by considering the details of the training procedure, in particular, whether it involves prolonged training at threshold using a single staircase procedure or multiple staircases. Here, we examine a hierarchical neural network model of the visual pathway, built upon previously proposed integrated reweighting models of perceptual learning, to understand how retinotopic transfer depends on the training procedure adopted. We propose that the transfer and specificity of learning between retinal locations can be explained by considering the task-difficulty and confidence during training. In our model, difficult tasks lead to higher learning of weights from early visual cortex to the decision unit, and thus to specificity, while easy tasks lead to higher learning of weights from later stages of the visual hierarchy and thus to more transfer. To model interindividual difference in task-difficulty, we relate task-difficulty to the confidence of subjects. We show that our confidence-based reweighting model can account for the results of Hung and Seitz (2014) and makes testable predictions.

Charles Bonnet syndrome: evidence for a generative model in the cortex?

Several theories propose that the cortex implements an internal model to explain, predict, and learn about sensory data, but the nature of this model is unclear. One condition that could be highly informative here is Charles Bonnet syndrome (CBS), where loss of vision leads to complex, vivid visual hallucinations of objects, people, and whole scenes. CBS could be taken as indication that there is a generative model in the brain, specifically one that can synthesise rich, consistent visual representations even in the absence of actual visual input. The processes that lead to CBS are poorly understood. Here, we argue that a model recently introduced in machine learning, the deep Boltzmann machine (DBM), could capture the relevant aspects of (hypothetical) generative processing in the cortex. The DBM carries both the semantics of a probabilistic generative model and of a neural network. The latter allows us to model a concrete neural mechanism that could underlie CBS, namely, homeostatic regulation of neuronal activity. We show that homeostatic plasticity could serve to make the learnt internal model robust against e.g. degradation of sensory input, but overcompensate in the case of CBS, leading to hallucinations. We demonstrate how a wide range of features of CBS can be explained in the model and suggest a potential role for the neuromodulator acetylcholine. This work constitutes the first concrete computational model of CBS and the first application of the DBM as a model in computational neuroscience. Our results lend further credence to the hypothesis of a generative model in the brain.

Can computational models help elucidate the link between complex trauma and hallucinations?

Recently, a number of predictive coding models have been proposed to account for post-traumatic stress disorder (PTSD)'s symptomatology, including intrusions, flashbacks and hallucinations. These models were usually developed to account for traditional/type-1 PTSD. We here discuss whether these models also apply or can be translated to the case of complex/type-2 PTSD and childhood trauma (cPTSD). The distinction between PTSD and cPTSD is important because the disorders differ in terms of symptomatology and potential mechanisms, how they relate to developmental stages, but also in terms of illness trajectory and treatment. Models of complex trauma could give us insights on hallucinations in physiological/pathological conditions or more generally on the development of intrusive experiences across diagnostic classes.

Designing optimal behavioral experiments using machine learning.

Computational models are powerful tools for understanding human cognition and behavior. They let us express our theories clearly and precisely and offer predictions that can be subtle and often counter-intuitive. However, this same richness and ability to surprise means our scientific intuitions and traditional tools are ill-suited to designing experiments to test and compare these models. To avoid these pitfalls and realize the full potential of computational modeling, we require tools to design experiments that provide clear answers about what models explain human behavior and the auxiliary assumptions those models must make. Bayesian optimal experimental design (BOED) formalizes the search for optimal experimental designs by identifying experiments that are expected to yield informative data. In this work, we provide a tutorial on leveraging recent advances in BOED and machine learning to find optimal experiments for any kind of model that we can simulate data from, and show how by-products of this procedure allow for quick and straightforward evaluation of models and their parameters against real experimental data. As a case study, we consider theories of how people balance exploration and exploitation in multi-armed bandit decision-making tasks. We validate the presented approach using simulations and a real-world experiment. As compared to experimental designs commonly used in the literature, we show that our optimal designs more efficiently determine which of a set of models best account for individual human behavior, and more efficiently characterize behavior given a preferred model. At the same time, formalizing a scientific question such that it can be adequately addressed with BOED can be challenging and we discuss several potential caveats and pitfalls that practitioners should be aware of. We provide code to replicate all analyses as well as tutorial notebooks and pointers to adapt the methodology to different experimental settings.

Attention as reward-driven optimization of sensory processing.

Attention causes diverse changes to visual neuron responses, including alterations in receptive field structure, and firing rates. A common theoretical approach to investigate why sensory neurons behave as they do is based on the efficient coding hypothesis: that sensory processing is optimized toward the statistics of the received input. We extend this approach to account for the influence of task demands, hypothesizing that the brain learns a probabilistic model of both the sensory input and reward received for performing different actions. Attention-dependent changes to neural responses reflect optimization of this internal model to deal with changes in the sensory environment (stimulus statistics) and behavioral demands (reward statistics). We use this framework to construct a simple model of visual processing that is able to replicate a number of attention-dependent changes to the responses of neurons in the midlevel visual cortices. The model is consistent with and provides a normative explanation for recent divisive normalization models of attention (Reynolds & Heeger, 2009).

Similarity-based extraction of individual networks from gray matter MRI scans.

The characterization of gray matter morphology of individual brains is an important issue in neuroscience. Graph theory has been used to describe cortical morphology, with networks based on covariation of gray matter volume or thickness between cortical areas across people. Here, we extend this research by proposing a new method that describes the gray matter morphology of an individual cortex as a network. In these large-scale morphological networks, nodes represent small cortical regions, and edges connect regions that have a statistically similar structure. The method was applied to a healthy sample (n = 14, scanned at 2 different time points). For all networks, we described the spatial degree distribution, average minimum path length, average clustering coefficient, small world property, and betweenness centrality (BC). Finally, we studied the reproducibility of all these properties. The networks showed more clustering than random networks and a similar minimum path length, indicating that they were "small world." The spatial degree and BC distributions corresponded closely to those from group-derived networks. All network property values were reproducible over the 2 time points examined. Our results demonstrate that intracortical similarities can be used to provide a robust statistical description of individual gray matter morphology.

Complexity and specificity of experimentally-induced expectations in motion perception.

Our perceptions are fundamentally altered by our expectations, i.e., priors about the world. In previous statistical learning experiments (Chalk, Seitz, & Seriès, 2010), we investigated how such priors are formed by presenting subjects with white low contrast moving dots on a blank screen and using a bimodal distribution of motion directions such that two directions were more frequently presented than the others. We found that human observers quickly and automatically developed expectations for the most frequently presented directions of motion. Here, we examine the specificity of these expectations. Can one learn simultaneously to expect different motion directions for dots of different colors? We interleaved moving dot displays of two different colors, either red or green, with different motion direction distributions. When one distribution was bimodal while the other was uniform, we found that subjects learned a single bimodal prior for the two stimuli. On the contrary, when both distributions were similarly structured, we found evidence for the formation of two distinct priors, which significantly influenced the subjects' behavior when no stimulus was presented. Our results can be modeled using a Bayesian framework and discussed in terms of a suboptimality of the statistical learning process under some conditions.

10 years of Bayesian theories of autism: A comprehensive review.

Ten years ago, Pellicano and Burr published one of the most influential articles in the study of autism spectrum disorders, linking them to aberrant Bayesian inference processes in the brain. In particular, they proposed that autistic individuals are less influenced by their brains' prior beliefs about the environment. In this systematic review, we investigate if this theory is supported by the experimental evidence. To that end, we collect all studies which included comparisons across diagnostic groups or autistic traits and categorise them based on the investigated priors. Our results are highly mixed, with a slight majority of studies finding no difference in the integration of Bayesian priors. We find that priors developed during the experiments exhibited reduced influences more frequently than priors acquired previously, with various studies providing evidence for learning differences between participant groups. Finally, we focus on the methodological and computational aspects of the included studies, showing low statistical power and often inconsistent approaches. Based on our findings, we propose guidelines for future research.

The effect of neural adaptation on population coding accuracy.

Most neurons in the primary visual cortex initially respond vigorously when a preferred stimulus is presented, but adapt as stimulation continues. The functional consequences of adaptation are unclear. Typically a reduction of firing rate would reduce single neuron accuracy as less spikes are available for decoding, but it has been suggested that on the population level, adaptation increases coding accuracy. This question requires careful analysis as adaptation not only changes the firing rates of neurons, but also the neural variability and correlations between neurons, which affect coding accuracy as well. We calculate the coding accuracy using a computational model that implements two forms of adaptation: spike frequency adaptation and synaptic adaptation in the form of short-term synaptic plasticity. We find that the net effect of adaptation is subtle and heterogeneous. Depending on adaptation mechanism and test stimulus, adaptation can either increase or decrease coding accuracy. We discuss the neurophysiological and psychophysical implications of the findings and relate it to published experimental data.

Fisher and Shannon information in finite neural populations.

The precision of the neural code is commonly investigated using two families of statistical measures: Shannon mutual information and derived quantities when investigating very small populations of neurons and Fisher information when studying large populations. These statistical tools are no longer the preserve of theorists and are being applied by experimental research groups in the analysis of empirical data. Although the relationship between information-theoretic and Fisher-based measures in the limit of infinite populations is relatively well understood, how these measures compare in finite-size populations has not yet been systematically explored. We aim to close this gap. We are particularly interested in understanding which stimuli are best encoded by a given neuron within a population and how this depends on the chosen measure. We use a novel Monte Carlo approach to compute a stimulus-specific decomposition of the mutual information (the SSI) for populations of up to 256 neurons and show that Fisher information can be used to accurately estimate both mutual information and SSI for populations of the order of 100 neurons, even in the presence of biologically realistic variability, noise correlations, and experimentally relevant integration times. According to both measures, the stimuli that are best encoded are those falling at the flanks of the neuron's tuning curve. In populations of fewer than around 50 neurons, however, Fisher information can be misleading.