Perceptual instability in schizophrenia: Probing predictive coding accounts of delusions with ambiguous stimuli

Effect of prior beliefs and cognitive deficits on learning in first-episode schizophrenia patients

Introduction

It is known that cognitive deficits are a core feature of schizophrenia and that in the general population, prior beliefs significantly influence learning and reasoning processes. However, the interaction of prior beliefs with cognitive deficits and their impact on performance in schizophrenia patients is still poorly understood. This study investigates the role of beliefs and cognitive variables (CVs) like working memory, associative learning, and processing speed on learning processes in individuals with schizophrenia. We hypothesize that beliefs will influence the ability to learn correct predictions and that first-episode schizophrenia patients (FEP) will show impaired learning due to cognitive deficits.

Methods

We used a predictive-learning task to examine how FEP (n = 23) and matched controls (n = 23) adjusted their decisional criteria concerning physical properties during the learning process when predicting the sinking behavior of two transparent containers filled with aluminum discs when placed in water.

Results

On accuracy, initial differences by group, trial type, and interaction effects of these variables disappeared when CVs were controlled. The differences by conditions, associated with differential beliefs about why the objects sink slower or faster, were seen in patients and controls, despite controlling the CVs' effect.

Conclusions

Differences between groups were mainly explained by CVs, proving that they play an important role than what is assumed in this type of task. However, beliefs about physical events were not affected by CVs, and beliefs affect in the same way the decisional criteria of the control or FEP patients' groups.

Different learning aberrations relate to delusion-like beliefs with different contents

The prediction error account of delusions has had success. However, its explanation of delusions with different contents has been lacking. Persecutory delusions and paranoia are the common unfounded beliefs that others have harmful intentions towards us. Other delusions include believing that one's thoughts or actions are under external control or that events in the world have specific personal meaning. We compare learning in two different cognitive tasks, probabilistic reversal learning and Kamin blocking, that have relationships to paranoid and non-paranoid delusion-like beliefs, respectively. We find that clinical high-risk status alone does not result in different behavioural results in the probabilistic reversal learning task but that an individual's level of paranoia is associated with excessive switching behaviour. During the Kamin blocking task, paranoid individuals learned inappropriately about the blocked cue. However, they also had decreased learning about the control cue, suggesting more general learning impairments. Non-paranoid delusion-like belief conviction (but not paranoia) was associated with aberrant learning about the blocked cue but intact learning about the control cue, suggesting specific impairments in learning related to cue combination. We fit task-specific computational models separately to behavioural data to explore how latent parameters vary within individuals between tasks and how they can explain symptom-specific effects. We find that paranoia is associated with low learning rates in the probabilistic reversal learning task and the blocking task. Non-paranoid delusion-like belief conviction is instead related to parameters controlling the degree and direction of similarity between cue updating during simultaneous cue presentation. These results suggest that paranoia and other delusion-like beliefs involve dissociable deficits in learning and belief updating, which, given the transdiagnostic status of paranoia, might have differential utility in predicting psychosis.

Faster bi-stable visual switching in psychosis

Bi-stable stimuli evoke two distinct perceptual interpretations that alternate and compete for dominance. Bi-stable perception is thought to be driven at least in part by mutual suppression between distinct neural populations that represent each percept. Abnormal visual perception has been observed among people with psychotic psychopathology (PwPP), and there is evidence to suggest that these visual deficits may depend on impaired neural suppression in the visual cortex. However, it is not yet clear whether bi-stable visual perception is abnormal among PwPP. Here, we examined bi-stable perception in a visual structure-from-motion task using a rotating cylinder illusion in a group of 65 PwPP, 44 first-degree biological relatives, and 43 healthy controls. Data from a 'real switch' task, in which physical depth cues signaled real switches in rotation direction were used to exclude individuals who did not show adequate task performance. In addition, we measured concentrations of neurochemicals, including glutamate, glutamine, and γ-amino butyric acid (GABA), involved in excitatory and inhibitory neurotransmission. These neurochemicals were measured non-invasively in the visual cortex using 7 tesla MR spectroscopy. We found that PwPP and their relatives showed faster bi-stable switch rates than healthy controls. Faster switch rates also correlated with significantly higher psychiatric symptom levels, specifically disorganization, across all participants. However, we did not observe any significant relationships across individuals between neurochemical concentrations and SFM switch rates. Our results are consistent with a reduction in suppressive neural processes during structure-from-motion perception in PwPP, and suggest that genetic liability for psychosis is associated with disrupted bi-stable perception.

Visualizing threat and trustworthiness prior beliefs in face perception in high versus low paranoia

Predictive processing accounts of psychosis conceptualize delusions as overly strong learned expectations (prior beliefs) that shape cognition and perception. Paranoia, the most prevalent form of delusions, involves threat prior beliefs that are inherently social. Here, we investigated whether paranoia is related to overly strong threat prior beliefs in face perception. Participants with subclinical levels of high (n = 109) versus low (n = 111) paranoia viewed face stimuli paired with written descriptions of threatening versus trustworthy behaviors, thereby activating their threat versus trustworthiness prior beliefs. Subsequently, they completed an established social-psychological reverse correlation image classification (RCIC) paradigm. This paradigm used participants' responses to randomly varying face stimuli to generate individual classification images (ICIs) that intend to visualize either facial prior belief (threat vs. trust). An independent sample (n = 76) rated these ICIs as more threatening in the threat compared to the trust condition, validating the causal effect of prior beliefs on face perception. Contrary to expectations derived from predictive processing accounts, there was no evidence for a main effect of paranoia. This finding suggests that paranoia was not related to stronger threat prior beliefs that directly affected face perception, challenging the assumption that paranoid beliefs operate on a perceptual level.

Tug-of-Peace: Visual Rivalry and Atypical Visual Motion Processing in MECP2 Duplication Syndrome of Autism

Extracting common patterns of neural circuit computations in the autism spectrum and confirming them as a cause of specific core traits of autism is the first step toward identifying cell-level and circuit-level targets for effective clinical intervention. Studies in humans with autism have identified functional links and common anatomic substrates between core restricted behavioral repertoire, cognitive rigidity, and overstability of visual percepts during visual rivalry. To study these processes with single-cell precision and comprehensive neuronal population coverage, we developed the visual bistable perception paradigm for mice based on ambiguous moving plaid patterns consisting of two transparent gratings drifting at an angle of 120°. This results in spontaneous reversals of the perception between local component motion (plaid perceived as two separate moving grating components) and integrated global pattern motion (plaid perceived as a fused moving texture). This robust paradigm does not depend on the explicit report of the mouse, since the direction of the optokinetic nystagmus (OKN) is used to infer the dominant percept. Using this paradigm, we found that the rate of perceptual reversals between global and local motion interpretations is reduced in the methyl-CpG-binding protein 2 duplication syndrome (MECP2-ds) mouse model of autism. Moreover, the stability of local motion percepts is greatly increased in MECP2-ds mice at the expense of global motion percepts. Thus, our model reproduces a subclass of the core features in human autism (reduced rate of visual rivalry and atypical perception of visual motion). This further offers a well-controlled approach for dissecting neuronal circuits underlying these core features.

Expectation Cues and False Percepts Generate Stimulus-Specific Activity in Distinct Layers of the Early Visual Cortex

Perception has been proposed to result from the integration of feedforward sensory signals with internally generated feedback signals. Feedback signals are believed to play an important role in driving false percepts, that is, seeing things that are not actually there. Feedforward and feedback influences on perception can be studied using layer-specific fMRI, which we used here to interrogate neural activity underlying high-confidence false percepts while healthy human participants (N = 25, male and female) performed a perceptual orientation discrimination task. Auditory cues implicitly signaled the most likely upcoming orientation (referred to here as expectations). These expectations induced orientation-specific templates in the deep and superficial layers of V2, without affecting perception. In contrast, the orientation of falsely perceived stimuli with high confidence was reflected in the middle input layers of V2, suggesting a feedforward signal contributing to false percepts. The prevalence of high-confidence false percepts was related to everyday hallucination severity in a separate online sample (N = 100), suggesting a possible link with abnormal perceptual experiences. These results reveal a potential feedforward mechanism underlying false percepts, reflected by spontaneous stimulus-like activity in the input layers of the visual cortex, independent of top-down signals reflecting cued orientations.SIGNIFICANCE STATEMENT False percepts have been suggested to arise through excessive feedback signals. However, feedforward contributions to false percepts have remained largely understudied. Laminar fMRI has been shown to be useful in distinguishing feedforward from feedback activity as it allows the imaging of different cortical layers. In the present study we demonstrate that although cued orientations are encoded in the feedback layers of the visual cortex, the content of the false percepts are encoded in the feedforward layers and did not rely on these cued orientations. This shows that false percepts can in principle emerge from random feedforward signals in the visual cortex, with possible implications for disorders hallmarked by hallucinations like schizophrenia and Parkinson's disease.

Bayesian accounts of perceptual decisions in the nonclinical continuum of psychosis: Greater imprecision in both top-down and bottom-up processes

Neurocomputational accounts of psychosis propose mechanisms for how information is integrated into a predictive model of the world, in attempts to understand the occurrence of altered perceptual experiences. Conflicting Bayesian theories postulate aberrations in either top-down or bottom-up processing. The top-down theory predicts an overreliance on prior beliefs or expectations resulting in aberrant perceptual experiences, whereas the bottom-up theory predicts an overreliance on current sensory information, as aberrant salience is directed towards objectively uninformative stimuli. This study empirically adjudicates between these models. We use a perceptual decision-making task in a neurotypical population with varying degrees of psychotic-like experiences. Bayesian modelling was used to compute individuals' reliance on prior relative to sensory information. Across two datasets (discovery dataset n = 363; independent replication in validation dataset n = 782) we showed that psychotic-like experiences were associated with an overweighting of sensory information relative to prior expectations, which seem to be driven by decreased precision afforded to prior information. However, when prior information was more uncertain, participants with greater psychotic-like experiences encoded sensory information with greater noise. Greater psychotic-like experiences were associated with aberrant precision in the encoding both prior and likelihood information, which we suggest may be related to generally heightened perceptions of task instability. Our study lends empirical support to notions of both weaker bottom-up and weaker (rather than stronger) top-down perceptual processes, as well as aberrancies in belief updating that extend into the non-clinical continuum of psychosis.

Spontaneous Necker-cube reversals may not be that spontaneous

Introduction

During observation of the ambiguous Necker cube, our perception suddenly reverses between two about equally possible 3D interpretations. During passive observation, perceptual reversals seem to be sudden and spontaneous. A number of theoretical approaches postulate destabilization of neural representations as a pre-condition for reversals of ambiguous figures. In the current study, we focused on possible Electroencephalogram (EEG) correlates of perceptual destabilization, that may allow prediction of an upcoming perceptual reversal.

Methods

We presented ambiguous Necker cube stimuli in an onset-paradigm and investigated the neural processes underlying endogenous reversals as compared to perceptual stability across two consecutive stimulus presentations. In a separate experimental condition, disambiguated cube variants were alternated randomly, to exogenously induce perceptual reversals. We compared the EEG immediately before and during endogenous Necker cube reversals with corresponding time windows during exogenously induced perceptual reversals of disambiguated cube variants.

Results

For the ambiguous Necker cube stimuli, we found the earliest differences in the EEG between reversal trials and stability trials already 1 s before a reversal occurred, at bilateral parietal electrodes. The traces remained similar until approximately 1100 ms before a perceived reversal, became maximally different at around 890 ms (p = 7.59 × 10^-6, Cohen's d = 1.35) and remained different until shortly before offset of the stimulus preceding the reversal. No such patterns were found in the case of disambiguated cube variants.

Discussion

The identified EEG effects may reflect destabilized states of neural representations, related to destabilized perceptual states preceding a perceptual reversal. They further indicate that spontaneous Necker cube reversals are most probably not as spontaneous as generally thought. Rather, the destabilization may occur over a longer time scale, at least 1 s before a reversal event, despite the reversal event as such being perceived as spontaneous by the viewer.

The Role of Shame in Schizophrenia Delusion: The Interplay Between Cognitive-Perceptual and Emotional Traits

ABSTRACT

Through a strictly dimensional approach, the present study aimed at evaluating the interplay between cognitive-perceptual disturbances and emotional dispositions, particularly shame proneness, in schizophrenia delusion. One hundred one outpatients with schizophrenia were administered the Peters et al. Delusions Inventory, the Referential Thinking Scale (REF), the Magical Ideation Scale (MIS), the Perceptual Aberration Scale (PAS), the Positive and Negative Affect Schedule and the Experiences of Shame Scale (ESS). The severity of delusional ideation was positively related to all the cognitive-perceptual scales (REF, MIS, and PAS) and to shame proneness (ESS). Referential thinking (REF) emerged as the strongest predictor of delusion severity. The experience of shame played a mediation role in the relationship between cognitive-perceptual traits and delusional severity. These data suggest that severity delusion in schizophrenia depends, at least in part, on a complex interplay between cognitive-perceptual disturbances and experiences of shame.

The role of the salience network in cognitive and affective deficits

Analysis and interpretation of studies on cognitive and affective dysregulation often draw upon the network paradigm, especially the Triple Network Model, which consists of the default mode network (DMN), the frontoparietal network (FPN), and the salience network (SN). DMN activity is primarily dominant during cognitive leisure and self-monitoring processes. The FPN peaks during task involvement and cognitive exertion. Meanwhile, the SN serves as a dynamic "switch" between the DMN and FPN, in line with salience and cognitive demand. In the cognitive and affective domains, dysfunctions involving SN activity are connected to a broad spectrum of deficits and maladaptive behavioral patterns in a variety of clinical disorders, such as depression, insomnia, narcissism, PTSD (in the case of SN hyperactivity), chronic pain, and anxiety, high degrees of neuroticism, schizophrenia, epilepsy, autism, and neurodegenerative illnesses, bipolar disorder (in the case of SN hypoactivity). We discuss behavioral and neurological data from various research domains and present an integrated perspective indicating that these conditions can be associated with a widespread disruption in predictive coding at multiple hierarchical levels. We delineate the fundamental ideas of the brain network paradigm and contrast them with the conventional modular method in the first section of this article. Following this, we outline the interaction model of the key functional brain networks and highlight recent studies coupling SN-related dysfunctions with cognitive and affective impairments.

A Computational Analysis of Abnormal Belief Updating Processes and Their Association With Psychotic Experiences and Childhood Trauma in a UK Birth Cohort

BACKGROUND

Psychotic experiences emerge from abnormalities in perception and belief formation and occur more commonly in those experiencing childhood trauma. However, which precise aspects of belief formation are atypical in psychosis is not well understood. We used a computational modeling approach to characterize belief updating in young adults in the general population, examine their relationship with psychotic outcomes and trauma, and determine the extent to which they mediate the trauma-psychosis relationship.

METHODS

We used data from 3360 individuals from the Avon Longitudinal Study of Parents and Children birth cohort who completed assessments for psychotic outcomes, depression, anxiety, and two belief updating tasks at age 24 and had data available on traumatic events assessed from birth to late adolescence. Unadjusted and adjusted regression and counterfactual mediation methods were used for the analyses.

RESULTS

Basic behavioral measures of belief updating (draws-to-decision and disconfirmatory updating) were not associated with psychotic experiences. However, computational modeling revealed an association between increased decision noise with both psychotic experiences and trauma exposure, although <3% of the trauma-psychotic experience association was mediated by decision noise. Belief updating measures were also associated with intelligence and sociodemographic characteristics, confounding most of the associations with psychotic experiences. There was little evidence that belief updating parameters were differentially associated with delusions compared with hallucinations or that they were differentially associated with psychotic outcomes compared with depression or anxiety.

CONCLUSIONS

These findings challenge the hypothesis that atypical belief updating mechanisms (as indexed by the computational models and behavioral measures we used) underlie the development of psychotic phenomena.

Perceptual inference, accuracy, and precision in temporal reproduction in schizophrenia

Accumulating evidence suggests that deficits in perceptual inference account for symptoms of schizophrenia. One manifestation of perceptual inference is the central bias, i.e., the tendency to put emphasis on prior experiences over actual events in perceiving incoming sensory stimuli. Using an interval reproduction task, this study aimed to determine whether patients with schizophrenia show a stronger central bias than participants without schizophrenia. In the interval reproduction task, participants were shown a cross on a screen. The cross was replaced with a Gaussian patch for a predetermined time interval, and participants were required to reproduce the interval duration by pressing and releasing the space key. We manipulated the uncertainty of prior information using different interval distributions. We found no difference in the influence of prior information on interval reproduction between patients and controls. However, patients with SZ showed a stronger central bias than healthy participants in the intermediate interval range (approximately 450 ms to 900 ms). It is possible that the patients in SZ have non-uniform deficits associated with interval range or uncertainty of prior information in perceptual inference. Further, the severity of avolition and alogia was correlated with the strength of central bias in SZ. This study provides some insights into the mechanisms underlying the association between schizophrenic symptoms and perceptual inference.

Top-down resolution of visual ambiguity - knowledge from the future or footprints from the past?

Current theories about visual perception assume that our perceptual system weights the a priori incomplete, noisy and ambiguous sensory information with previous, memorized perceptual experiences in order to construct stable and reliable percepts. These theories are supported by numerous experimental findings. Theories about precognition have an opposite point of view. They assume that information from the future can have influence on perception, thoughts, and behavior. Several experimental studies provide evidence for precognition effects, other studies found no such effects. One problem may be that the vast majority of precognition paradigms did not systematically control for potential effects from the perceptual history. In the present study, we presented ambiguous Necker cube stimuli and disambiguated cube variants and systematically tested in two separate experiments whether perception of a currently observed ambiguous Necker cube stimulus can be influenced by a disambiguated cube variant, presented in the immediate perceptual past (perceptual history effects) and/or in the immediate perceptual future (precognition effects). We found perceptual history effects, which partly depended on the length of the perceptual history trace but were independent of the perceptual future. Results from some individual participants suggest on the first glance a precognition pattern, but results from our second experiment make a perceptual history explanation more probable. On the group level, no precognition effects were statistically indicated. The perceptual history effects found in the present study are in confirmation with related studies from the literature. The precognition analysis revealed some interesting individual patterns, which however did not allow for general conclusions. Overall, the present study demonstrates that any future experiment about sensory or extrasensory perception urgently needs to control for potential perceptual history effects and that temporal aspects of stimulus presentation are of high relevance.

Cross-modal auditory priors drive the perception of bistable visual stimuli with reliable differences between individuals

It is a widely held assumption that the brain performs perceptual inference by combining sensory information with prior expectations, weighted by their uncertainty. A distinction can be made between higher- and lower-level priors, which can be manipulated with associative learning and sensory priming, respectively. Here, we simultaneously investigate priming and the differential effect of auditory vs. visual associative cues on visual perception, and we also examine the reliability of individual differences. Healthy individuals (N = 29) performed a perceptual inference task twice with a one-week delay. They reported the perceived direction of motion of dot pairs, which were preceded by a probabilistic visuo-acoustic cue. In 30% of the trials, motion direction was ambiguous, and in half of these trials, the auditory versus the visual cue predicted opposing directions. Cue-stimulus contingency could change every 40 trials. On ambiguous trials where the visual and the auditory cue predicted conflicting directions of motion, participants made more decisions consistent with the prediction of the acoustic cue. Increased predictive processing under stimulus uncertainty was indicated by slower responses to ambiguous (vs. non-ambiguous) stimuli. Furthermore, priming effects were also observed in that perception of ambiguous stimuli was influenced by perceptual decisions on the previous ambiguous and unambiguous trials as well. Critically, behavioural effects had substantial inter-individual variability which showed high test-retest reliability (intraclass correlation coefficient (ICC) > 0.78). Overall, higher-level priors based on auditory (vs. visual) information had greater influence on visual perception, and lower-level priors were also in action. Importantly, we observed large and stable differences in various aspects of task performance. Computational modelling combined with neuroimaging could allow testing hypotheses regarding the potential mechanisms causing these behavioral effects. The reliability of the behavioural differences implicates that such perceptual inference tasks could be valuable tools during large-scale biomarker and neuroimaging studies.

Unification by Fiat: Arrested Development of Predictive Processing

Predictive processing (PP) has been repeatedly presented as a unificatory account of perception, action, and cognition. In this paper, we argue that this is premature: As a unifying theory, PP fails to deliver general, simple, homogeneous, and systematic explanations. By examining its current trajectory of development, we conclude that PP remains only loosely connected both to its computational framework and to its hypothetical biological underpinnings, which makes its fundamentals unclear. Instead of offering explanations that refer to the same set of principles, we observe systematic equivocations in PP‐based models, or outright contradictions with its avowed principles. To make matters worse, PP‐based models are seldom empirically validated, and they are frequently offered as mere just‐so stories. The large number of PP‐based models is thus not evidence of theoretical progress in unifying perception, action, and cognition. On the contrary, we maintain that the gap between theory and its biological and computational bases contributes to the arrested development of PP as a unificatory theory. Thus, we urge the defenders of PP to focus on its critical problems instead of offering mere re‐descriptions of known phenomena, and to validate their models against possible alternative explanations that stem from different theoretical assumptions. Otherwise, PP will ultimately fail as a unified theory of cognition.

Disentangling predictive processing in the brain: a meta-analytic study in favour of a predictive network

According to the predictive coding (PC) theory, the brain is constantly engaged in predicting its upcoming states and refining these predictions through error signals. Despite extensive research investigating the neural bases of this theory, to date no previous study has systematically attempted to define the neural mechanisms of predictive coding across studies and sensory channels, focussing on functional connectivity. In this study, we employ a coordinate-based meta-analytical approach to address this issue. We first use the Activation Likelihood Estimation (ALE) algorithm to detect spatial convergence across studies, related to prediction error and encoding. Overall, our ALE results suggest the ultimate role of the left inferior frontal gyrus and left insula in both processes. Moreover, we employ a meta-analytic connectivity method (Seed-Voxel Correlations Consensus). This technique reveals a large, bilateral predictive network, which resembles large-scale networks involved in task-driven attention and execution. In sum, we find that: (i) predictive processing seems to occur more in certain brain regions than others, when considering different sensory modalities at a time; (ii) there is no evidence, at the network level, for a distinction between error and prediction processing.

Retinal ganglion cell dysfunction is correlated with disturbed visual cognition in schizophrenia patients with visual hallucinations

Patients with schizophrenia have altered visual cognition and retinal functions. No studies have explored if retinal anomalies are related to visual cognition and the presence of visual hallucinations (VH). We explored functional responses of the retinal ganglion cells in schizophrenia patients with or without VH and conducted a neuropsychological evaluation to explore the links between cognition and retinal function. The VH+ group showed poorer visual cognition and we found correlations between the amplitudes of the P50 and the N95 waves and visual cognition. Our results provide arguments for a link between retinal dysfunction, impaired visual processing and VH in schizophrenia.

Illusions, Delusions, and Your Backwards Bayesian Brain: A Biased Visual Perspective

The retinal image is insufficient for determining what is "out there," because many different real-world geometries could produce any given retinal image. Thus, the visual system must infer which external cause is most likely, given both the sensory data and prior knowledge that is either innate or learned via interactions with the environment. We will describe a general framework of "hierarchical Bayesian inference" that we and others have used to explore the role of cortico-cortical feedback in the visual system, and we will further argue that this approach to "seeing" makes our visual systems prone to perceptual errors in a variety of different ways. In this deliberately provocative and biased perspective, we argue that the neuromodulator, dopamine, may be a crucial link between neural circuits performing Bayesian inference and the perceptual idiosyncrasies of people with schizophrenia.

Advances in the computational understanding of mental illness

Computational psychiatry is a rapidly growing field attempting to translate advances in computational neuroscience and machine learning into improved outcomes for patients suffering from mental illness. It encompasses both data-driven and theory-driven efforts. Here, recent advances in theory-driven work are reviewed. We argue that the brain is a computational organ. As such, an understanding of the illnesses arising from it will require a computational framework. The review divides work up into three theoretical approaches that have deep mathematical connections: dynamical systems, Bayesian inference and reinforcement learning. We discuss both general and specific challenges for the field, and suggest ways forward.

A functional theory of bistable perception based on dynamical circular inference

When we face ambiguous images, the brain cannot commit to a single percept; instead, it switches between mutually exclusive interpretations every few seconds, a phenomenon known as bistable perception. While neuromechanistic models, e.g., adapting neural populations with lateral inhibition, may account for the dynamics of bistability, a larger question remains unresolved: how this phenomenon informs us on generic perceptual processes in less artificial contexts. Here, we propose that bistable perception is due to our prior beliefs being reverberated in the cortical hierarchy and corrupting the sensory evidence, a phenomenon known as “circular inference”. Such circularity could occur in a hierarchical brain where sensory responses trigger activity in higher-level areas but are also modulated by feedback projections from these same areas. We show that in the face of ambiguous sensory stimuli, circular inference can change the dynamics of the perceptual system and turn what should be an integrator of inputs into a bistable attractor switching between two highly trusted interpretations. The model captures various aspects of bistability, including Levelt’s laws and the stabilizing effects of intermittent presentation of the stimulus. Since it is related to the generic perceptual inference and belief updating mechanisms, this approach can be used to predict the tendency of individuals to form aberrant beliefs from their bistable perception behavior. Overall, we suggest that feedforward/feedback information loops in hierarchical neural networks, a phenomenon that could lead to psychotic symptoms when overly strong, could also underlie perception in nonclinical populations.

In cases of high ambiguity, our perceptual system cannot commit to a single percept and switches between different interpretations, giving rise to bistable perception. In this paper we outline a computational model of bistability based on the notion of circular inference, i.e. a form of suboptimal hierarchical inference in which priors and / or sensory inputs are reverberated and over-counted. We suggest that descending loops (i.e. reverberated priors) transform our perceptual system from a simple accumulator of sensory inputs into a bistable attractor, that switches between two highly-trusted interpretations. Using analytical methods we derive the necessary conditions for bistable perception to occur. We show that our dynamical circular inference model is able to capture many features of bistability, such as Levelt’s laws and the stabilizing effects of intermittent presentation of the stimulus. Finally we make novel predictions about the behavior of psychotic patients.

Psychotic Experiences in Schizophrenia and Sensitivity to Sensory Evidence

Perceptual inference depends on an optimal integration of current sensory evidence with prior beliefs about the environment. Alterations of this process have been related to the emergence of positive symptoms in schizophrenia. However, it has remained unclear whether delusions and hallucinations arise from an increased or decreased weighting of prior beliefs relative to sensory evidence. To investigate the relation of this prior-to-likelihood ratio to positive symptoms in schizophrenia, we devised a novel experimental paradigm which gradually manipulates perceptually ambiguous visual stimuli by disambiguating stimulus information. As a proxy for likelihood precision, we assessed the sensitivity of individual participants to sensory evidence. As a surrogate for the precision of prior beliefs in perceptual stability, we measured phase duration in ambiguity. Relative to healthy controls, patients with schizophrenia showed a stronger increment in congruent perceptual states for increasing levels of disambiguating stimulus evidence. Sensitivity to sensory evidence correlated positively with the individual patients' severity of perceptual anomalies and hallucinations. Moreover, the severity of such experiences correlated negatively with phase duration. Our results indicate that perceptual anomalies and hallucinations are associated with a shift of perceptual inference toward sensory evidence and away from prior beliefs. This reduced prior-to-likelihood ratio in sensory processing may contribute to the phenomenon of aberrant salience, which has been suggested to give rise to the false inferences underlying psychotic experiences.

Prospection does not imply predictive processing

Predictive processing models of psychopathologies are not explanatorily consistent with the present account of abstract thought. These models are based on latent variables probabilistically mapping the structure of the world. As such, they cannot be informed by representational ontology based on mental objects and states. What actually is the case is merely some terminological affinity between subjective and informational uncertainty.

Scene Representations Conveyed by Cortical Feedback to Early Visual Cortex Can Be Described by Line Drawings

Human behavior is dependent on the ability of neuronal circuits to predict the outside world. Neuronal circuits in early visual areas make these predictions based on internal models that are delivered via non-feedforward connections. Despite our extensive knowledge of the feedforward sensory features that drive cortical neurons, we have a limited grasp on the structure of the brain's internal models. Progress in neuroscience therefore depends on our ability to replicate the models that the brain creates internally. Here we record human fMRI data while presenting partially occluded visual scenes. Visual occlusion allows us to experimentally control sensory input to subregions of visual cortex while internal models continue to influence activity in these regions. Because the observed activity is dependent on internal models, but not on sensory input, we have the opportunity to map visual features conveyed by the brain's internal models. Our results show that activity related to internal models in early visual cortex are more related to scene-specific features than to categorical or depth features. We further demonstrate that behavioral line drawings provide a good description of internal model structure representing scene-specific features. These findings extend our understanding of internal models, showing that line drawings provide a window into our brains' internal models of vision.

SIGNIFICANCE STATEMENT We find that fMRI activity patterns corresponding to occluded visual information in early visual cortex fill in scene-specific features. Line drawings of the missing scene information correlate with our recorded activity patterns, and thus to internal models. Despite our extensive knowledge of the sensory features that drive cortical neurons, we have a limited grasp on the structure of our brains' internal models. These results therefore constitute an advance to the field of neuroscience by extending our knowledge about the models that our brains construct to efficiently represent and predict the world. Moreover, they link a behavioral measure to these internal models, which play an active role in many components of human behavior, including visual predictions, action planning, and decision making.

Towards a Unifying Cognitive, Neurophysiological, and Computational Neuroscience Account of Schizophrenia

Psychotic experiences may be understood as altered information processing due to aberrant neural computations. A prominent example of such neural computations is the computation of prediction errors (PEs), which signal the difference between expected and experienced events. Among other areas showing PE coding, hippocampal-prefrontal-striatal neurocircuits play a prominent role in information processing. Dysregulation of dopaminergic signaling, often secondary to psychosocial stress, is thought to interfere with the processing of biologically important events (such as reward prediction errors) and result in the aberrant attribution of salience to irrelevant sensory stimuli and internal representations. Bayesian hierarchical predictive coding offers a promising framework for the identification of dysfunctional neurocomputational processes and the development of a mechanistic understanding of psychotic experience. According to this framework, mismatches between prior beliefs encoded at higher levels of the cortical hierarchy and lower-level (sensory) information can also be thought of as PEs, with important consequences for belief updating. Low levels of precision in the representation of prior beliefs relative to sensory data, as well as dysfunctional interactions between prior beliefs and sensory data in an ever-changing environment, have been suggested as a general mechanism underlying psychotic experiences. Translating the promise of the Bayesian hierarchical predictive coding into patient benefit will come from integrating this framework with existing knowledge of the etiology and pathophysiology of psychosis, especially regarding hippocampal-prefrontal-striatal network function and neural mechanisms of information processing and belief updating.

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.

Intrinsic Rivalry. Can White Bears Help Us With the Other Side of Consciousness?

Studies of consciousness have traditionally been based mainly upon the perceptual domains of consciousness. However, there is another side of consciousness, represented by various types of intrinsic conscious experiences. Even though intrinsic experiences can represent up to 50% of our conscious experiences, they are still largely neglected in conscious studies. We assume there are two reasons for this. First, the field of intrinsic conscious experiences is methodologically far more problematic than any other. Second, specific paradigms for capturing the correlates of intrinsic conscious experiences are almost nonexistent. Nevertheless, we expect the intrinsic side of consciousness to soon take its place in conscious studies, but first new experimental paradigms will have to be devised, which would be of a similar design to the paradigms used in studies of perceptual consciousness. In this hypothesis and theory article, we propose such a hypothetical paradigm, presenting the exploratory data of our proof-of-concept study, discussing its use, and addressing its shortcomings and their possible remediation.

Parsing components of auditory predictive coding in schizophrenia using a roving standard mismatch negativity paradigm

BACKGROUND

Mismatch negativity (MMN) is an event-related potential (ERP) component reflecting auditory predictive coding. Repeated standard tones evoke increasing positivity ('repetition positivity'; RP), reflecting strengthening of the standard's memory trace and the prediction it will recur. Likewise, deviant tones preceded by more standard repetitions evoke greater negativity ('deviant negativity'; DN), reflecting stronger prediction error signaling. These memory trace effects are also evident in MMN difference wave. Here, we assess group differences and test-retest reliability of these indices in schizophrenia patients (SZ) and healthy controls (HC).

METHODS

Electroencephalography was recorded twice, 2 weeks apart, from 43 SZ and 30 HC, during a roving standard paradigm. We examined ERPs to the third, eighth, and 33rd standards (RP), immediately subsequent deviants (DN), and the corresponding MMN. Memory trace effects were assessed by comparing amplitudes associated with the three standard repetition trains.

RESULTS

Compared with controls, SZ showed reduced MMNs and DNs, but normal RPs. Both groups showed memory trace effects for RP, MMN, and DN, with a trend for attenuated DNs in SZ. Intraclass correlations obtained via this paradigm indicated good-to-moderate reliabilities for overall MMN, DN and RP, but moderate to poor reliabilities for components associated with short, intermediate, and long standard trains, and poor reliability of their memory trace effects.

CONCLUSION

MMN deficits in SZ reflected attenuated prediction error signaling (DN), with relatively intact predictive code formation (RP) and memory trace effects. This roving standard MMN paradigm requires additional development/validation to obtain suitable levels of reliability for use in clinical trials.

Hallucinations and Strong Priors

Hallucinations, perceptions in the absence of objectively identifiable stimuli, illustrate the constructive nature of perception. Here, we highlight the role of prior beliefs as a critical elicitor of hallucinations. Recent empirical work from independent laboratories shows strong, overly precise priors can engender hallucinations in healthy subjects and that individuals who hallucinate in the real world are more susceptible to these laboratory phenomena. We consider these observations in light of work demonstrating apparently weak, or imprecise, priors in psychosis. Appreciating the interactions within and between hierarchies of inference can reconcile this apparent disconnect. Data from neural networks, human behavior, and neuroimaging support this contention. This work underlines the continuum from normal to aberrant perception, encouraging a more empathic approach to clinical hallucinations.

Models of persecutory delusions: a mechanistic insight into the early stages of psychosis

Identifying robust markers for predicting the onset of psychosis has been a key challenge for early detection research. Persecutory delusions are core symptoms of psychosis, and social cognition is particularly impaired in first-episode psychosis patients and individuals at risk for developing psychosis. Here, we propose new avenues for translation provided by hierarchical Bayesian models of behaviour and neuroimaging data applied in the context of social learning to target persecutory delusions. As it comprises a mechanistic model embedded in neurophysiology, the findings of this approach may shed light onto inference and neurobiological causes of transition to psychosis.

Delusion Proneness is Linked to a Reduced Usage of Prior Beliefs in Perceptual Decisions

Predictive coding theories state an aberrant weighting of prior beliefs and present sensory information as a core computational pathology in psychosis. Specifically, it has been proposed that the influence of prior beliefs which attenuate improbable sensory information is weakened, resulting in an overweighing of this potentially misleading information. However, it is currently unclear whether this alteration is specific to perceptual processes or whether it represents a more pervasive deficit that extends to cognitive processes. Here, we carried out 2 behavioral experiments that probed the usage of priors during perceptual and cognitive processes, respectively, in 123 healthy individuals with varying degrees of delusion proneness. In an audio-visual perceptual discrimination task, participants had to judge the global motion direction of random dot kinematograms. Prior beliefs were induced by auditory cues that probabilistically predicted the global motion direction of the dot kinematograms, allowing us to measure the impact of prior beliefs on perceptual decision making. A control experiment paralleled the design of the perceptual decision making task in the domain of cognitive decision making. By fitting the participants' responses with a probabilistic decision model, we quantified the impact of prior beliefs on participants' decisions in both tasks. With growing delusion proneness, we found a decreased impact of prior beliefs on perceptual but not on cognitive decision making. Our results show that delusion proneness is linked to a specifically reduced usage of prior beliefs in perceptual decisions, thereby empirically substantiating predictive coding theories of psychosis.

The Predictive Coding Account of Psychosis

Fueled by developments in computational neuroscience, there has been increasing interest in the underlying neurocomputational mechanisms of psychosis. One successful approach involves predictive coding and Bayesian inference. Here, inferences regarding the current state of the world are made by combining prior beliefs with incoming sensory signals. Mismatches between prior beliefs and incoming signals constitute prediction errors that drive new learning. Psychosis has been suggested to result from a decreased precision in the encoding of prior beliefs relative to the sensory data, thereby garnering maladaptive inferences. Here, we review the current evidence for aberrant predictive coding and discuss challenges for this canonical predictive coding account of psychosis. For example, hallucinations and delusions may relate to distinct alterations in predictive coding, despite their common co-occurrence. More broadly, some studies implicate weakened prior beliefs in psychosis, and others find stronger priors. These challenges might be answered with a more nuanced view of predictive coding. Different priors may be specified for different sensory modalities and their integration, and deficits in each modality need not be uniform. Furthermore, hierarchical organization may be critical. Altered processes at lower levels of a hierarchy need not be linearly related to processes at higher levels (and vice versa). Finally, canonical theories do not highlight active inference-the process through which the effects of our actions on our sensations are anticipated and minimized. It is possible that conflicting findings might be reconciled by considering these complexities, portending a framework for psychosis more equipped to deal with its many manifestations.

Intelligence and uncertainty: Implications of hierarchical predictive processing for the neuroscience of cognitive ability

Hierarchical predictive processing (PP) has recently emerged as a candidate theoretical paradigm for neurobehavioral research. To date, PP has found support through its success in offering compelling explanations for a number of perceptual, cognitive, and psychiatric phenomena, as well as from accumulating neurophysiological evidence. However, its implications for understanding intelligence and its neural basis have received relatively little attention. The present review outlines the key tenets and evidence for PP, and assesses its implications for intelligence research. It is argued that PP suggests indeterminacy as a unifying principle from which to investigate the cognitive hierarchy and brain-ability correlations. The resulting framework not only accommodates prominent psychometric models of intelligence, but also incorporates key findings from neuroanatomical and functional activation research, and motivates new predictions via the mechanisms of prediction-error minimization. Because PP also suggests unique neural signatures of experience-dependent activity, it may also help clarify environmental contributions to intellectual development. It is concluded that PP represents a plausible, integrative framework that could enhance progress in the neuroscience of intelligence.

Autistic traits, but not schizotypy, predict increased weighting of sensory information in Bayesian visual integration

Recent theories propose that schizophrenia/schizotypy and autistic spectrum disorder are related to impairments in Bayesian inference that is, how the brain integrates sensory information (likelihoods) with prior knowledge. However existing accounts fail to clarify: (i) how proposed theories differ in accounts of ASD vs. schizophrenia and (ii) whether the impairments result from weaker priors or enhanced likelihoods. Here, we directly address these issues by characterizing how 91 healthy participants, scored for autistic and schizotypal traits, implicitly learned and combined priors with sensory information. This was accomplished through a visual statistical learning paradigm designed to quantitatively assess variations in individuals' likelihoods and priors. The acquisition of the priors was found to be intact along both traits spectra. However, autistic traits were associated with more veridical perception and weaker influence of expectations. Bayesian modeling revealed that this was due, not to weaker prior expectations, but to more precise sensory representations.

A hierarchical stochastic model for bistable perception

Viewing of ambiguous stimuli can lead to bistable perception alternating between the possible percepts. During continuous presentation of ambiguous stimuli, percept changes occur as single events, whereas during intermittent presentation of ambiguous stimuli, percept changes occur at more or less regular intervals either as single events or bursts. Response patterns can be highly variable and have been reported to show systematic differences between patients with schizophrenia and healthy controls. Existing models of bistable perception often use detailed assumptions and large parameter sets which make parameter estimation challenging. Here we propose a parsimonious stochastic model that provides a link between empirical data analysis of the observed response patterns and detailed models of underlying neuronal processes. Firstly, we use a Hidden Markov Model (HMM) for the times between percept changes, which assumes one single state in continuous presentation and a stable and an unstable state in intermittent presentation. The HMM captures the observed differences between patients with schizophrenia and healthy controls, but remains descriptive. Therefore, we secondly propose a hierarchical Brownian model (HBM), which produces similar response patterns but also provides a relation to potential underlying mechanisms. The main idea is that neuronal activity is described as an activity difference between two competing neuronal populations reflected in Brownian motions with drift. This differential activity generates switching between the two conflicting percepts and between stable and unstable states with similar mechanisms on different neuronal levels. With only a small number of parameters, the HBM can be fitted closely to a high variety of response patterns and captures group differences between healthy controls and patients with schizophrenia. At the same time, it provides a link to mechanistic models of bistable perception, linking the group differences to potential underlying mechanisms.

Patients suffering from schizophrenia show specific cognitive deficits. One way to study these cognitive phenomena works with the presentation of ambiguous stimuli. During viewing of an ambiguous stimulus, perception alters spontaneously between different percepts. Percept changes occur as single events during continuous presentation, whereas during intermittent presentation, percept changes occur at regular intervals either as single events or bursts. Here we investigate perceptual responses to continuous and intermittent stimulation in healthy control subjects and patients with schizophrenia. Interestingly, the response patterns can be highly variable but show systematic group differences. We propose a model that connects these perceptual responses to underlying neuronal processes. The model mainly describes the activity difference between competing neuronal populations on different neuronal levels. In a hierarchical manner, the differential activity generates switching between the conflicting percepts as well as between states of higher and lower perceptual stability. By fitting the model directly to empirical responses, a high variety of patterns can be reproduced, and group differences between healthy controls and patients with schizophrenia can be captured. This helps to link the observed group differences to potential neuronal mechanisms, suggesting that patients with schizophrenia tend to spend more time in neuronal states of lower perceptual stability.

Psychotic Experiences and Overhasty Inferences Are Related to Maladaptive Learning

Theoretical accounts suggest that an alteration in the brain’s learning mechanisms might lead to overhasty inferences, resulting in psychotic symptoms. Here, we sought to elucidate the suggested link between maladaptive learning and psychosis. Ninety-eight healthy individuals with varying degrees of delusional ideation and hallucinatory experiences performed a probabilistic reasoning task that allowed us to quantify overhasty inferences. Replicating previous results, we found a relationship between psychotic experiences and overhasty inferences during probabilistic reasoning. Computational modelling revealed that the behavioral data was best explained by a novel computational learning model that formalizes the adaptiveness of learning by a non-linear distortion of prediction error processing, where an increased non-linearity implies a growing resilience against learning from surprising and thus unreliable information (large prediction errors). Most importantly, a decreased adaptiveness of learning predicted delusional ideation and hallucinatory experiences. Our current findings provide a formal description of the computational mechanisms underlying overhasty inferences, thereby empirically substantiating theories that link psychosis to maladaptive learning.

Predictive coding theories represent a unifying account of psychosis, stating that the central psychosis-related alteration affects the interplay between prior predictions and incoming information. Since every incoming information is imprecise and potentially allows for different interpretations, prior expectations achieve the enforcement of interpretations with a higher prior probability. Disturbances in this basic framework might let unlikely interpretations come into effect, resulting in proneness for delusions and hallucinations. Here, we contribute to these theories by devising a novel computational model for behavior in a reasoning task that quantifies the participants' readiness to draw inferences from very surprising information. We thereby demonstrate that precisely this increased learning from surprising and thus potentially spurious information, as opposed to non-specific alterations in the general learning speed, predispose healthy individuals for delusions and hallucinations. The present results hence speak for the hypothesis that hallucinations and delusions arise when noisy information is considered as precise and is thus not suppressed by opposing prior beliefs. In this sense, our findings also tie with recent neurophysiological models of psychosis that posit aberrations in modulatory neurotransmitters such as dopamine (or its interactions with GABAergic interneurons) as a correlate of perturbed computations of information precision in the cortex.

Enhanced predictive signalling in schizophrenia

Positive symptoms of schizophrenia such as delusions and hallucinations are thought to arise from an alteration in predictive mechanisms of the brain. Here, we empirically tested the hypothesis that schizophrenia is associated with an enhanced signalling of higher-level predictions that shape perception into conformity with acquired beliefs. Twenty-one patients with schizophrenia and twenty-eight healthy controls matched for age and gender took part in a functional magnetic resonance imaging (fMRI) experiment that assessed the effect of an experimental manipulation of cognitive beliefs on the perception of an ambiguous visual motion stimulus. At the behavioural level, there was a generally weaker effect of experimentally induced beliefs on perception in schizophrenia patients compared with controls, but a positive correlation between the effect of beliefs on perception and the severity of positive symptoms. At the neural level, belief-related connectivity between a region encoding beliefs in the orbitofrontal cortex (OFC) and a region encoding visual motion in the visual cortex (V5) was higher in patients compared with controls, indicating a stronger impact of cognitive beliefs on visual processing in schizophrenia. We suggest that schizophrenia might be associated with a generally weaker acquisition of externally generated beliefs and a compensatory increase in the effect of beliefs on sensory processing. Our current results are in line with the notion that enhanced signalling of higher-level predictions that shape perception into conformity with acquired beliefs might underlie positive symptoms in schizophrenia. Hum Brain Mapp 38:1767-1779, 2017. © 2017 Wiley Periodicals, Inc.

Thought Insertion as a Self-Disturbance: An Integration of Predictive Coding and Phenomenological Approaches

Current theories in the framework of hierarchical predictive coding propose that positive symptoms of schizophrenia, such as delusions and hallucinations, arise from an alteration in Bayesian inference, the term inference referring to a process by which learned predictions are used to infer probable causes of sensory data. However, for one particularly striking and frequent symptom of schizophrenia, thought insertion, no plausible account has been proposed in terms of the predictive-coding framework. Here we propose that thought insertion is due to an altered experience of thoughts as coming from “nowhere”, as is already indicated by the early 20th century phenomenological accounts by the early Heidelberg School of psychiatry. These accounts identified thought insertion as one of the self-disturbances (from German: “Ichstörungen”) of schizophrenia and used mescaline as a model-psychosis in healthy individuals to explore the possible mechanisms. The early Heidelberg School (Gruhle, Mayer-Gross, Beringer) first named and defined the self-disturbances, and proposed that thought insertion involves a disruption of the inner connectedness of thoughts and experiences, and a “becoming sensory” of those thoughts experienced as inserted. This account offers a novel way to integrate the phenomenology of thought insertion with the predictive coding framework. We argue that the altered experience of thoughts may be caused by a reduced precision of context-dependent predictions, relative to sensory precision. According to the principles of Bayesian inference, this reduced precision leads to increased prediction-error signals evoked by the neural activity that encodes thoughts. Thus, in analogy with the prediction-error related aberrant salience of external events that has been proposed previously, “internal” events such as thoughts (including volitions, emotions and memories) can also be associated with increased prediction-error signaling and are thus imbued with aberrant salience. We suggest that the individual’s attempt to explain the aberrant salience of thoughts results in their interpretation as being inserted by an alien agent, similarly to the emergence of delusions in response to the aberrant salience of sensory stimuli.

Perceptual Stability of the Lissajous Figure Is Modulated by the Speed of Illusory Rotation

Lissajous figures represent ambiguous structure-from-motion stimuli rotating in depth and have proven to be a versatile tool to explore the cognitive and neural mechanisms underlying bistable perception. They are generated by the intersection of two sinusoids with perpendicular axes and increasing phase-shift whose frequency determines the speed of illusory 3D rotation. Recently, we found that Lissajous figures of higher shifting frequencies elicited longer perceptual phase durations and tentatively proposed a “representational momentum” account. In this study, our aim was twofold. First, we aimed to gather more behavioral evidence related to the perceptual dynamics of the Lissajous figure by simultaneously varying its shifting frequency and size. Using a conventional analysis, we investigated the effects of our experimental manipulations on transition probability (i.e., the probability that the current percept will change at the next critical stimulus configuration). Second, we sought to test the impact of our experimental factors on the occurrence of transitions in bistable perception by means of a Bayesian approach that can be used to directly quantify the impact of contextual cues on perceptual stability. We thereby estimated the implicit prediction of perceptual stability and how it is modulated by experimental manipulations.