Evidence for impaired visual prediction error in schizophrenia

Predictive coding in neuropsychiatric disorders: A systematic transdiagnostic review

The predictive coding framework postulates that the human brain continuously generates predictions about the environment, maximizing successes and minimizing failures based on prior experiences and beliefs. This PRISMA-compliant systematic review aims to comprehensively and transdiagnostically examine the differences in predictive coding between individuals with neuropsychiatric disorders and healthy controls. We included 72 articles including case-control studies investigating predictive coding as the primary outcome and reporting behavioral, neuroimaging, or electrophysiological findings. Thirty-three studies investigated predictive coding in the schizophrenia spectrum, 33 in neurodevelopmental disorders, 5 in mood disorders, 4 in neurocognitive disorders, 1 in post-traumatic stress disorder, and 1 in substance use disorders. Oddball and oddball-like paradigms were most frequently used to quantify predictive coding performance. Evidence showed heterogeneous impairments in the predictive coding abilities of the brain across neuropsychiatric disorders, particularly in schizophrenia and autism. Patients within the schizophrenia spectrum showed a consistent pattern of impaired non-social predictive coding. Conversely, predictive coding deficits were more selective for social cues in the autism spectrum. Predictive coding impairments were correlated with clinical symptom severity. These findings underscore the potential utility of predictive coding as a framework for understanding cognitive dysfunctions in the neuropsychiatric population, even though more evidence is needed on underexplored conditions, also considering potential confounders such as medication use and sex/gender. The potential role of predictive coding as a determinant of treatment response may also be considered to tailor personalized interventions.

Automatic sensory change processing in adults with attention deficit and hyperactivity disorder: a visual mismatch negativity study

In addition to higher-order executive functions, underlying sensory processing ability is also thought to play an important role in Attention-Deficit/Hyperactivity Disorder (AD/HD). An event-related potential feature, the mismatch negativity, reflects the ability of automatic sensory change processing and may be correlated with AD/HD symptoms and executive functions. This study aims to investigate the characteristics of visual mismatch negativity (vMMN) in adults with AD/HD. Twenty eight adults with AD/HD and 31 healthy controls were included in this study. These two groups were matched in age, IQ and sex. In addition, both groups completed psychiatric evaluations, a visual ERP task used to elicit vMMN, and psychological measures about AD/HD symptoms and day-to-day executive functions. Compared to trols, the late vMMN (230-330 ms) was significantly reduced in the AD/HD group. Correlation analyses showed that late vMMN was correlated with executive functions but not AD/HD symptoms. However, further mediation analyses showed that different executive functions had mediated the relationships between late vMMN and AD/HD symptoms. Our findings indicate that the late vMMN, reflecting automatic sensory change processing ability, was impaired in adults with AD/HD. This impairment could have negative impact on AD/HD symptoms via affecting day-to-day executive functions.

Systematic review and meta-analysis of the visual mismatch negativity in schizophrenia

Mismatch negativity (MMN) is an event-related potential component automatically elicited by events that violate predictions based on prior events. To elicit this component, researchers use stimulus repetition to induce predictions, and the MMN is obtained by subtracting the brain response to rare or unpredicted stimuli from that of frequent stimuli. Under the Predictive Processing framework, one increasingly popular interpretation of the mismatch response postulates that MMN represents a prediction error. In this context, the reduced MMN amplitude to auditory stimuli has been considered a potential biomarker of Schizophrenia, representing a reduced prediction error and the inability to update the mental model of the world based on the sensory signals. It is unclear, however, whether this amplitude reduction is specific for auditory events or if the visual MMN reveals a similar pattern in schizophrenia spectrum disorder. This review and meta-analysis aimed to summarise the available literature on the vMMN in schizophrenia. A systematic literature search resulted in 10 eligible studies that resulted in a combined effect size of g = -.63, CI [-.86, -.41], reflecting lower vMMN amplitudes in patients. These results are in line with the findings in the auditory domain. This component offers certain advantages, such as less susceptibility to overlap with components generated by attentional demands. Future studies should use vMMN to explore abnormalities in the Predictive Processing framework in different stages and groups of the SSD and increase the knowledge in the search for biomarkers in schizophrenia.

Working memory and sensory memory in subclinical high schizotypy: An avenue for understanding schizophrenia?

The search for robust, reliable biomarkers of schizophrenia remains a high priority in psychiatry. Biomarkers are valuable because they can reveal the underlying mechanisms of symptoms and monitor treatment progress and may predict future risk of developing schizophrenia. Despite the existence of various promising biomarkers that relate to symptoms across the schizophrenia spectrum, and despite published recommendations encouraging multivariate metrics, they are rarely investigated simultaneously within the same individuals. In those with schizophrenia, the magnitude of purported biomarkers is complicated by comorbid diagnoses, medications and other treatments. Here, we argue three points. First, we reiterate the importance of assessing multiple biomarkers simultaneously. Second, we argue that investigating biomarkers in those with schizophrenia-related traits (schizotypy) in the general population can accelerate progress in understanding the mechanisms of schizophrenia. We focus on biomarkers of sensory and working memory in schizophrenia and their smaller effects in individuals with nonclinical schizotypy. Third, we note irregularities across research domains leading to the current situation in which there is a preponderance of data on auditory sensory memory and visual working memory, but markedly less in visual (iconic) memory and auditory working memory, particularly when focusing on schizotypy where data are either scarce or inconsistent. Together, this review highlights opportunities for researchers without access to clinical populations to address gaps in knowledge. We conclude by highlighting the theory that early sensory memory deficits contribute negatively to working memory and vice versa. This presents a mechanistic perspective where biomarkers may interact with one another and impact schizophrenia-related symptoms.

Perceptual Functioning

Perceptual disorders are not part of the diagnosis criteria for schizophrenia. Yet, a considerable amount of work has been conducted, especially on visual perception abnormalities, and there is little doubt that visual perception is altered in patients. There are several reasons why such perturbations are of interest in this pathology. They are observed during the prodromal phase of psychosis, they are related to the pathophysiology (clinical disorganization, disorders of the sense of self), and they are associated with neuronal connectivity disorders. Perturbations occur at different levels of processing and likely affect how patients interact and adapt to their surroundings. The literature has become very large, and here we try to summarize different models that have guided the exploration of perception in patients. We also illustrate several lines of research by showing how perception has been investigated and by discussing the interpretation of the results. In addition to discussing domains such as contrast sensitivity, masking, and visual grouping, we develop more recent fields like processing at the level of the retina, and the timing of perception.

A neural mechanism underlying predictive visual motion processing in patients with schizophrenia

Psychotic symptoms may be traced back to sensory sensitivity. Thereby, visual motion (VM) processing particularly has been suggested to be impaired in schizophrenia (SCZ). In healthy brains, VM underlies predictive processing within hierarchically structured systems. However, less is known about predictive VM processing in SCZ. Therefore, we performed fMRI during a VM paradigm with three conditions of varying predictability, i.e., Predictable-, Random-, and Arbitrary motion. The study sample comprised 17 SCZ patients and 23 healthy controls. We calculated general linear model (GLM) analysis to assess group differences in VM processing across motion conditions. Here, we identified significantly lower activity in right temporoparietal junction (TPJ) for SCZ patients. Therefore, right TPJ was set as seed for connectivity analyses. For patients, across conditions we identified increased connections to higher regions, namely medial prefrontal cortex, or paracingulate gyrus. Healthy subjects activated sensory regions as area V5, or superior parietal lobule. Reduced TPJ activity may reflect both a failure in the bottom-up flow of visual information and a decrease of signal processing as consequence of increased top-down input from frontal areas. In sum, these altered neural patterns provide a framework for future studies focusing on predictive VM processing to identify potential biomarkers of psychosis.

The Relationship Between Affective Visual Mismatch Negativity and Interpersonal Difficulties Across Autism and Schizotypal Traits

Sensory deficits are a feature of autism and schizophrenia, as well as the upper end of their non-clinical spectra. The mismatch negativity (MMN), an index of pre-attentive auditory processing, is particularly sensitive in detecting such deficits; however, little is known about the relationship between the visual MMN (vMMN) to facial emotions and autism and schizophrenia spectrum symptom domains. We probed the vMMN to happy, sad, and neutral faces in 61 healthy adults (18-40 years, 32 female), and evaluated their degree of autism and schizophrenia spectrum traits using the Autism Spectrum Quotient (AQ) and Schizotypal Personality Questionnaire (SPQ). The vMMN to happy faces was significantly larger than the vMMNs to sad and neutral faces. The vMMN to happy faces was associated with interpersonal difficulties as indexed by AQ Communication and Attention to Detail subscales, and SPQ associated with more interpersonal difficulties. These data suggest that pre-attentive processing of positive affect might be more specific to the interpersonal features associated with autism and schizophrenia. These findings add valuable insights into the growing body of literature investigating symptom-specific neurobiological markers of autism and schizophrenia spectrum conditions.

Predictive waves in the autism-schizophrenia continuum: A novel biobehavioral model

The brain is a predictive machine. Converging data suggests a diametric predictive strategy from autism spectrum disorders (ASD) to schizophrenic spectrum disorders (SSD). Whereas perceptual inference in ASD is rigidly shaped by incoming sensory information, the SSD population is prone to overestimate the precision of their priors' models. Growing evidence considers brain oscillations pivotal biomarkers to understand how top-down predictions integrate bottom-up input. Starting from the conceptualization of ASD and SSD as oscillopathies, we introduce an integrated perspective that ascribes the maladjustments of the predictive mechanism to dysregulation of neural synchronization. According to this proposal, disturbances in the oscillatory profile do not allow the appropriate trade-off between descending predictive signal, overweighted in SSD, and ascending prediction errors, overweighted in ASD. These opposing imbalances both result in an ill-adapted reaction to external challenges. This approach offers a neuro-computational model capable of linking predictive coding theories with electrophysiological findings, aiming to increase knowledge on the neuronal foundations of the two spectra features and stimulate hypothesis-driven rehabilitation/research perspectives.

Auditory Mismatch Negativity Under Predictive Coding Framework and Its Role in Psychotic Disorders

Traditional neuroscience sees sensory perception as a simple feedforward process. This view is challenged by the predictive coding model in recent years due to the robust evidence researchers had found on how our prediction could influence perception. In the first half of this article, we reviewed the concept of predictive brain and some empirical evidence of sensory prediction in visual and auditory processing. The predictive function along the auditory pathway was mainly studied by mismatch negativity (MMN)-a brain response to an unexpected disruption of regularity. We summarized a range of MMN paradigms and discussed how they could contribute to the theoretical development of the predictive coding neural network by the mechanism of adaptation and deviance detection. Such methodological and conceptual evolution sharpen MMN as a tool to better understand the structural and functional brain abnormality for neuropsychiatric disorder such as schizophrenia.

Significant repetition probability effects in schizophrenia

BACKGROUND

A growing body of evidence suggests that the comparison of expected and incoming sensory stimuli (the prediction-error (ε) processing) is impaired in schizophrenia patients (SZ). For example, in studies of mismatch negativity, an ERP component that signals ε, SZ patients show deficits in the auditory and visual modalities. To test the role of impaired ε processing further in SZ, using neuroimaging methods, we applied a repetition-suppression (RS) paradigm.

METHODS

Patients diagnosed with SZ (n = 17) as well as age- and sex- matched healthy control subjects (HC, n = 17) were presented with pairs of faces, which could either repeat or alternate. Additionally, the likelihood of repetition/alternation trials was modulated in individual blocks of fMRI recordings, testing the effects of repetition probability (P(rep)) on RS.

RESULTS

We found a significant RS in the fusiform and occipital face areas as well as in the lateral occipital cortex that was similar in healthy controls and SZ patients SZ. More importantly, we observed similar P(rep) effects (larger RS in blocks with high frequency of repetitions than in blocks with low repetition likelihood) in both the control and the patient group.

CONCLUSION

Our findings suggest that repetition_probability modulations affect the neural responses in schizophrenia patients and healthy participants similarly. This suggests that the neural mechanisms determining perceptual inferences based on stimulus probabilities remain unimpaired in schizophrenia.

Emotional context restores cortical prediction error responses in schizophrenia

The mismatch negativity (MMN) deficit in schizophrenia is a consistently replicated finding and is considered a potential biomarker. From the cognitive neuroscience perspective, MMN represents a cortical correlate of the prediction error, a fundamental computational operator that may be at the core of various cognitive and clinical deficits observed in schizophrenia. The impact of emotion on cognitive processes in schizophrenia is insufficiently understood, and its impact on basic operators of cortical computation is largely unknown. In the visual domain, the facial expression mismatch negativity (EMMN) offers an opportunity to investigate basic computational operators in purely cognitive and in emotional contexts. In this study, we asked whether emotional context enhances cortical prediction error responses in patients with schizophrenia, as is the case in normal subjects. Therefore, seventeen patients with schizophrenia and eighteen controls completed a visual sequence oddball task, which allows for directly comparing MMN components evoked by deviants with high, intermediate and low emotional engagement. Interestingly, patients with schizophrenia showed pronounced deficits in response to neutral stimuli, but almost normal responses to emotional stimuli. The dissociation between impaired MMN and normal EMMN suggests that emotional context not only enhances, but restores cortical prediction error responses in patients with schizophrenia to near-normal levels. Our results show that emotional processing in schizophrenia is not necessarily defect; more likely, emotional processing heterogeneously impacts on cognition in schizophrenia. In fact, this study suggests that emotional context may even compensate for cognitive deficits in schizophrenia that are, in a different sensory domain, discussed as biomarkers.

Automatic detection advantage of network information among Internet addicts: behavioral and ERP evidence

Converging evidence has proved the attentional bias of Internet addicts (IAs) on network information. However, previous studies have neither explained how characteristics of network information are detected by IAs with priority nor proved whether this advantage is in line with the unconscious and automatic process. To answer the two questions, this study aims to investigate whether IAs prioritize automatic detection of network information from the behavior and cognitive neuroscience aspects. 15 severe IAs and 15 matching healthy controls were selected using Internet Addiction Test (IAT). Dot-probe task with mask was used in the behavioral experiment, while deviant-standard reverse oddball paradigm was used in the event-related potential (ERP) experiment to induce mismatch negativity (MMN). In the dot-probe task, when the probe location appeared on the Internet-related picture's position, the IAs had significantly shorter reaction time than do the controls; in the ERP experiment, when Internet-related picture appeared, MMN was significantly induced in the IAs relative to the controls. Both experiments show that IAs can automatically detect network information.

Impairment in Mismatch Negativity but not Repetition Suppression in Schizophrenia

Schizophrenia is characterized by impaired auditory-evoked potentials (AEPs), mismatch negativity (MMN), and sensory gating of AEPs to repeated stimuli (repetition suppression, RS). In the predictive modeling framework, MMN and RS reflect encoding of prediction error and model sharpening, respectively. We compared P50, N100, P200 RS, and pitch and duration MMN in 26 participants diagnosed with schizophrenia (SZ) and 26 matched healthy controls (HC), and assessed relationships between MMN, RS, and SZ diagnosis. RS was measured by comparing responses to individual tones presented as 5-tone groups (1 kHz, 75 dB, 50 ms, 5 ms rise/fall times, 330 ms SOA), separated by a 750 ms inter-trial interval. For MMN, the same tones were presented, with occasional pitch (1.2 kHz, 10%) or duration deviants (100 ms, 10%) interspersed. Pitch and duration MMN were reduced in SZ (p < 0.01). There were no group differences in P50 RS, N100 RS, or P200 RS (p's > 0.1). Importantly, although pitch and duration MMN both correlated with RS of AEPs within the MMN time range (p's < 0.01), SZ diagnosis predicted MMN over and above RS (p < 0.05) and shared little variance with RS in prediction of MMN amplitude (tolerance > 0.93). We suggest that reduced MMN in SZ is related to deficits in encoding prediction error but not repetition suppression.

Functional connectivity when detecting rare visual targets in schizophrenia

Individuals with schizophrenia demonstrate difficulties in attending to important stimuli (e.g., targets) and ignoring distractors (e.g., non-targets). We used a visual oddball task during fMRI to examine functional connectivity within and between the ventral and dorsal attention networks to determine the relative contribution of each network to detection of rare visual targets in schizophrenia. The sample comprised 25 schizophrenia patients and 27 healthy controls. Psychophysiological interaction analysis was used to examine whole-brain functional connectivity in response to targets. We used the right temporo parietal junction (TPJ) as the seed region for the ventral network and the right medial intraparietal sulcus (IPS) as the seed region for the dorsal network. We found that connectivity between right IPS and right anterior insula (AI; a component of the ventral network) was significantly greater in controls than patients. Expected patterns of within- and between-network connectivity for right TPJ were observed in controls, and not significantly different in patients. These findings indicate functional connectivity deficits between the dorsal and ventral attention networks in schizophrenia that may create problems in processing relevant versus irrelevant stimuli. Understanding the nature of network disruptions underlying cognitive deficits of schizophrenia may help shed light on the pathophysiology of this disorder.

Elementary sensory deficits in schizophrenia indexed by impaired visual mismatch negativity

INTRODUCTION

Mismatch negativity (MMN) is an automatic brain response to unexpected events. It represents a prediction error (PE) response, reflecting the difference between the sensory input and predictions. While deficits in auditory MMN are well known in schizophrenia, only few studies investigated impairments in predictive visual processing in schizophrenia. These studies used complex stimuli such as motion direction and emotional facial expressions. Here we studied whether automatic predictive processing of elementary features such as orientation is also impaired in schizophrenia.

METHODS

Altogether 28 patients with schizophrenia and 27 healthy controls matched in age, gender, and education participated in the study. EEG was recorded using 128 channels in the two experimental blocks. Using an oddball paradigm, horizontal stripes of Gabor patches were presented as frequent standards and vertical stripes as rare deviants in one block. Stimulus probabilities were swapped in the other block. Mismatch responses were obtained by subtracting responses to standard from those to deviant stimuli.

RESULTS

We found significant mismatch responses in healthy controls but not in patients in the prefrontal and occipital-parietal regions in the 90-200ms interval. Furthermore patients showed significantly decreased deviant minus standard difference waveforms relative to controls in the same regions with moderate to large effect sizes.

CONCLUSIONS

Our findings demonstrate that predictive processing of unattended low-level visual features such as orientation is impaired in schizophrenia. Our results complement reports of sensory deficits found in tasks requiring attentive processing and suggest that deficits are present in automatic visual sensory processes putatively mediated by glutamatergic functioning.

Emotional context facilitates cortical prediction error responses

In the predictive coding framework, mismatch negativity (MMN) is regarded a correlate of the prediction error that occurs when top-down predictions conflict with bottom-up sensory inputs. Expression-related MMN is a relatively novel construct thought to reflect a prediction error specific to emotional processing, which, however, has not yet been tested directly. Our paradigm includes both neutral and emotional deviants, thereby allowing for investigating whether expression-related MMN is emotion-specific or unspecifically arises from violations of a given sequence. Twenty healthy participants completed a visual sequence oddball task where they were presented with (1) sequence deviants, (2) emotional sequence deviants, and (3) emotional deviants. Mismatch components were assessed at ventral occipitotemporal scalp sites and analyzed regarding their amplitudes, spatiotemporal profiles, and neuronal sources. Expression-related MMN could be clearly separated from its neutral counterpart in all investigated aspects. Specifically, expression-related MMN showed enhanced amplitude, shorter latency, and different neuronal sources. Our results, therefore, provide converging evidence for a quantitative specificity of expression-related MMN and seems to provide an opportunity to study prediction error during preattentive emotional processing. Our neurophysiological evidence ultimately suggests that a basic cognitive operator, the prediction error, is enhanced at the cortical level by processing of emotionally salient stimuli.

Disruption of information processing in schizophrenia: The time perspective

We review studies suggesting time disorders on both automatic and subjective levels in patients with schizophrenia. Patients have difficulty explicitly discriminating between simultaneous and asynchronous events, and ordering events in time. We discuss the relationship between these difficulties and impairments on a more elementary level. We showed that for undetectable stimulus onset asynchronies below 20 ms, neither patients nor controls merge events in time, as previously believed. On the contrary, subjects implicitly distinguish between events even when evaluating them to be simultaneous. Furthermore, controls privilege the last stimulus, whereas patients seem to stay stuck on the first stimulus when asynchronies are sub-threshold. Combining previous results shows this to be true for patients even for asynchronies as short as 8 ms. Moreover, this peculiarity predicts difficulties with detecting asynchronies longer than 50 ms, suggesting an impact on the conscious ability to time events. Difficulties on the subjective level are also correlated with clinical disorganization. The results are interpreted within the framework of predictive coding which can account for an implicit ability to update events. These results complement a range of other results, by suggesting a difficulty with binding information in time as well as space, and by showing that information processing lacks continuity and stability in patients. The time perspective may help bridge the gap between cognitive impairments and clinical symptoms, by showing how the innermost structure of thought and experience is disrupted.

Auditory mismatch negativity and repetition suppression deficits in schizophrenia explained by irregular computation of prediction error

BACKGROUND

The predictive coding model is rapidly gaining attention in schizophrenia research. It posits the neuronal computation of residual variance ('prediction error') between sensory information and top-down expectation through multiple hierarchical levels. Event-related potentials (ERP) reflect cortical processing stages that are increasingly interpreted in the light of the predictive coding hypothesis. Both mismatch negativity (MMN) and repetition suppression (RS) measures are considered a prediction error correlates based on error detection and error minimization, respectively.

METHODS

Twenty-five schizophrenia patients and 25 healthy controls completed auditory tasks designed to elicit MMN and RS responses that were investigated using repeated measures models and strong spatio-temporal a priori hypothesis based on previous research. Separate correlations were performed for controls and schizophrenia patients, using age and clinical variables as covariates.

RESULTS

MMN and RS deficits were largely replicated in our sample of schizophrenia patients. Moreover, MMN and RS measures were strongly correlated in healthy controls, while no correlation was found in schizophrenia patients. Single-trial analyses indicated significantly lower signal-to-noise ratio during prediction error computation in schizophrenia.

CONCLUSIONS

This study provides evidence that auditory ERP components relevant for schizophrenia research can be reconciled in the light of the predictive coding framework. The lack of any correlation between the investigated measures in schizophrenia patients suggests a disruption of predictive coding mechanisms in general. More specifically, these results suggest that schizophrenia is associated with an irregular computation of residual variance between sensory input and top-down models, i.e. prediction error.