Cortical Layer-Dependent Signaling in Cognition: Three Computational Modes of the Canonical Circuit.

The cerebral cortex performs computations via numerous six-layer modules. The operational dynamics of these modules were studied primarily in early sensory cortices using bottom-up computation for response selectivity as a model, which has been recently revolutionized by genetic approaches in mice. However, cognitive processes such as recall and imagery require top-down generative computation. The question of whether the layered module operates similarly in top-down generative processing as in bottom-up sensory processing has become testable by advances in the layer identification of recorded neurons in behaving monkeys. This review examines recent advances in laminar signaling in these two computations, using predictive coding computation as a common reference, and shows that each of these computations recruits distinct laminar circuits, particularly in layer 5, depending on the cognitive demands. These findings highlight many open questions, including how different interareal feedback pathways, originating from and terminating at different layers, convey distinct functional signals.

Cortical-subcortical interactions underlie processing of auditory predictions measured with 7T fMRI.

Perception integrates both sensory inputs and internal models of the environment. In the auditory domain, predictions play a critical role because of the temporal nature of sounds. However, the precise contribution of cortical and subcortical structures in these processes and their interaction remain unclear. It is also unclear whether these brain interactions are specific to abstract rules or if they also underlie the predictive coding of local features. We used high-field 7T functional magnetic resonance imaging to investigate interactions between cortical and subcortical areas during auditory predictive processing. Volunteers listened to tone sequences in an oddball paradigm where the predictability of the deviant was manipulated. Perturbations in periodicity were also introduced to test the specificity of the response. Results indicate that both cortical and subcortical auditory structures encode high-order predictive dynamics, with the effect of predictability being strongest in the auditory cortex. These predictive dynamics were best explained by modeling a top-down information flow, in contrast to unpredicted responses. No error signals were observed to deviations of periodicity, suggesting that these responses are specific to abstract rule violations. Our results support the idea that the high-order predictive dynamics observed in subcortical areas propagate from the auditory cortex.

Collective predictive coding hypothesis: symbol emergence as decentralized Bayesian inference.

Understanding the emergence of symbol systems, especially language, requires the construction of a computational model that reproduces both the developmental learning process in everyday life and the evolutionary dynamics of symbol emergence throughout history. This study introduces the collective predictive coding (CPC) hypothesis, which emphasizes and models the interdependence between forming internal representations through physical interactions with the environment and sharing and utilizing meanings through social semiotic interactions within a symbol emergence system. The total system dynamics is theorized from the perspective of predictive coding. The hypothesis draws inspiration from computational studies grounded in probabilistic generative models and language games, including the Metropolis-Hastings naming game. Thus, playing such games among agents in a distributed manner can be interpreted as a decentralized Bayesian inference of representations shared by a multi-agent system. Moreover, this study explores the potential link between the CPC hypothesis and the free-energy principle, positing that symbol emergence adheres to the society-wide free-energy principle. Furthermore, this paper provides a new explanation for why large language models appear to possess knowledge about the world based on experience, even though they have neither sensory organs nor bodies. This paper reviews past approaches to symbol emergence systems, offers a comprehensive survey of related prior studies, and presents a discussion on CPC-based generalizations. Future challenges and potential cross-disciplinary research avenues are highlighted.

High segregation and diminished global integration in large-scale brain functional networks enhances the perceptual binding of cross-modal stimuli.

Speech perception requires the binding of spatiotemporally disjoint auditory-visual cues. The corresponding brain network-level information processing can be characterized by two complementary mechanisms: functional segregation which refers to the localization of processing in either isolated or distributed modules across the brain, and integration which pertains to cooperation among relevant functional modules. Here, we demonstrate using functional magnetic resonance imaging recordings that subjective perceptual experience of multisensory speech stimuli, real and illusory, are represented in differential states of segregation-integration. We controlled the inter-subject variability of illusory/cross-modal perception parametrically, by introducing temporal lags in the incongruent auditory-visual articulations of speech sounds within the McGurk paradigm. The states of segregation-integration balance were captured using two alternative computational approaches. First, the module responsible for cross-modal binding of sensory signals defined as the perceptual binding network (PBN) was identified using standardized parametric statistical approaches and their temporal correlations with all other brain areas were computed. With increasing illusory perception, the majority of the nodes of PBN showed decreased cooperation with the rest of the brain, reflecting states of high segregation but reduced global integration. Second, using graph theoretic measures, the altered patterns of segregation-integration were cross-validated.

Dysfunction of the magnocellular subdivision of the visual thalamus in developmental dyslexia.

Developmental dyslexia (DD) is one of the most common learning disorders, affecting millions of children and adults worldwide. To date, scientific research has attempted to explain DD primarily based on pathophysiological alterations in the cerebral cortex. In contrast, several decades ago, pioneering research on five post-mortem human brains suggested that a core characteristic of DD might be morphological alterations in a specific subdivision of the visual thalamus - the magnocellular LGN (M-LGN). However, due to considerable technical challenges in investigating LGN subdivisions non-invasively in humans, this finding was never confirmed in-vivo, and its relevance for DD pathology remained highly controversial. Here, we leveraged recent advances in high-resolution magnetic resonance imaging (MRI) at high field strength (7 Tesla) to investigate the M-LGN in DD in-vivo. Using a case-control design, we acquired data from a large sample of young adults with DD (n = 26; age 28 ± 7 years; 13 females) and matched control participants (n = 28; age 27 ± 6 years; 15 females). Each participant completed a comprehensive diagnostic behavioral test battery and participated in two MRI sessions, including three functional MRI experiments and one structural MRI acquisition. We measured blood-oxygen-level-dependent responses and longitudinal relaxation rates to compare both groups on LGN subdivision function and myelination. Based on previous research, we hypothesized that the M-LGN is altered in DD and that these alterations are associated with a key DD diagnostic score, i.e., rapid letter and number naming (RANln). The results showed aberrant responses of the M-LGN in DD compared to controls, which was reflected in a different functional lateralization of this subdivision between groups. These alterations were associated with RANln performance, specifically in male DD. We also found lateralization differences in the longitudinal relaxation rates of the M-LGN in DD relative to controls. Conversely, the other main subdivision of the LGN, the parvocellular LGN (P-LGN), showed comparable blood-oxygen-level-dependent responses and longitudinal relaxation rates between groups. The present study is the first to unequivocally show that M-LGN alterations are a hallmark of DD, affecting both the function and microstructure of this subdivision. It further provides a first functional interpretation of M-LGN alterations and a basis for a better understanding of sex-specific differences in DD with implications for prospective diagnostic and treatment strategies.

Associations of Cognitive Expectancies With Auditory Hallucinations and Hallucinatory-Like Experiences in Patients With Schizophrenia.

BACKGROUND: Various neurocognitive models explore perceptual distortions and hallucinations in schizophrenia and the general population. A variant of predictive coding account suggests that strong priors, like cognitive expectancy, may influence perception. This study examines if stronger cognitive expectancies result in more auditory false percepts in clinical and healthy control groups, investigates group differences, and explores the association between false percepts and hallucinations. STUDY DESIGN: Patients diagnosed with schizophrenia with current auditory hallucinations (n = 51) and without hallucinations (n = 66) and healthy controls (n = 51) underwent the False Perception Task under various expectancy conditions. All groups were examined for the presence and severity of hallucinations or hallucinatory-like experiences. STUDY RESULTS: We observed a main effect of condition across all groups, ie, the stronger the cognitive expectancy, the greater the ratio of auditory false percepts. However, there was no group effect for the ratio of auditory false percepts. Despite modest pairwise correlations in the hallucinating group, the ratio of auditory false percepts was not predicted by levels of hallucinations and hallucinatory-like experiences in a linear mixed model. CONCLUSIONS: The current study demonstrates that strong priors in the form of cognitive expectancies affect perception and play a role in perceptual disturbances. There is also a tentative possibility that overreliance on strong priors may be associated with hallucinations in currently hallucinating subjects. Possible, avoidable confounding factors are discussed in detail.

Cognitive simulation along with neural adaptation explain effects of suggestions: a novel theoretical framework.

Hypnosis is an effective intervention with proven efficacy that is employed in clinical settings and for investigating various cognitive processes. Despite their practical success, no consensus exists regarding the mechanisms underlying well-established hypnotic phenomena. Here, we suggest a new framework called the Simulation-Adaptation Theory of Hypnosis (SATH). SATH expands the predictive coding framework by focusing on (a) redundancy elimination in generative models using intrinsically generated prediction errors, (b) adaptation due to amplified or prolonged neural activity, and (c) using internally generated predictions as a venue for learning new associations. The core of our treatise is that simulating proprioceptive, interoceptive, and exteroceptive signals, along with the top-down attenuation of the precision of sensory prediction errors due to neural adaptation, can explain objective and subjective hypnotic phenomena. Based on these postulations, we offer mechanistic explanations for critical categories of direct verbal suggestions, including (1) direct-ideomotor, (2) challenge-ideomotor, (3) perceptual, and (4) cognitive suggestions. Notably, we argue that besides explaining objective responses, SATH accounts for the subjective effects of suggestions, i.e., the change in the sense of agency and reality. Finally, we discuss individual differences in hypnotizability and how SATH accommodates them. We believe that SATH is exhaustive and parsimonious in its scope, can explain a wide range of hypnotic phenomena without contradiction, and provides a host of testable predictions for future research.

Understanding Cognitive Behavioral Therapy for Psychosis Through the Predictive Coding Framework.

Psychological treatments for persecutory delusions, particularly cognitive behavioral therapy for psychosis, are efficacious; however, mechanistic theories explaining why they work rarely bridge to the level of cognitive neuroscience. Predictive coding, a general brain processing theory rooted in cognitive and computational neuroscience, has increasing experimental support for explaining symptoms of psychosis, including the formation and maintenance of delusions. Here, we describe recent advances in cognitive behavioral therapy for psychosis-based psychotherapy for persecutory delusions, which targets specific psychological processes at the computational level of information processing. We outline how Bayesian learning models employed in predictive coding are superior to simple associative learning models for understanding the impact of cognitive behavioral interventions at the algorithmic level. We review hierarchical predictive coding as an account of belief updating rooted in prediction error signaling. We examine how this process is abnormal in psychotic disorders, garnering noisy sensory data that is made sense of through the development of overly strong delusional priors. We argue that effective cognitive behavioral therapy for psychosis systematically targets the way sensory data are selected, experienced, and interpreted, thus allowing for the strengthening of alternative beliefs. Finally, future directions based on these arguments are discussed.

Delusions are distressing and disabling psychiatric symptoms. Cognitive behavioral therapy for psychosis (CBTp) is the leading psychotherapeutic approach for treating delusions. Predictive coding is a contemporary cognitive neuroscience framework that is increasingly being used to explain mechanisms of delusions. In this article, we attempt to integrate CBTp within the predictive coding framework, outlining how effective CBTp techniques impact aspects of the predictive coding model to contribute to cutting-edge treatment and cognitive neuroscience research on delusions and inform recommendations for treatment advancement.

Brain to brain musical interaction: A systematic review of neural synchrony in musical activities.

The use of hyperscanning technology has revealed the neural mechanisms underlying multi-person interaction in musical activities. However, there is currently a lack of integration among various research findings. This systematic review aims to provide a comprehensive understanding of the social dynamics and brain synchronization in music activities through the analysis of 32 studies. The findings illustrate a strong correlation between inter-brain synchronization (IBS) and various musical activities, with the frontal, central, parietal, and temporal lobes as the primary regions involved. The application of hyperscanning not only advances theoretical research but also holds practical significance in enhancing the effectiveness of music-based interventions in therapy and education. The review also utilizes Predictive Coding Models (PCM) to provide a new perspective for interpreting neural synchronization in music activities. To address the limitations of current research, future studies could integrate multimodal data, adopt novel technologies, use non-invasive techniques, and explore additional research directions.

Active sensing with predictive coding and uncertainty minimization.

We present an end-to-end architecture for embodied exploration inspired by two biological computations: predictive coding and uncertainty minimization. The architecture can be applied to any exploration setting in a task-independent and intrinsically driven manner. We first demonstrate our approach in a maze navigation task and show that it can discover the underlying transition distributions and spatial features of the environment. Second, we apply our model to a more complex active vision task, whereby an agent actively samples its visual environment to gather information. We show that our model builds unsupervised representations through exploration that allow it to efficiently categorize visual scenes. We further show that using these representations for downstream classification leads to superior data efficiency and learning speed compared to other baselines while maintaining lower parameter complexity. Finally, the modular structure of our model facilitates interpretability, allowing us to probe its internal mechanisms and representations during exploration.

The fear-avoidance model as an embodied prediction of threat.

The fear-avoidance model is a well-established framework in the understanding of persistent pain. It proposes a dichotomous path: either the context is interpreted as safe; there is no fear reaction and, therefore, the individual engages in active (positive) coping; or the context is interpreted as threatening, leading to a self-reinforcing vicious circle of fear and (negative) avoidance. We propose an embodied interpretation of this phenomenon employing the joint framework of predictive coding and active inference. The key idea is that multisensory integration of exteroceptive, proprioceptive, and interoceptive sensory inputs can lead to dysfunctional experiences of threat in nonthreatening situations. Threat inference can promote fear responses, maladaptive strategies (i.e., avoidance) and self-provides evidence for threat in associated or future contexts, or both. Under this treatment, the prediction of nonrealized threat becomes self-evidencing and context-invariant, and hence self-perpetuating. Safety cues are unable to attenuate the interpretation of the negative context as the dominant inference of the context is threatful and gains more precision and becomes resistant over time. Our model provides an explanation for the emergence of a dysfunctional fear response in the clinical setting despite apparent safety based on modern concepts from theoretical (computational) neuroscience.

Sense of ownership influence on tactile perception: Is the predictive coding account valid for the somatic rubber hand Illusion?

According to the predictive coding account, the attenuation of tactile perception on the hand exposed to the visuo-tactile Rubber Hand Illusion (vtRHI) relies on a weight increase of visual information deriving from the fake hand and a weight decrease of tactile information deriving from the individual's hand. To explore if this diametrical modulation persists in the absence of vision when adopting the somatic RHI (sRHI), we recorded tactile acuity measures before and after both RHI paradigms in 31 healthy individuals, hypothesizing a weight decrease for somatosensory information deriving from the hand undergoing the illusion and a weight increase for those deriving from the contralateral hand in the sRHI. Our results showed a significant overall decrease in tactile acuity on the hand undergoing the illusion whilst no changes emerged on the contralateral hand during sRHI. Since the sRHI was not accompanied by the hand spatial remapping, despite the generation of the feeling of ownership toward the fake hand, we hypothesized spatial remapping might play a pivotal role in determining sensory information weight attribution.

Contextual expectations shape the motor coding of movement kinematics during the prediction of observed actions: A TMS study.

Contextual information may shape motor resonance and support intention understanding during observation of incomplete, ambiguous actions. It is unclear, however, whether this effect is contingent upon kinematics ambiguity or contextual information is continuously integrated with kinematics to predict the overarching action intention. Moreover, a differentiation between the motor mapping of the intention suggested by context or kinematics has not been clearly demonstrated. In a first action execution phase, 29 participants were asked to perform reaching-to-grasp movements towards big or small food objects with the intention to eat or to move; electromyography from the First Dorsal Interosseous (FDI) and Abductor Digiti Minimi (ADM) was recorded. Depending on object size, the intentions to eat or to move were differently implemented by a whole-hand or a precision grip kinematics, thus qualifying an action-muscle dissociation. Then, in a following action prediction task, the same participants were asked to observe an actor performing the same actions and to predict his/her intention while motor resonance was assessed for the same muscles. Of note, videos were interrupted at early or late action phases, and actions were embedded in contexts pointing toward an eating or a moving intention, congruently or incongruently with kinematics. We found greater involvement of the FDI or ADM in the execution of precision or whole-hand grips, respectively. Crucially, this pattern of activation was mirrored during observation of the same actions in congruent contexts, but it was cancelled out or reversed in the incongruent ones, either when videos were interrupted at either early or long phases of action deployment. Our results extend previous evidence by showing that contextual information shapes motor resonance not only under conditions of perceptual uncertainty but also when more informative kinematics is available.

Understanding music and aging through the lens of Bayesian inference.

Bayesian inference has recently gained momentum in explaining music perception and aging. A fundamental mechanism underlying Bayesian inference is the notion of prediction. This framework could explain how predictions pertaining to musical (melodic, rhythmic, harmonic) structures engender action, emotion, and learning, expanding related concepts of music research, such as musical expectancies, groove, pleasure, and tension. Moreover, a Bayesian perspective of music perception may shed new insights on the beneficial effects of music in aging. Aging could be framed as an optimization process of Bayesian inference. As predictive inferences refine over time, the reliance on consolidated priors increases, while the updating of prior models through Bayesian inference attenuates. This may affect the ability of older adults to estimate uncertainties in their environment, limiting their cognitive and behavioral repertoire. With Bayesian inference as an overarching framework, this review synthesizes the literature on predictive inferences in music and aging, and details how music could be a promising tool in preventive and rehabilitative interventions for older adults through the lens of Bayesian inference.

Chronic pain - A maladaptive compensation to unbalanced hierarchical predictive processing.

The ability to perceive pain presents an interesting evolutionary advantage to adapt to an ever-changing environment. However, in the case of chronic pain (CP), pain perception hinders the capacity of the system to adapt to changing sensory environments. Similar to other chronic perceptual disorders, CP is also proposed to be a maladaptive compensation to aberrant sensory predictive processing. The local-global oddball paradigm relies on learning hierarchical rules and processing environmental irregularities at a local and global level. Prediction errors (PE) between actual and predicted input typically trigger an update of the forward model to limit the probability of encountering future PEs. It has been hypothesised that CP hinders forward model updating, reflected in increased local deviance and decreased global deviance. In the present study, we used the local-global paradigm to examine how CP influences hierarchical learning relative to healthy controls. As hypothesised, we observed that deviance in the stimulus characteristics evoked heightened local deviance and decreased global deviance of the stimulus-driven PE. This is also accompanied by respective changes in theta phase locking that is correlated with the subjective pain perception. Changes in the global deviant in the stimulus-driven-PE could also be explained by dampened attention-related responses. Changing the context of the auditory stimulus did not however show a difference in the context-driven PE. These findings suggest that CP is accompanied by maladaptive forward model updating where the constant presence of pain perception disrupts local deviance in non-nociceptive domains. Furthermore, we hypothesise that the auditory-processing based biomarker identified here could be a marker of domain-general dysfunction that could be confirmed by future research.

Neuromodulation of safety and surprise in the early stages of infant development: affective homeostatic regulation in bodily and mental functions.

Developing a sense of internal safety and security depends mainly on others: numerous neuromodulators play a significant role in the homeostatic process, regulating the importance of proximity to a caregiver and experiencing feelings that enable us to regulate our interdependence with our conspecifics since birth. This array of neurofunctional structures have been called the SEPARATION DISTRESS system (now more commonly known as the PANIC/ GRIEF system). This emotional system is mainly involved in the production of depressive symptoms. The disruption of this essential emotional balance leads to the onset of feelings of panic followed by depression. We will focus on the neuropeptides that play a crucial role in social approach behavior in mammals, which enhance prosocial behavior and facilitate the consolidation of social bonds. We propose that most prosocial behaviors are regulated through the specific neuromodulators acting on salient intersubjective stimuli, reflecting an increased sense of inner confidence (safety) in social relationships. This review considers the neurofunctional link between the feelings that may ultimately be at the base of a sense of inner safety and the central neuromodulatory systems. This link may shed light on the clinical implications for the development of early mother-infant bonding and the depressive clinical consequences when this bond is disrupted, such as in post-partum depression, depressive feelings connected to, addiction, neurofunctional disorders, and psychological trauma.

Adaptation in the visual system: Networked fatigue or suppressed prediction error signalling?

Our brains are constantly adapting to changes in our visual environments. Neural adaptation exerts a persistent influence on the activity of sensory neurons and our perceptual experience, however there is a lack of consensus regarding how adaptation is implemented in the visual system. One account describes fatigue-based mechanisms embedded within local networks of stimulus-selective neurons (networked fatigue models). Another depicts adaptation as a product of stimulus expectations (predictive coding models). In this review, I evaluate neuroimaging and psychophysical evidence that poses fundamental problems for predictive coding models of neural adaptation. Specifically, I discuss observations of distinct repetition and expectation effects, as well as incorrect predictions of repulsive adaptation aftereffects made by predictive coding accounts. Based on this evidence, I argue that networked fatigue models provide a more parsimonious account of adaptation effects in the visual system. Although stimulus expectations can be formed based on recent stimulation history, any consequences of these expectations are likely to co-occur (or interact) with effects of fatigue-based adaptation. I conclude by proposing novel, testable hypotheses relating to interactions between fatigue-based adaptation and other predictive processes, focusing on stimulus feature extrapolation phenomena.

Predictive sequence learning in the hippocampal formation.

The hippocampus receives sequences of sensory inputs from the cortex during exploration and encodes the sequences with millisecond precision. We developed a predictive autoencoder model of the hippocampus including the trisynaptic and monosynaptic circuits from the entorhinal cortex (EC). CA3 was trained as a self-supervised recurrent neural network to predict its next input. We confirmed that CA3 is predicting ahead by analyzing the spike coupling between simultaneously recorded neurons in the dentate gyrus, CA3, and CA1 of the mouse hippocampus. In the model, CA1 neurons signal prediction errors by comparing CA3 predictions to the next direct EC input. The model exhibits the rapid appearance and slow fading of CA1 place cells and displays replay and phase precession from CA3. The model could be learned in a biologically plausible way with error-encoding neurons. Similarities between the hippocampal and thalamocortical circuits suggest that such computation motif could also underlie self-supervised sequence learning in the cortex.

An EEG-Based Neuroplastic Approach to Predictive Coding in People With Schizophrenia or Traumatic Brain Injury.

Despite different etiologies, people with schizophrenia (SCZ) or with traumatic brain injury (TBI) both show aberrant neuroplasticity. One neuroplastic mechanism that may be affected is prediction error coding. We used a roving mismatch negativity (rMMN) paradigm which uses different lengths of standard tone trains and is optimized to assess predictive coding. Twenty-five SCZ, 22 TBI (mild to moderate), and 25 healthy controls were assessed. We used a frequency-deviant rMMN in which the number of standards preceding the deviant was either 2, 6, or 36. We evaluated repetition positivity to the standard tone immediately preceding a deviant tone (repetition positivity [RP], to assess formation of the memory trace), deviant negativity to the deviant stimulus (deviant negativity [DN], which reflects signaling of a prediction error), and the difference wave between the 2 (the MMN). We found that SCZ showed reduced DN and MMN compared with healthy controls and with people with mild to moderate TBI. We did not detect impairments in any index (RP, DN, or MMN) in people with TBI compared to controls. Our findings suggest that prediction error coding assessed with rMMN is aberrant in SCZ but intact in TBI, though there is a suggestion that severity of head injury results in poorer prediction error coding.