Cerebral responses to local and global auditory novelty under general anesthesia

Transient brain activity dynamics discriminate levels of consciousness during anesthesia

The awake mammalian brain is functionally organized in terms of large-scale distributed networks that are constantly interacting. Loss of consciousness might disrupt this temporal organization leaving patients unresponsive. We hypothesize that characterizing brain activity in terms of transient events may provide a signature of consciousness. For this, we analyze temporal dynamics of spatiotemporally overlapping functional networks obtained from fMRI transient activity across different anesthetics and levels of anesthesia. We first show a striking homology in spatial organization of networks between monkeys and humans, indicating cross-species similarities in resting-state fMRI structure. We then track how network organization shifts under different anesthesia conditions in macaque monkeys. While the spatial aspect of the networks is preserved, their temporal dynamics are highly affected by anesthesia. Networks express for longer durations and co-activate in an anesthetic-specific configuration. Additionally, hierarchical brain organization is disrupted with a consciousness-level-signature role of the default mode network. In conclusion, large-scale brain network temporal dynamics capture differences in anesthetic-specific consciousness-level, paving the way towards a clinical translation of these cortical signature.

A measure centrality index for systematic empirical comparison of consciousness theories

Consciousness science is marred by disparate constructs and methodologies, making it challenging to systematically compare theories. This foundational crisis casts doubts on the scientific character of the field itself. Addressing it, we propose a framework for systematically comparing consciousness theories by introducing a novel inter-theory classification interface, the Measure Centrality Index (MCI). Recognizing its gradient distribution, the MCI assesses the degree of importance a specific empirical measure has for a given consciousness theory. We apply the MCI to probe how the empirical measures of the Global Neuronal Workspace Theory (GNW), Integrated Information Theory (IIT), and Temporospatial Theory of Consciousness (TTC) would fare within the context of the other two. We demonstrate that direct comparison of IIT, GNW, and TTC is meaningful and valid for some measures like Lempel-Ziv Complexity (LZC), Autocorrelation Window (ACW), and possibly Mutual Information (MI). In contrast, it is problematic for others like the anatomical and physiological neural correlates of consciousness (NCC) due to their MCI-based differential weightings within the structure of the theories. In sum, we introduce and provide proof-of-principle of a novel systematic method for direct inter-theory empirical comparisons, thereby addressing isolated evolution of theories and confirmatory bias issues in the state-of-the-art neuroscience of consciousness.

An integrative view of the role of prefrontal cortex in consciousness

The involvement of the prefrontal cortex (PFC) in consciousness is an ongoing focus of intense investigation. An important question is whether representations of conscious contents and experiences in the PFC are confounded by post-perceptual processes related to cognitive functions. Here, I review recent findings suggesting that neuronal representations of consciously perceived contents-in the absence of post-perceptual processes-can indeed be observed in the PFC. Slower ongoing fluctuations in the electrophysiological state of the PFC seem to control the stability and updates of these prefrontal representations of conscious awareness. In addition to conscious perception, the PFC has been shown to play a critical role in controlling the levels of consciousness as observed during anesthesia, while prefrontal lesions can result in severe loss of perceptual awareness. Together, the convergence of these processes in the PFC suggests its integrative role in consciousness and highlights the complex nature of consciousness itself.

Local orchestration of distributed functional patterns supporting loss and restoration of consciousness in the primate brain

A central challenge of neuroscience is to elucidate how brain function supports consciousness. Here, we combine the specificity of focal deep brain stimulation with fMRI coverage of the entire cortex, in awake and anaesthetised non-human primates. During propofol, sevoflurane, or ketamine anaesthesia, and subsequent restoration of responsiveness by electrical stimulation of the central thalamus, we investigate how loss of consciousness impacts distributed patterns of structure-function organisation across scales. We report that distributed brain activity under anaesthesia is increasingly constrained by brain structure across scales, coinciding with anaesthetic-induced collapse of multiple dimensions of hierarchical cortical organisation. These distributed signatures are observed across different anaesthetics, and they are reversed by electrical stimulation of the central thalamus, coinciding with recovery of behavioural markers of arousal. No such effects were observed upon stimulating the ventral lateral thalamus, demonstrating specificity. Overall, we identify consistent distributed signatures of consciousness that are orchestrated by specific thalamic nuclei.

Multiple Concurrent Predictions Inform Prediction Error in the Human Auditory Pathway

The key assumption of the predictive coding framework is that internal representations are used to generate predictions on how the sensory input will look like in the immediate future. These predictions are tested against the actual input by the so-called prediction error units, which encode the residuals of the predictions. What happens to prediction errors, however, if predictions drawn by different stages of the sensory hierarchy contradict each other? To answer this question, we conducted two fMRI experiments while female and male human participants listened to sequences of sounds: pure tones in the first experiment and frequency-modulated sweeps in the second experiment. In both experiments, we used repetition to induce predictions based on stimulus statistics (stats-informed predictions) and abstract rules disclosed in the task instructions to induce an orthogonal set of (task-informed) predictions. We tested three alternative scenarios: neural responses in the auditory sensory pathway encode prediction error with respect to (1) the stats-informed predictions, (2) the task-informed predictions, or (3) a combination of both. Results showed that neural populations in all recorded regions (bilateral inferior colliculus, medial geniculate body, and primary and secondary auditory cortices) encode prediction error with respect to a combination of the two orthogonal sets of predictions. The findings suggest that predictive coding exploits the non-linear architecture of the auditory pathway for the transmission of predictions. Such non-linear transmission of predictions might be crucial for the predictive coding of complex auditory signals like speech.Significance Statement Sensory systems exploit our subjective expectations to make sense of an overwhelming influx of sensory signals. It is still unclear how expectations at each stage of the processing pipeline are used to predict the representations at the other stages. The current view is that this transmission is hierarchical and linear. Here we measured fMRI responses in auditory cortex, sensory thalamus, and midbrain while we induced two sets of mutually inconsistent expectations on the sensory input, each putatively encoded at a different stage. We show that responses at all stages are concurrently shaped by both sets of expectations. The results challenge the hypothesis that expectations are transmitted linearly and provide for a normative explanation of the non-linear physiology of the corticofugal sensory system.

Monkey Dorsolateral Prefrontal Cortex Represents Abstract Visual Sequences during a No-Report Task

Monitoring sequential information is an essential component of our daily lives. Many of these sequences are abstract, in that they do not depend on the individual stimuli, but do depend on an ordered set of rules (e.g., chop then stir when cooking). Despite the ubiquity and utility of abstract sequential monitoring, little is known about its neural mechanisms. Human rostrolateral prefrontal cortex (RLPFC) exhibits specific increases in neural activity (i.e., "ramping") during abstract sequences. Monkey dorsolateral prefrontal cortex (DLPFC) has been shown to represent sequential information in motor (not abstract) sequence tasks, and contains a subregion, area 46, with homologous functional connectivity to human RLPFC. To test the prediction that area 46 may represent abstract sequence information, and do so with parallel dynamics to those found in humans, we conducted functional magnetic resonance imaging (fMRI) in three male monkeys. When monkeys performed no-report abstract sequence viewing, we found that left and right area 46 responded to abstract sequential changes. Interestingly, responses to rule and number changes overlapped in right area 46 and left area 46 exhibited responses to abstract sequence rules with changes in ramping activation, similar to that observed in humans. Together, these results indicate that monkey DLPFC monitors abstract visual sequential information, potentially with a preference for different dynamics in the two hemispheres. More generally, these results show that abstract sequences are represented in functionally homologous regions across monkeys and humans.SIGNIFICANCE STATEMENT Daily, we complete sequences that are "abstract" because they depend on an ordered set of rules (e.g., chop then stir when cooking) rather than the identity of individual items. Little is known about how the brain tracks, or monitors, this abstract sequential information. Based on previous human work showing abstract sequence related dynamics in an analogous area, we tested whether monkey dorsolateral prefrontal cortex (DLPFC), specifically area 46, represents abstract sequential information using awake monkey functional magnetic resonance imaging (fMRI). We found that area 46 responded to abstract sequence changes, with a preference for more general responses on the right and dynamics similar to humans on the left. These results suggest that abstract sequences are represented in functionally homologous regions across monkeys and humans.

Not with a “zap” but with a “beep”: measuring the origins of perinatal experience

When does the mind begin? Infant psychology is mysterious in part because we cannot remember our first months of life, nor can we directly communicate with infants. Even more speculative is the possibility of mental life prior to birth. The question of when consciousness, or subjective experience, begins in human development thus remains incompletely answered, though boundaries can be set using current knowledge from developmental neurobiology and recent investigations of the perinatal brain. Here, we offer our perspective on how the development of a sensory perturbational complexity index (sPCI) based on auditory ("beep-and-zip"), visual ("flash-and-zip"), or even olfactory ("sniff-and-zip") cortical perturbations in place of electromagnetic perturbations ("zap-and-zip") might be used to address this question. First, we discuss recent studies of perinatal cognition and consciousness using techniques such as functional magnetic resonance imaging (fMRI), electroencephalography (EEG), and, in particular, magnetoencephalography (MEG). While newborn infants are the archetypal subjects for studying early human development, researchers may also benefit from fetal studies, as the womb is, in many respects, a more controlled environment than the cradle. The earliest possible timepoint when subjective experience might begin is likely the establishment of thalamocortical connectivity at 26 weeks gestation, as the thalamocortical system is necessary for consciousness according to most theoretical frameworks. To infer at what age and in which behavioral states consciousness might emerge following the initiation of thalamocortical pathways, we advocate for the development of the sPCI and similar techniques, based on EEG, MEG, and fMRI, to estimate the perinatal brain's state of consciousness.

Passive functional mapping of receptive language cortex during general anesthesia using electrocorticography

OBJECTIVE

To investigate the feasibility of passive functional mapping in the receptive language cortex during general anesthesia using electrocorticographic (ECoG) signals.

METHODS

We used subdurally placed ECoG grids to record cortical responses to speech stimuli during awake and anesthesia conditions. We identified the cortical areas with significant responses to the stimuli using the spectro-temporal consistency of the brain signal in the broadband gamma (BBG) frequency band (70-170 Hz).

RESULTS

We found that ECoG BBG responses during general anesthesia effectively identify cortical regions associated with receptive language function. Our analyses demonstrated that the ability to identify receptive language cortex varies across different states and depths of anesthesia. We confirmed these results by comparing them to receptive language areas identified during the awake condition. Quantification of these results demonstrated an average sensitivity and specificity of passive language mapping during general anesthesia to be 49±7.7% and 100%, respectively.

CONCLUSION

Our results demonstrate that mapping receptive language cortex in patients during general anesthesia is feasible.

SIGNIFICANCE

Our proposed protocol could greatly expand the population of patients that can benefit from passive language mapping techniques, and could eliminate the risks associated with electrocortical stimulation during an awake craniotomy.

Temporal hierarchies in the predictive processing of melody - From pure tones to songs

Listening to musical melodies is a complex task that engages perceptual and memoryrelated processes. The processes underlying melody cognition happen simultaneously on different timescales, ranging from milliseconds to minutes. Although attempts have been made, research on melody perception is yet to produce a unified framework of how melody processing is achieved in the brain. This may in part be due to the difficulty of integrating concepts such as perception, attention and memory, which pertain to different temporal scales. Recent theories on brain processing, which hold prediction as a fundamental principle, offer potential solutions to this problem and may provide a unifying framework for explaining the neural processes that enable melody perception on multiple temporal levels. In this article, we review empirical evidence for predictive coding on the levels of pitch formation, basic pitch-related auditory patterns,more complex regularity processing extracted from basic patterns and long-term expectations related to musical syntax. We also identify areas that would benefit from further inquiry and suggest future directions in research on musical melody perception.

Alterations in whole-brain dynamic functional stability during memory tasks under dexmedetomidine sedation

Purpose

This study aimed to explore the neurological effects of dexmedetomidine-induced sedation on memory using functional stability, a whole-brain voxel-wise dynamic functional connectivity approach.

Methods

A total of 16 participants (10 men) underwent auditory memory task-related fMRI in the awake state and under dexmedetomidine sedation. Explicit and implicit memory tests were conducted 4 h after ceasing dexmedetomidine administration. One-sample Wilcoxon signed rank test was applied to determine the formation of explicit and implicit memory in the two states. Functional stability was calculated and compared voxel-wise between the awake and sedated states. The association between functional stability and memory performance was also assessed.

Results

In the awake baseline tests, explicit and implicit memory scores were significantly different from zero (p < 0.05). In the tests under sedation, explicit and implicit memory scores were not significantly different from zero. Compared to that at wakeful baseline, functional stability during light sedation was reduced in the medial prefrontal cortex, left angular gyrus, and right hippocampus (all clusters, p < 0.05, GRF-corrected), whereas the left superior temporal gyrus exhibited higher functional stability (cluster p < 0.05, GRF-corrected). No significant associations were observed between functional stability and memory test scores.

Conclusions

The distribution and patterns of alterations in functional stability during sedation illustrate the modulation of functional architecture by dexmedetomidine from a dynamic perspective. Our findings provide novel insight into the dynamic brain functional networks underlying consciousness and memory in humans.

Selective corticocortical connectivity suppression during propofol-induced anesthesia in healthy volunteers

We comprehensively studied directional feedback and feedforward connectivity to explore potential connectivity changes that underlie propofol-induced deep sedation. We further investigated the corticocortical connectivity patterns within and between hemispheres. Sixty-channel electroencephalographic data were collected from 19 healthy volunteers in a resting wakefulness state and propofol-induced deep unconsciousness state defined by a bispectral index value of 40. A source analysis was employed to locate cortical activity. The Desikan-Killiany atlas was used to partition cortices, and directional functional connectivity was assessed by normalized symbolic transfer entropy between higher-order (prefrontal and frontal) and lower-order (auditory, sensorimotor and visual) cortices and between hot-spot frontal and parietal cortices. We found that propofol significantly suppressed feedforward connectivity from the left parietal to right frontal cortex and bidirectional connectivity between the left frontal and left parietal cortex, between the frontal and auditory cortex, and between the frontal and sensorimotor cortex. However, there were no significant changes in either feedforward or feedback connectivity between the prefrontal and all the lower-order cortices and between the frontal and visual cortices or in feedback connectivity from the frontal to parietal cortex. Propofol anesthetic selectively decreased the unidirectional interaction between higher-order frontoparietal cortices and bidirectional interactions between the higher-order frontal cortex and lower-order auditory and sensorimotor cortices, which indicated that both feedback and feedforward connectivity were suppressed under propofol-induced deep sedation. Our findings provide critical insights into the connectivity changes underlying the top-down mechanism of propofol anesthesia at deep sedation.

Supplementary Information

The online version contains supplementary material available at 10.1007/s11571-021-09775-x.

Cerebro-cerebellar interactions in non-human primates examined by optogenetic functional magnetic resonance imaging

Functional magnetic resonance imaging (fMRI) is a promising approach for the simultaneous and extensive scanning of whole-brain activities. Optogenetics is free from electrical and magnetic artifacts and is an ideal stimulation method for combined use with fMRI. However, the application of optogenetics in nonhuman primates (NHPs) remains limited. Recently, we developed an efficient optogenetic intracortical microstimulation method of the primary motor cortex (M1), which successfully induced forelimb movements in macaque monkeys. Here, we aimed to investigate how optogenetic M1 stimulation causes neural modulation in the local and remote brain regions in anesthetized monkeys using 7-tesla fMRI. We demonstrated that optogenetic stimulation of the M1 forelimb and hindlimb regions successfully evoked robust direct and remote fMRI activities. Prominent remote activities were detected in the anterior and posterior lobes in the contralateral cerebellum, which receive projections polysynaptically from the M1. We further demonstrated that the cerebro-cerebellar projections from these M1 regions were topographically organized, which is concordant with the somatotopic map in the cerebellar cortex previously reported in macaques and humans. The present study significantly enhances optogenetic fMRI in NHPs, resulting in profound understanding of the brain network, thereby accelerating the translation of findings from animal models to humans.

Deep brain stimulation of the thalamus restores signatures of consciousness in a nonhuman primate model

Loss of consciousness is associated with the disruption of long-range thalamocortical and corticocortical brain communication. We tested the hypothesis that deep brain stimulation (DBS) of central thalamus might restore both arousal and awareness following consciousness loss. We applied anesthesia to suppress consciousness in nonhuman primates. During anesthesia, central thalamic stimulation induced arousal in an on-off manner and increased functional magnetic resonance imaging activity in prefrontal, parietal, and cingulate cortices. Moreover, DBS restored a broad dynamic repertoire of spontaneous resting-state activity, previously described as a signature of consciousness. None of these effects were obtained during the stimulation of a control site in the ventrolateral thalamus. Last, DBS restored a broad hierarchical response to auditory violations that was disrupted under anesthesia. Thus, DBS restored the two dimensions of consciousness, arousal and conscious access, following consciousness loss, paving the way to its therapeutical translation in patients with disorders of consciousness.

Interconnected sub-networks of the macaque monkey gustatory connectome

Macroscopic taste processing connectivity was investigated using functional magnetic resonance imaging during the presentation of sour, salty, and sweet tastants in anesthetized macaque monkeys. This examination of taste processing affords the opportunity to study the interactions between sensory regions, central integrators, and effector areas. Here, 58 brain regions associated with gustatory processing in primates were aggregated, collectively forming the gustatory connectome. Regional regression coefficients (or β-series) obtained during taste stimulation were correlated to infer functional connectivity. This connectivity was then evaluated by assessing its laterality, modularity and centrality. Our results indicate significant correlations between same region pairs across hemispheres in a bilaterally interconnected scheme for taste processing throughout the gustatory connectome. Using unbiased community detection, three bilateral sub-networks were detected within the graph of the connectome. This analysis revealed clustering of 16 medial cortical structures, 24 lateral structures, and 18 subcortical structures. Across the three sub-networks, a similar pattern was observed in the differential processing of taste qualities. In all cases, the amplitude of the response was greatest for sweet, but the network connectivity was strongest for sour and salty tastants. The importance of each region in taste processing was computed using node centrality measures within the connectome graph, showing centrality to be correlated across hemispheres and, to a smaller extent, region volume. Connectome hubs exhibited varying degrees of centrality with a prominent leftward increase in insular cortex centrality. Taken together, these criteria illustrate quantifiable characteristics of the macaque monkey gustatory connectome and its organization as a tri-modular network, which may reflect the general medial-lateral-subcortical organization of salience and interoception processing networks.

Consciousness as a multidimensional phenomenon: implications for the assessment of disorders of consciousness

Disorders of consciousness (DoCs) pose a significant clinical and ethical challenge because they allow for complex forms of conscious experience in patients where intentional behaviour and communication are highly limited or non-existent. There is a pressing need for brain-based assessments that can precisely and accurately characterize the conscious state of individual DoC patients. There has been an ongoing research effort to develop neural measures of consciousness. However, these measures are challenging to validate not only due to our lack of ground truth about consciousness in many DoC patients but also because there is an open ontological question about consciousness. There is a growing, well-supported view that consciousness is a multidimensional phenomenon that cannot be fully described in terms of the theoretical construct of hierarchical, easily ordered conscious levels. The multidimensional view of consciousness challenges the utility of levels-based neural measures in the context of DoC assessment. To examine how these measures may map onto consciousness as a multidimensional phenomenon, this article will investigate a range of studies where they have been applied in states other than DoC and where more is known about conscious experience. This comparative evidence suggests that measures of conscious level are more sensitive to some dimensions of consciousness than others and cannot be assumed to provide a straightforward hierarchical characterization of conscious states. Elevated levels of brain complexity, for example, are associated with conscious states characterized by a high degree of sensory richness and minimal attentional constraints, but are suboptimal for goal-directed behaviour and external responsiveness. Overall, this comparative analysis indicates that there are currently limitations to the use of these measures as tools to evaluate consciousness as a multidimensional phenomenon and that the relationship between these neural signatures and phenomenology requires closer scrutiny.

Implicit auditory perception of local and global irregularities in passive listening condition

The auditory system detects differences in sounds at an implicit level, but data on this difference might not be sufficient to make explicit discrimination. The biomarkers of implicit auditory memory of ambiguous stimuli could shed light on unconscious auditory processing and implicit auditory learning. Mismatch negativity (MMN) and P3a, components of event-related potentials (ERPs) reflecting stimuli discrimination without direct attention, were previously detected in response to the local (short-term) irregularity in the auditory sequence even in an unconscious state. At the same time, P3b was elicited only in case of direct attention in response to the global (long-term) irregularity. In this study, we applied the local-global auditory paradigm to obtain possible electrophysiological signatures of implicit detection of hardly distinguishable auditory stimuli. ERPs were recorded from 20 healthy volunteers during active discrimination of deviant sounds in the old-ball sequence and passive listening of the same sounds in the sequence with local-global irregularity. The discrimination task consisted of two blocks with different deviant sounds targeted to respond. The sound discrimination accuracy was at an average of 40%, implying the difficulty of explicit sound recognition. Comparing ERPs to standard and deviant sounds, we found posterior negativity in ERP around 450-600 ms in response to targeted deviant sounds. MMN was significant only in response to non-target deviants. In the passive local-global paradigm, we observed an anterior positivity (284-412 ms), compatible with P3a, in response to a violation of local regularity. Violation of global regularity elicited an anterior negative response (228-586 ms), resembling the N400 component of ERPs. Importantly, the other indexes of auditory discrimination, such as MMN and P3b, were insignificant in ERPs to both regularity violations. The observed P3a and N400 components of ERPs may reflect prediction error signals in the implicit perception of sound patterns even if behavioral recognition was poor.

Arousal State-Dependence of Interactions Between Short- and Long-Term Auditory Novelty Responses in Human Subjects

In everyday life, predictable sensory stimuli are generally not ecologically informative. By contrast, novel or unexpected stimuli signal ecologically salient changes in the environment. This idea forms the basis of the predictive coding hypothesis: efficient sensory encoding minimizes neural activity associated with predictable backgrounds and emphasizes detection of changes in the environment. In real life, the brain must resolve multiple unexpected sensory events occurring over different time scales. The local/global deviant experimental paradigm examines auditory predictive coding over multiple time scales. For short-term novelty [hundreds of milliseconds; local deviance (LD)], sequences of identical sounds (/xxxxx/) are interspersed with sequences that contain deviants (/xxxxy/). Long-term novelty [several seconds; global deviance (GD)] is created using either (a) frequent /xxxxx/ and infrequent /xxxxy/ sequences, or (b) frequent /xxxxy/ and infrequent /xxxxx/ sequences. In scenario (a), there is both an LD and a GD effect (LDGD, "double surprise"). In (b), the global deviant is a local standard, i.e., sequence of identical sounds (LSGD). Cortical responses reflecting LD and GD originate in different brain areas, have a different time course, and are differentially sensitive to general anesthesia. Neural processes underlying LD and GD have been shown to interact, reflecting overlapping networks subserving the detection of novel auditory stimuli. This study examined these interactions using intracranial electroencephalography in neurosurgical patients. Subjects performed a GD target detection task before and during induction of anesthesia with propofol. Recordings were made from the auditory cortex, surrounding auditory-related and prefrontal cortex in awake, sedated, and unresponsive states. High gamma activity was used to measure the neural basis of local-by-global novelty interactions. Positive interaction was defined as a greater response to the double surprise LDGD condition compared to LSGD. Negative interaction was defined as a weaker response to LDGD. Positive interaction was more frequent than negative interaction and was primarily found in auditory cortex. Negative interaction typically occurred in prefrontal cortex and was more sensitive to general anesthesia. Temporo-parietal auditory-related areas exhibited both types of interaction. These interactions may have relevance in a clinical setting as biomarkers of conscious perception in the assessment of depth of anesthesia and disorders of consciousness.

Automatic Sensory Predictions: A Review of Predictive Mechanisms in the Brain and Their Link to Conscious Processing

The human brain has the astonishing capacity of integrating streams of sensory information from the environment and forming predictions about future events in an automatic way. Despite being initially developed for visual processing, the bulk of predictive coding research has subsequently focused on auditory processing, with the famous mismatch negativity signal as possibly the most studied signature of a surprise or prediction error (PE) signal. Auditory PEs are present during various consciousness states. Intriguingly, their presence and characteristics have been linked with residual levels of consciousness and return of awareness. In this review we first give an overview of the neural substrates of predictive processes in the auditory modality and their relation to consciousness. Then, we focus on different states of consciousness - wakefulness, sleep, anesthesia, coma, meditation, and hypnosis - and on what mysteries predictive processing has been able to disclose about brain functioning in such states. We review studies investigating how the neural signatures of auditory predictions are modulated by states of reduced or lacking consciousness. As a future outlook, we propose the combination of electrophysiological and computational techniques that will allow investigation of which facets of sensory predictive processes are maintained when consciousness fades away.

Combining brain perturbation and neuroimaging in non-human primates

Brain perturbation studies allow detailed causal inferences of behavioral and neural processes. Because the combination of brain perturbation methods and neural measurement techniques is inherently challenging, research in humans has predominantly focused on non-invasive, indirect brain perturbations, or neurological lesion studies. Non-human primates have been indispensable as a neurobiological system that is highly similar to humans while simultaneously being more experimentally tractable, allowing visualization of the functional and structural impact of systematic brain perturbation. This review considers the state of the art in non-human primate brain perturbation with a focus on approaches that can be combined with neuroimaging. We consider both non-reversible (lesions) and reversible or temporary perturbations such as electrical, pharmacological, optical, optogenetic, chemogenetic, pathway-selective, and ultrasound based interference methods. Method-specific considerations from the research and development community are offered to facilitate research in this field and support further innovations. We conclude by identifying novel avenues for further research and innovation and by highlighting the clinical translational potential of the methods.

Macroscopic Quantities of Collective Brain Activity during Wakefulness and Anesthesia

The study of states of arousal is key to understand the principles of consciousness. Yet, how different brain states emerge from the collective activity of brain regions remains unknown. Here, we studied the fMRI brain activity of monkeys during wakefulness and anesthesia-induced loss of consciousness. We showed that the coupling between each brain region and the rest of the cortex provides an efficient statistic to classify the two brain states. Based on this and other statistics, we estimated maximum entropy models to derive collective, macroscopic properties that quantify the system's capabilities to produce work, to contain information, and to transmit it, which were all maximized in the awake state. The differences in these properties were consistent with a phase transition from critical dynamics in the awake state to supercritical dynamics in the anesthetized state. Moreover, information-theoretic measures identified those parameters that impacted the most the network dynamics. We found that changes in the state of consciousness primarily depended on changes in network couplings of insular, cingulate, and parietal cortices. Our findings suggest that the brain state transition underlying the loss of consciousness is predominantly driven by the uncoupling of specific brain regions from the rest of the network.

Human Brain Ages With Hierarchy-Selective Attenuation of Prediction Errors

From the perspective of predictive coding, our brain embodies a hierarchical generative model to realize perception, which proactively predicts the statistical structure of sensory inputs. How are these predictive processes modified as we age? Recent research suggested that aging leads to decreased weighting of sensory inputs and increased reliance on predictions. Here we investigated whether this age-related shift from sensorium to predictions occurs at all levels of hierarchical message passing. We recorded the electroencephalography responses with an auditory local-global paradigm in a cohort of 108 healthy participants from 3 groups: seniors, adults, and adolescents. The detection of local deviancy seems largely preserved in older individuals at earlier latency (including the mismatch negativity followed by the P3a but not the reorienting negativity). In contrast, the detection of global deviancy is clearly compromised in older individuals, as they showed worse task performance and attenuated P3b. Our findings demonstrate that older brains show little decline in sensory (i.e., first-order) prediction errors but significant diminution in contextual (i.e., second-order) prediction errors. Age-related deficient maintenance of auditory information in working memory might affect whether and how lower-level prediction errors propagate to the higher level.

Using non-invasive neuroimaging to enhance the care, well-being and experimental outcomes of laboratory non-human primates (monkeys)

Over the past 10-20 years, neuroscience witnessed an explosion in the use of non-invasive imaging methods, particularly magnetic resonance imaging (MRI), to study brain structure and function. Simultaneously, with access to MRI in many research institutions, MRI has become an indispensable tool for researchers and veterinarians to guide improvements in surgical procedures and implants and thus, experimental as well as clinical outcomes, given that access to MRI also allows for improved diagnosis and monitoring for brain disease. As part of the PRIMEatE Data Exchange, we gathered expert scientists, veterinarians, and clinicians who treat humans, to provide an overview of the use of non-invasive imaging tools, primarily MRI, to enhance experimental and welfare outcomes for laboratory non-human primates engaged in neuroscientific experiments. We aimed to provide guidance for other researchers, scientists and veterinarians in the use of this powerful imaging technology as well as to foster a larger conversation and community of scientists and veterinarians with a shared goal of improving the well-being and experimental outcomes for laboratory animals.

Validating the use of functional Near-Infrared Spectroscopy in monkeys: The case of brain activation lateralization in Papio anubis

Hemispheric asymmetries have long been seen as characterizing the human brain; yet, an increasing number of reports suggest the presence of such brain asymmetries in our closest primate relatives. However, most available data in non-human primates have so far been acquired as part of neurostructural approaches such as MRI, while comparative data in humans are often dynamically acquired as part of neurofunctional studies. In the present exploratory study in baboons (Papio anubis), we tested whether brain lateralization could be recorded non-invasively using a functional Near-Infrared Spectroscopy (fNIRS) device in two contexts: motor and auditory passive stimulations. Under light propofol anaesthesia monitoring, three adult female baboons were exposed to a series of (1) left- versus right-arm passive movement stimulations; and (2) left- versus right-ear versus stereo auditory stimulations while recording fNIRS signals in the related brain areas (i.e., motor central sulcus and superior temporal cortices respectively). For the sensorimotor condition our results show that left-arm versus right-arm stimulations induced typical contralateral difference in hemispheric activation asymmetries in the three subjects. For the auditory condition, we also revealed typical human-like patterns of hemispheric asymmetries in one subject, namely a leftward lateralization for right ear stimulations for all three channels. Overall, our findings support the use of fNIRS to investigate brain processing in non-human primates from a functional perspective, opening the way for the development of non-invasive procedures in non-human primate brain research.

Hierarchical disruption in the cortex of anesthetized monkeys as a new signature of consciousness loss

Anesthesia induces a reconfiguration of the repertoire of functional brain states leading to a high function-structure similarity. However, it is unclear how these functional changes lead to loss of consciousness. Here we suggest that the mechanism of conscious access is related to a general dynamical rearrangement of the intrinsic hierarchical organization of the cortex. To measure cortical hierarchy, we applied the Intrinsic Ignition analysis to resting-state fMRI data acquired in awake and anesthetized macaques. Our results reveal the existence of spatial and temporal hierarchical differences of neural activity within the macaque cortex, with a strong modulation by the depth of anesthesia and the employed anesthetic agent. Higher values of Intrinsic Ignition correspond to rich and flexible brain dynamics whereas lower values correspond to poor and rigid, structurally driven brain dynamics. Moreover, spatial and temporal hierarchical dimensions are disrupted in a different manner, involving different hierarchical brain networks. All together suggest that disruption of brain hierarchy is a new signature of consciousness loss.

Sedation Modulates Frontotemporal Predictive Coding Circuits and the Double Surprise Acceleration Effect

Two important theories in cognitive neuroscience are predictive coding (PC) and the global workspace (GW) theory. A key research task is to understand how these two theories relate to one another, and particularly, how the brain transitions from a predictive early state to the eventual engagement of a brain-scale state (the GW). To address this question, we present a source-localization of EEG responses evoked by the local-global task—an experimental paradigm that engages a predictive hierarchy, which encompasses the GW. The results of our source reconstruction suggest three phases of processing. The first phase involves the sensory (here auditory) regions of the superior temporal lobe and predicts sensory regularities over a short timeframe (as per the local effect). The third phase is brain-scale, involving inferior frontal, as well as inferior and superior parietal regions, consistent with a global neuronal workspace (GNW; as per the global effect). Crucially, our analysis suggests that there is an intermediate (second) phase, involving modulatory interactions between inferior frontal and superior temporal regions. Furthermore, sedation with propofol reduces modulatory interactions in the second phase. This selective effect is consistent with a PC explanation of sedation, with propofol acting on descending predictions of the precision of prediction errors; thereby constraining access to the GNW.

The effectiveness of PC6 acupuncture in the prevention of postoperative nausea and vomiting in children: A systematic review and meta-analysis

BACKGROUND

A growing number of studies have demonstrated the effectiveness of acupuncture in preventing and treating postoperative nausea and vomiting. Here, we used meta-analysis to confirm these benefits in children and to determine the optimal time to perform this treatment.

METHODS

Four databases (MEDLINE, EMBASE, CENTRAL, and Chinese Database of Biology and Medicine) were searched from inception until January 16, 2019. We included randomized controlled trials for evaluating the effectiveness of acupuncture in the prevention and treatment of postoperative nausea and vomiting during the early stage (0-4 hours) and within 24 hours postoperatively in pediatrics. Control groups received standardized care control or standardized care combined with sham control.

RESULTS

Sixteen literatures and 1773 patients undergoing general anesthesia were included in the study. The results indicated that acupuncture was effective in reducing postoperative vomiting, both during the first 4 hours (RR = 0.47, 95% CI 0.26, 0.84; low quality) and within 24 hours postoperatively (RR = 0.74, 95% CI 0.60, 0.91; low quality). Stratifying by the timing of acupuncture, acupuncture was effective in reducing the first 4 hours (RR = 0.34, 95% CI 0.18, 0.64; moderate quality), and 0-24 hours postoperative vomiting (RR = 0.81, 95% CI 0.70, 0.93; moderate quality) when performed before and during anesthesia, respectively. Further, the RR value was more robust when acupuncture was performed before anesthesia. Acupuncture was also effective in treating 0-24 hours postoperative nausea (RR = 0.73, 95% CI 0.60, 0.88; moderate quality) and in reducing the utilization of remedies during the first 4 hours (RR = 0.64, 95% CI 0.45, 0.89; moderate quality).

CONCLUSION

Acupuncture reduces the incidence of postoperative nausea and vomiting as well as the utilization of antiemetic remedies, particularly during the first 4 hours following the operation. Acupuncture performed before anesthesia was demonstrated to be the most ideal intervention time for children.

Cortical responses to auditory novelty across task conditions: An intracranial electrophysiology study

Elucidating changes in sensory processing across attentional and arousal states is a major focus in neuroscience. The local/global deviant (LGD) stimulus paradigm engages auditory predictive coding over short (local deviance, LD) and long (global deviance, GD) time scales, and has been used to assay disruption of auditory predictive coding upon loss of consciousness. Our previous work (Nourski et al., 2018, J Neurosci 38:8441-52) examined effects of general anesthesia on short- and long-term novelty detection. GD effects were suppressed at subhypnotic doses of propofol, suggesting that they may be more related to task engagement than consciousness per se. The present study addressed this hypothesis by comparing cortical responses to auditory novelty during passive versus active listening conditions in awake listeners. Subjects were seven adult neurosurgical patients undergoing chronic invasive monitoring for medically intractable epilepsy. LGD stimuli were sequences of four identical vowels followed by a fifth identical or different vowel. In the passive condition, the stimuli were presented to subjects as they watched a silent TV program and were instructed to attend to its content. In the active condition, stimuli were presented in the absence of a TV program, and subjects were instructed to press a button in response to GD target stimuli. Intracranial recordings were made from multiple brain regions, including core and non-core auditory, auditory-related, prefrontal and sensorimotor cortex. Metrics of task performance included hit rate, sensitivity index, and reaction times. Cortical activity was measured as averaged auditory evoked potentials (AEPs) and event-related band power in high gamma (70-150 Hz) and alpha (8-14 Hz) frequency bands. The vowel stimuli and LD elicited robust AEPs in all studied brain areas in both passive and active conditions. High gamma responses to stimulus onset and LD were localized predominantly to the auditory cortex in the superior temporal plane and had a comparable prevalence and spatial extent between the two conditions. In contrast, GD effects (AEPs, high gamma and alpha suppression) were greatly enhanced during the active condition in all studied brain areas. The prevalence of high gamma GD effects was positively correlated with individual subjects' task performance. The data demonstrate distinct task engagement-related effects on responses to auditory novelty across the auditory cortical processing hierarchy. The results motivate a closer examination of effective connectivity underlying attentional modulation of cortical sensory responses, and serve as a foundation for examining changes in sensory processing associated with general anesthesia, sleep and disorders of consciousness.

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.

Resting-state Dynamics as a Cortical Signature of Anesthesia in Monkeys

WHAT WE ALREADY KNOW ABOUT THIS TOPIC

WHAT THIS ARTICLE TELLS US THAT IS NEW: BACKGROUND:: The mechanism by which anesthetics induce a loss of consciousness remains a puzzling problem. We hypothesized that a cortical signature of anesthesia could be found in an increase in similarity between the matrix of resting-state functional correlations and the anatomical connectivity matrix of the brain, resulting in an increased function-structure similarity.

METHODS

We acquired resting-state functional magnetic resonance images in macaque monkeys during wakefulness (n = 3) or anesthesia with propofol (n = 3), ketamine (n = 3), or sevoflurane (n = 3). We used the k-means algorithm to cluster dynamic resting-state data into independent functional brain states. For each condition, we performed a regression analysis to quantify function-structure similarity and the repertoire of functional brain states.

RESULTS

Seven functional brain states were clustered and ranked according to their similarity to structural connectivity, with higher ranks corresponding to higher function-structure similarity and lower ranks corresponding to lower correlation between brain function and brain anatomy. Anesthesia shifted the brain state composition from a low rank (rounded rank [mean ± SD]) in the awake condition (awake rank = 4 [3.58 ± 1.03]) to high ranks in the different anesthetic conditions (ketamine rank = 6 [6.10 ± 0.32]; moderate propofol rank = 6 [6.15 ± 0.76]; deep propofol rank = 6 [6.16 ± 0.46]; moderate sevoflurane rank = 5 [5.10 ± 0.81]; deep sevoflurane rank = 6 [5.81 ± 1.11]; P < 0.0001).

CONCLUSIONS

Whatever the molecular mechanism, anesthesia led to a massive reconfiguration of the repertoire of functional brain states that became predominantly shaped by brain anatomy (high function-structure similarity), giving rise to a well-defined cortical signature of anesthesia-induced loss of consciousness.

Mismatch negativity to predict subsequent awakening in deeply sedated critically ill patients

BACKGROUND

Mismatch negativity (MMN) is the neurophysiological correlate of cognitive integration of novel stimuli. Although MMN is a well-established predictor of awakening in non-sedated comatose patients, its prognostic value in deeply sedated critically ill patients remains unknown. The aim of this prospective, observational pilot study was to investigate the prognostic value of MMN for subsequent awakening in deeply sedated critically ill patients.

METHODS

MMN was recorded in 43 deeply sedated critically ill patients on Day 3 of ICU admission using a classical 'odd-ball' paradigm that delivers rare deviant sounds in a train of frequent standard sounds. Individual visual analyses and a group level analysis of recordings were performed. MMN amplitudes were then analysed according to the neurological status (awake vs not awake) at Day 28.

RESULTS

Median (inter-quartile range) Richmond Assessment Sedation Scale (RASS) at the time of recording was -5 (range, from -5 to -4.5). Visual detection of MMN revealed a poor inter-rater agreement [kappa=0.17, 95% confidence interval (0.07-0.26)]. On Day 28, 30 (70%) patients had regained consciousness while 13 (30%) had not. Quantitative group level analysis revealed a significantly greater MMN amplitude for patients who awakened compared with those who had not [mean (standard deviation) = -0.65 (1.4) vs 0.08 (0.17) μV, respectively; P=0.003).

CONCLUSIONS

MMN can be observed in deeply sedated critically ill patients and could help predict subsequent awakening. However, visual analysis alone is unreliable and should be systematically completed with individual level statistics.

Auditory Predictive Coding across Awareness States under Anesthesia: An Intracranial Electrophysiology Study

The systems-level mechanisms underlying loss of consciousness (LOC) under anesthesia remain unclear. General anesthetics suppress sensory responses within higher-order cortex and feedback connections, both critical elements of predictive coding hypotheses of conscious perception. Responses to auditory novelty may offer promise as biomarkers for consciousness. This study examined anesthesia-induced changes in auditory novelty responses over short (local deviant [LD]) and long (global deviant [GD]) time scales, envisioned to engage preattentive and conscious levels of processing, respectively. Electrocorticographic recordings were obtained in human neurosurgical patients (3 male, 3 female) from four hierarchical processing levels: core auditory cortex, non-core auditory cortex, auditory-related, and PFC. Stimuli were vowel patterns incorporating deviants within and across stimuli (LD and GD). Subjects were presented with stimuli while awake, and during sedation (responsive) and following LOC (unresponsive) under propofol anesthesia. LD and GD effects were assayed as the averaged evoked potential and high gamma (70-150 Hz) activity. In the awake state, LD and GD effects were present in all recorded regions, with averaged evoked potential effects more broadly distributed than high gamma activity. Under sedation, LD effects were preserved in all regions, except PFC. LOC was accompanied by loss of LD effects outside of auditory cortex. By contrast, GD effects were markedly suppressed under sedation in all regions and were absent following LOC. Thus, although the presence of GD effects is indicative of being awake, its absence is not indicative of LOC. Loss of LD effects in higher-order cortical areas may constitute an alternative biomarker of LOC.SIGNIFICANCE STATEMENT Development of a biomarker that indexes changes in the brain upon loss of consciousness (LOC) under general anesthesia has broad implications for elucidating the neural basis of awareness and clinical relevance to mechanisms of sleep, coma, and disorders of consciousness. Using intracranial recordings from neurosurgery patients, we investigated changes in the activation of cortical networks involved in auditory novelty detection over short (local deviance) and long (global deviance) time scales associated with sedation and LOC under propofol anesthesia. Our results indicate that, whereas the presence of global deviance effects can index awareness, their loss cannot serve as a biomarker for LOC. The dramatic reduction of local deviance effects in areas beyond auditory cortex may constitute an alternative biomarker of LOC.

Processing of auditory novelty across the cortical hierarchy: An intracranial electrophysiology study

Under the predictive coding hypothesis, specific spatiotemporal patterns of cortical activation are postulated to occur during sensory processing as expectations generate feedback predictions and prediction errors generate feedforward signals. Establishing experimental evidence for this information flow within cortical hierarchy has been difficult, especially in humans, due to spatial and temporal limitations of non-invasive measures of cortical activity. This study investigated cortical responses to auditory novelty using the local/global deviant paradigm, which engages the hierarchical network underlying auditory predictive coding over short ('local deviance'; LD) and long ('global deviance'; GD) time scales. Electrocorticographic responses to auditory stimuli were obtained in neurosurgical patients from regions of interest (ROIs) including auditory, auditory-related and prefrontal cortex. LD and GD effects were assayed in averaged evoked potential (AEP) and high gamma (70-150 Hz) signals, the former likely dominated by local synaptic currents and the latter largely reflecting local spiking activity. AEP LD effects were distributed across all ROIs, with greatest percentage of significant sites in core and non-core auditory cortex. High gamma LD effects were localized primarily to auditory cortex in the superior temporal plane and on the lateral surface of the superior temporal gyrus (STG). LD effects exhibited progressively longer latencies in core, non-core, auditory-related and prefrontal cortices, consistent with feedforward signaling. The spatial distribution of AEP GD effects overlapped that of LD effects, but high gamma GD effects were more restricted to non-core areas. High gamma GD effects had shortest latencies in STG and preceded AEP GD effects in most ROIs. This latency profile, along with the paucity of high gamma GD effects in the superior temporal plane, suggest that the STG plays a prominent role in initiating novelty detection signals over long time scales. Thus, the data demonstrate distinct patterns of information flow in human cortex associated with auditory novelty detection over multiple time scales.

Deep temporal models and active inference

How do we navigate a deeply structured world? Why are you reading this sentence first - and did you actually look at the fifth word? This review offers some answers by appealing to active inference based on deep temporal models. It builds on previous formulations of active inference to simulate behavioural and electrophysiological responses under hierarchical generative models of state transitions. Inverting these models corresponds to sequential inference, such that the state at any hierarchical level entails a sequence of transitions in the level below. The deep temporal aspect of these models means that evidence is accumulated over nested time scales, enabling inferences about narratives (i.e., temporal scenes). We illustrate this behaviour with Bayesian belief updating - and neuronal process theories - to simulate the epistemic foraging seen in reading. These simulations reproduce perisaccadic delay period activity and local field potentials seen empirically. Finally, we exploit the deep structure of these models to simulate responses to local (e.g., font type) and global (e.g., semantic) violations; reproducing mismatch negativity and P300 responses respectively.

Timescales of Intrinsic BOLD Signal Dynamics and Functional Connectivity in Pharmacologic and Neuropathologic States of Unconsciousness

Environmental events are processed on multiple timescales via hierarchical organization of temporal receptive windows (TRWs) in the brain. The dependence of neural timescales and TRWs on altered states of consciousness is unclear. States of reduced consciousness are marked by a shift toward slowing of neural dynamics (<1 Hz) in EEG/ECoG signals. We hypothesize that such prolongation of intrinsic timescales are also seen in blood-oxygen-level-dependent (BOLD) signals. To test this hypothesis, we measured the timescales of intrinsic BOLD signals using mean frequency (MF) and temporal autocorrelation (AC) in healthy volunteers (n = 23; male/female 14/9) during graded sedation with propofol. We further examined the relationship between the intrinsic timescales (local/voxel level) and its regional connectivity (across neighboring voxels; regional homogeneity, ReHo), global (whole-brain level) functional connectivity (GFC), and topographical similarity (Topo). Additional results were obtained from patients undergoing deep general anesthesia (n = 12; male/female: 5/7) and in patients with disorders of consciousness (DOC) (n = 21; male/female: 14/7). We found that MF, AC, and ReHo increased, whereas GFC and Topo decreased, during propofol sedation. The local alterations occur before changes of distant connectivity. Conversely, all of these parameters decreased in deep anesthesia and in patients with DOC. We conclude that propofol synchronizes local neuronal interactions and prolongs the timescales of intrinsic BOLD signals. These effects may impede communication among distant brain regions. Furthermore, the intrinsic timescales exhibit distinct dynamic signatures in sedation, deep anesthesia, and DOC. These results improve our understanding of the neural mechanisms of unconsciousness in pharmacologic and neuropathologic states.

SIGNIFICANCE STATEMENT Information processing in the brain occurs through a hierarchy of temporal receptive windows (TRWs) in multiple timescales. Anesthetic drugs induce a reversible suppression of consciousness and thus offer a unique opportunity to investigate the state dependence of neural timescales. Here, we demonstrate for the first time that sedation with propofol is accompanied by the prolongation of the timescales of intrinsic BOLD signals presumably reflecting enlarged TRWs. We show that this is accomplished by an increase of local and regional signal synchronization, effects that may disrupt information exchange among distant brain regions. Furthermore, we show that the timescales of intrinsic BOLD signals exhibit distinct dynamic signatures in sedation, deep anesthesia, and disorders of consciousness.

Neurons along the auditory pathway exhibit a hierarchical organization of prediction error

Perception is characterized by a reciprocal exchange of predictions and prediction error signals between neural regions. However, the relationship between such sensory mismatch responses and hierarchical predictive processing has not yet been demonstrated at the neuronal level in the auditory pathway. We recorded single-neuron activity from different auditory centers in anaesthetized rats and awake mice while animals were played a sequence of sounds, designed to separate the responses due to prediction error from those due to adaptation effects. Here we report that prediction error is organized hierarchically along the central auditory pathway. These prediction error signals are detectable in subcortical regions and increase as the signals move towards auditory cortex, which in turn demonstrates a large-scale mismatch potential. Finally, the predictive activity of single auditory neurons underlies automatic deviance detection at subcortical levels of processing. These results demonstrate that prediction error is a fundamental component of singly auditory neuron responses.

Great Expectations: Is there Evidence for Predictive Coding in Auditory Cortex?

Predictive coding is possibly one of the most influential, comprehensive, and controversial theories of neural function. While proponents praise its explanatory potential, critics object that key tenets of the theory are untested or even untestable. The present article critically examines existing evidence for predictive coding in the auditory modality. Specifically, we identify five key assumptions of the theory and evaluate each in the light of animal, human and modeling studies of auditory pattern processing. For the first two assumptions - that neural responses are shaped by expectations and that these expectations are hierarchically organized - animal and human studies provide compelling evidence. The anticipatory, predictive nature of these expectations also enjoys empirical support, especially from studies on unexpected stimulus omission. However, for the existence of separate error and prediction neurons, a key assumption of the theory, evidence is lacking. More work exists on the proposed oscillatory signatures of predictive coding, and on the relation between attention and precision. However, results on these latter two assumptions are mixed or contradictory. Looking to the future, more collaboration between human and animal studies, aided by model-based analyses will be needed to test specific assumptions and implementations of predictive coding - and, as such, help determine whether this popular grand theory can fulfill its expectations.

Deep temporal models and active inference

How do we navigate a deeply structured world? Why are you reading this sentence first - and did you actually look at the fifth word? This review offers some answers by appealing to active inference based on deep temporal models. It builds on previous formulations of active inference to simulate behavioural and electrophysiological responses under hierarchical generative models of state transitions. Inverting these models corresponds to sequential inference, such that the state at any hierarchical level entails a sequence of transitions in the level below. The deep temporal aspect of these models means that evidence is accumulated over nested time scales, enabling inferences about narratives (i.e., temporal scenes). We illustrate this behaviour with Bayesian belief updating - and neuronal process theories - to simulate the epistemic foraging seen in reading. These simulations reproduce perisaccadic delay period activity and local field potentials seen empirically. Finally, we exploit the deep structure of these models to simulate responses to local (e.g., font type) and global (e.g., semantic) violations; reproducing mismatch negativity and P300 responses respectively.