Skip to content

Anil K. Seth

Canadian Institute for Advanced Research, University of Sussex

54 papers in the library · 8,967 citations · publishing 2006-2026

Papers

Interoceptive inference, emotion, and the embodied self

Trends in Cognitive Sciences October 12, 2013 Anil K. Seth 1,812 citations

The brain functions as a prediction machine, a view recently revived by Bayesian brain and predictive coding theories. While this perspective has mainly been applied to external senses like vision and action, this work extends it to interoception—the perception of internal body states. The 'interoceptive inference' model proposes that subjective feelings and emotions arise from the brain's active inference about the causes of internal bodily signals. This approach generalizes appraisal theories, which hold that emotions result from cognitive evaluations of physiological changes. It also provides new insights into the neurocognitive mechanisms underlying body ownership and conscious selfhood, both in health and neuropsychiatric conditions.

An Interoceptive Predictive Coding Model of Conscious Presence

Frontiers in Psychology January 1, 2012 Anil K. Seth, Keisuke Suzuki, Hugo Critchley 971 citations

Conscious presence—the feeling of being a self located in a here and now—arises when the brain successfully suppresses interoceptive prediction errors via top-down predictions. A theoretical model links presence to the anterior insular cortex (AIC), which compares predicted versus actual bodily signals. When afferent sensory signals are judged as self-generated, predicted interoceptive signals differ from when they are externally caused, connecting presence to agency. The model integrates evidence from predictive coding, dopamine signaling, virtual reality studies, and depersonalization/derealization disorder. It proposes that emotion results from interoceptive inference and offers testable predictions for experiments manipulating both agency and presence.

Active interoceptive inference and the emotional brain

Philosophical Transactions of the Royal Society B Biological Sciences October 11, 2016 Anil K. Seth, Karl Friston 936 citations

Interoception—the perception of internal bodily states—is being reconceptualized through the lens of hierarchical inference in the brain. Bodily states are regulated by autonomic reflexes enslaved by descending predictions from deep generative models of our internal and external environment. This re-conceptualization, termed interoceptive inference, illuminates issues in cognitive and clinical neuroscience, including experiences of selfhood and emotion. The review contextualizes interoception in active Bayesian inference, highlighting its enactivist aspects, the role of uncertainty or precision in neuromodulation, implications for the functional anatomy of the emotional brain, and the role of interoception in shaping a sense of embodied self and feelings. Links are drawn between homeostasis, allostasis, cybernetic predictive control, and predictive processing. The explanatory scope ranges from autism and depression to consciousness.

Theories of consciousness.

Nature reviews. Neuroscience July 1, 2022 Anil K. Seth, Tim Bayne 836 citations

Four prominent theoretical approaches to consciousness are reviewed: higher-order theories, global workspace theories, re-entry and predictive processing theories, and integrated information theory. Each theory's key characteristics are described, including which aspects of consciousness they explain, their neurobiological commitments, and supporting empirical data. The review considers how empirical debates might distinguish among these theories and outlines three ways theories need to be developed for rigorous testing. Iterative development, testing, and comparison of these theories is expected to deepen understanding of consciousness.

Multisensory integration across exteroceptive and interoceptive domains modulates self-experience in the rubber-hand illusion

Neuropsychologia August 28, 2013 Keisuke Suzuki, Sarah N. Garfinkel, Hugo Critchley et al. 491 citations

Identifying with a body is central to being a conscious self. The classic rubber hand illusion shows that body-ownership can be modulated by timing visual and tactile feedback, and its strength relates to individual differences in sensitivity to internal bodily signals (interoception). Here, an innovative cardiac rubber hand illusion using augmented reality and feedback of cardiac information demonstrates that both subjective and objective measures of virtual-hand ownership are enhanced by cardio-visual feedback in-time with the actual heartbeat compared to asynchronous feedback. These measures correlate with interoceptive sensitivity and are also modulated by integrating proprioceptive signals via real-time visual remapping of finger movements. Interoceptive signals directly influence body ownership through multisensory integration, supporting models of conscious selfhood based on interoceptive predictive coding.

Increased spontaneous MEG signal diversity for psychoactive doses of ketamine, LSD and psilocybin

Scientific Reports April 19, 2017 Michael Schartner, Robin Carhart‐Harris, Adam B. Barrett et al. 450 citations

Measures of neural signal diversity, such as entropy and Lempel-Ziv complexity, are higher during wakeful rest than during anesthesia. In this study, these measures were computed for spontaneous magnetoencephalographic signals from humans under psilocybin, ketamine, and LSD. All three psychedelics produced reliably higher signal diversity, even after controlling for spectral changes, with the most pronounced increase in temporal (single-channel LZ complexity) rather than spatial diversity. Selective correlations emerged between changes in signal diversity and the intensity of psychedelic experience. This is the first time these measures have been applied to the psychedelic state and have yielded values exceeding normal waking consciousness, suggesting that psychedelic phenomenology constitutes an elevated level of consciousness.

A predictive processing theory of sensorimotor contingencies: Explaining the puzzle of perceptual presence and its absence in synesthesia

Cognitive Neuroscience January 21, 2014 Anil K. Seth 357 citations

Normal perception makes objects feel real and present in the world, a quality called 'perceptual presence.' Sensorimotor theories explain this by saying perception involves mastering how sensory input changes with movement, but they haven't specified the mechanism and struggle with synesthesia, where concurrents lack presence. Predictive processing theories see perception as probabilistic inference but haven't addressed presence or synesthesia. This paper proposes that generative models for perception include counterfactual elements—predictions of how sensory inputs would change with possible actions, even unperformed ones. Counterfactual richness determines perceptual presence: normal perception has rich models encoding many sensorimotor dependencies, while synesthetic concurrents arise from poor models. The theory also explains differences among dreaming, hallucination, and perception, and may reshape views on perceptual determinacy.

Theories and measures of consciousness: An extended framework

Proceedings of the National Academy of Sciences July 5, 2006 Anil K. Seth, Eugene M. Izhikevich, George N. Reeke et al. 327 citations

No single quantitative measure—neural complexity, information integration, or causal density—fully captures the multidimensional complexity of the neural systems underlying consciousness, and each has practical limitations. The analysis suggests that combining alternative measures may improve quantitative characterization, but some aspects of consciousness likely resist quantification entirely, so a satisfactory theory should blend qualitative and quantitative elements.

Consciousness in humans and non-human animals: recent advances and future directions

Frontiers in Psychology January 1, 2013 Melanie Boly, Anil K. Seth, Melanie Wilke et al. 295 citations

Consciousness research over the last decade has advanced through evidence that fronto-parietal connectivity and top-down processes enable information to travel across distant cortical areas, and through numerous dissociations between consciousness and cognitive functions such as attention in humans. Mental imagery paradigms now allow identification of covert awareness in non-responsive subjects. In non-human animals, technological enhancements have clarified the roles of cortical areas and higher-order thalamus for consciousness, and progress has been made in understanding non-vertebrate cognition relevant to possible conscious states. Major advances have also occurred in theories of consciousness and their comparison with available evidence.

Practical Measures of Integrated Information for Time-Series Data

PLoS Computational Biology January 20, 2011 Adam B. Barrett, Anil K. Seth 243 citations

Two new measures of integrated information, Φ(E) and Φ(AR), overcome limitations of the earlier Φ(DM) measure, which could rarely be applied to biological systems because it required discrete Markov dynamics. The new measures are easy to apply to time-series data, as demonstrated through simulations. They offer new opportunities for studying information integration in real and model systems and have implications for understanding consciousness and other neurocognitive processes. However, the findings also challenge theories that assign physical meaning to these measured quantities.

Expectations accelerate entry of visual stimuli into awareness

Journal of Vision June 26, 2015 Yair Pinto, Simon van Gaal, Floris P. de Lange et al. 177 citations

Expectations speed up conscious perception. In four experiments using predictive cueing and continuous flash suppression, expected images entered awareness faster than neutral or unexpected ones. This effect was not due to response priming, stimulus associations, or attentional timing biases. Top-down expectations had a stronger influence when visual input was ambiguous, consistent with predictive processing accounts. The findings support the idea that conscious access depends on verifying perceptual predictions.

Predictive processing as a systematic basis for identifying the neural correlates of consciousness

Philosophy and the Mind Sciences December 30, 2020 Jakob Hohwy, Anil K. Seth 162 citations

The search for the neural correlates of consciousness needs a systematic, principled foundation to give putative correlates greater predictive and explanatory value. The predictive processing framework for brain function is proposed as a promising candidate because it addresses three general challenges to identifying neural correlates and satisfies two constraints common to many theories of consciousness. Implementing the search through this lens can yield detailed, systematic mappings between neural substrates and phenomenological structure. The framework, precisely because it is not itself a theory of consciousness, has significant potential for advancing the neuroscience of consciousness.

Global and local complexity of intracranial EEG decreases during NREM sleep

Neuroscience of Consciousness January 27, 2017 Michael Schartner, Andrea Pigorini, Steve A. Gibbs et al. 158 citations

During non-rapid eye movement (NREM) sleep, neural dynamical complexity—measured by Lempel-Ziv complexity, amplitude coalition entropy, and synchrony coalition entropy—decreased substantially across broadly distributed brain regions in ten epilepsy patients. This decrease partially reversed during late-night NREM sleep and returned to wakeful rest levels during rapid eye movement (REM) sleep. Local brain regions mirrored this global pattern. The frontal lobe showed elevated signal complexity. These changes were not solely due to spectral power differences. The findings suggest that level of consciousness correlates with neural dynamical complexity.

Are There Islands of Awareness?

Trends in Neurosciences December 10, 2019 Tim Bayne, Anil K. Seth, Marcello Massimini 146 citations

Consciousness might persist in brain tissue that is fully causally isolated from the body and environment, forming an 'island of awareness'—a conscious state neither shaped by sensory input nor able to produce motor output. This opinion paper examines three potential cases: ex cranio brains (brains kept alive outside the body), the disconnected hemisphere after hemispherotomy surgery, and cerebral organoids (lab-grown brain tissue). The authors discuss methods for detecting consciousness in such disconnected brains, noting that current detection tools rely on sensory or motor interaction and may not apply. They argue that discovering islands of awareness would have profound ethical and legal implications and would challenge theories holding that consciousness requires embodiment or interaction with the world.

Trait phenomenological control predicts experience of mirror synaesthesia and the rubber hand illusion

Nature Communications September 25, 2020 Peter Lush, Vanessa Botan, R. B. Y. Scott et al. 136 citations

Hypnotisability—a normally distributed, stable trait—predicts experiential changes like hallucinations in response to imaginative suggestion, which feel involuntary. Such experiences can also be triggered by implicit suggestion outside hypnosis. In three large samples (156, 404, and 353 participants), substantial relationships were found between hypnotisability and experiential change in mirror-sensory synaesthesia and the rubber hand illusion, comparable to relationships with individual hypnosis scale items. The control of phenomenology to meet expectancies from perceived task requirements may explain experiential change in psychological experiments.

Rhythmic Influence of Top–Down Perceptual Priors in the Phase of Prestimulus Occipital Alpha Oscillations

Journal of Cognitive Neuroscience April 15, 2016 Maxine T. Sherman, Ryota Kanai, Anil K. Seth et al. 134 citations

Spontaneous alpha-band neural oscillations in the brain periodically transmit prior expectations to the visual cortex, biasing both objective decisions and subjective confidence before a stimulus appears. In a detection task with scalp EEG, prestimulus occipital alpha phase predicted the weighting of expectations on yes/no decisions and on confidence judgments, independent of attention. These findings suggest that alpha oscillations change the baseline from which evidence accumulation begins, shaping early visual processing and informing how expectations influence perception at the neural level.

A Deep-Dream Virtual Reality Platform for Studying Altered Perceptual Phenomenology

Scientific Reports November 16, 2017 Keisuke Suzuki, Warrick Roseboom, David J. Schwartzman et al. 111 citations

The Hallucination Machine combines deep convolutional neural networks with panoramic virtual reality to simulate visual hallucinatory experiences without drugs or psychosis. In one experiment, the system induced visual phenomenology similar to classical psychedelics. In a second experiment, the simulated hallucinations did not produce the temporal distortion typically linked to altered states. This tool allows researchers to study altered consciousness without the confounding physiological and cognitive effects of psychoactive substances or psychopathological conditions.

The strength of weak integrated information theory

Trends in Cognitive Sciences June 1, 2022 Pedro A. M. Mediano, Fernando E. Rosas, Daniel Bor et al. 109 citations

The integrated information theory of consciousness (IIT) proposes a universal mathematical formula derived from fundamental properties of conscious experience to describe the quality of consciousness. The theory is ambitious in that it aims to derive a mathematical measure of integrated information from consciousness itself, using properties of conscious experience to characterize the quality of consciousness. The text repeatedly states that IIT is a theory of consciousness that proposes a mathematical formula derived from fundamental properties of conscious experience to describe the quality of consciousness, but it does not provide specific findings, numbers, or empirical results.

The Enactive Torch: A New Tool for the Science of Perception

IEEE Transactions on Haptics October 27, 2011 Tom Froese, Marek Mcgann, William Bigge et al. 94 citations

The enactive approach to cognition emphasizes that meaningful perceptual experience arises from goal-directed bodily activity. Building on insights about practical tool-use, this work defines an 'enactive interface' as one that augments sense-making. The authors introduce the Enactive Torch, a custom-built device, and present a study of its experiential effects. Results show that the user experience is not captured by any standard perceptual modality; instead, it is a new feeling mediated by the device's design and shaped by the task situation. The findings suggest that engineering and cognitive science should collaborate to create experience-centered technology that serves as a transparent medium for augmenting natural interaction skills.

Decreased directed functional connectivity in the psychedelic state.

Neuroimage December 17, 2019 Lionel Barnett, Suresh D. Muthukumaraswamy, Robin L. Carhart-Harris et al. 88 citations

The psychedelic state involves a breakdown in patterns of information flow in the brain. Using MEG recordings, researchers measured functional connectivity after administration of LSD, psilocybin, and low-dose ketamine, and compared them with the non-psychedelic drug tiagabine. All three psychedelics decreased directed functional connectivity (information flow) throughout the brain, as measured by Granger causality. For LSD, this decrease was coupled with an increase in undirected functional connectivity (correlation and coherence), a surprising opposite movement that highlights the importance of comparing multiple measures of functional connectivity in neuroimaging data.

Split-Brain: What We Know Now and Why This is Important for Understanding Consciousness

Neuropsychology Review May 12, 2020 Edward H.F. de Haan, Paul M. Corballis, Steven A. Hillyard et al. 86 citations

Cutting the corpus callosum (split-brain) broadly disrupts functional integration across perception and attention, but some processes like action control remain unified. The key unresolved question is whether split-brain patients have split or unified consciousness. Current evidence is insufficient to decide, and future studies may need to consider intermediate conceptualizations beyond a simple yes or no.

Measuring Autonomy and Emergence via Granger Causality

Artificial Life January 12, 2010 Anil K. Seth 85 citations

Quantitative measures for autonomy and emergence, grounded in Granger causality and multivariate autoregression, are introduced and validated. G-autonomy quantifies how much a variable's past predicts its own future beyond external factors, while G-emergence measures a process's simultaneous dependence on and autonomy from its underlying causes. Applied to agent-based models, evolutionary adaptation increases autonomy in a predation model, and a flocking model demonstrates both emergence and downward causation. The work connects these measures to broader discussions of autonomy, emergence, and consciousness.

Models of consciousness

Scholarpedia January 1, 2007 Anil K. Seth 78 citations

A model of consciousness is a theoretical description that links brain properties—such as fast irregular electrical activity and widespread brain activation—to phenomenal properties like qualia, a first-person perspective, and the unity of a conscious scene. Because these properties are diverse, useful models can be either mathematical/logical or verbal/conceptual.

I overthink—Therefore I am not: An active inference account of altered sense of self and agency in depersonalisation disorder

Consciousness and Cognition April 28, 2022 Anna Ciaunica, Anil K. Seth, Jakub Limanowski et al. 76 citations

Depersonalisation disorder may arise when the brain fails to modulate sensory precision, leading to a loss of the normal sense of control over one's own perceptions and actions. The paper proposes that individuals with depersonalisation adopt the hypothesis that they are embodied perceivers who are not in control of their perception, while still recognizing that the controlling agent is themselves. This account uses predictive processing to explain the phenomenology of depersonalisation, including the feeling of being detached from one's own thoughts and actions. The analysis suggests specific psychophysical and physiological abnormalities that could underlie these experiences.

Greater than the parts: a review of the information decomposition approach to causal emergence

Philosophical Transactions of the Royal Society A Mathematical Physical and Engineering Sciences May 23, 2022 Pedro A. M. Mediano, Fernando E. Rosas, Andrea I. Luppi et al. 73 citations

Emergence—how galaxies form or consciousness arises from neurons—lacks formal tools for rigorous study. This article summarizes, elaborates, and extends a recent formal theory of causal emergence based on information decomposition, which is quantifiable and empirically testable. The theory links emergence to information about a system's temporal evolution that cannot be obtained from its parts separately. The article provides an accessible but rigorous introduction to this framework, discussing its merits in various scenarios, interpretation issues, and potential misunderstandings, highlighting the distinctive benefits of this formalism.