RESEARCH DOSSIER
Consciousness
Abstract. Consciousness is investigated as subjective experience and through empirically distinguishable constructs of global state, phenomenal content, access, and metacognitive report. This research monograph reviews the conceptual history of consciousness science, psychophysical and neural methods, global neuronal workspace theory, integrated information theory, recurrent processing, higher-order accounts, and predictive approaches. It critically evaluates studies of masking, attention, dreaming, general anesthesia, thalamocortical dynamics, and disorders of consciousness, including evidence of covert command following. Special attention is given to the 2025 preregistered adversarial collaboration comparing rival theoretical predictions. The synthesis differentiates neural correlates from prerequisites and consequences, reports from experience, and measures of causal complexity from theoretical constructs, while identifying unresolved questions about cross-state measurement, clinical interpretation, comparative consciousness, and the attribution of experience to artificial systems.
Keywords: consciousness; subjective experience; phenomenal consciousness; conscious access; neural correlates of consciousness; global neuronal workspace theory; integrated information theory; recurrent processing; higher-order theory; psychophysics; anesthesia; dreaming; disorders of consciousness; thalamocortical connectivity; covert awareness
Consciousness as a Scientific Object
Consciousness designates the presence and organization of subjective experience: the occurrence of seeing a color, hearing a sound, feeling pain, imagining an event, or undergoing a temporally extended stream of awareness. In contemporary psychology and neuroscience, the term also identifies measurable capacities and conditions associated with those experiences. A scientific account must specify its target before choosing its instruments. The level or state of consciousness describes whether and how an individual is conscious across wakefulness, sleep, anesthesia, and brain injury; the contents of consciousness concern which particular perceptions, thoughts, emotions, or bodily feelings occur within that state. Conscious access describes the availability of information for flexible reasoning, report, and action, while phenomenal consciousness names experiential character itself. The distinctions organize complementary research questions and protect interpretations from treating a single behavioral score as an exhaustive definition of experience. Seth & Bayne (2022) Koch et al. (2016)
A research dossier on consciousness integrates behavioral psychology, psychophysics, cognitive science, systems neuroscience, clinical neurology, philosophy of mind, and comparative inquiry. Each discipline supplies a different evidentiary route. Psychophysics manipulates stimulus visibility and asks how percepts change with physical input. Cognitive psychology examines report, decision, memory, and metacognitive judgment. Neuroscience measures or perturbs the circuits correlated with changes in conscious experience. Clinical investigation studies patients whose capacities for arousal, communication, and awareness separate under injury. Philosophical analysis clarifies the explanatory commitments that link first-person phenomena to third-person mechanisms. A productive synthesis does not assume that all these methods measure the same latent variable. It asks what can be inferred from each observation, what alternative explanations remain, and which findings survive comparisons among distinct tasks and populations. Dehaene et al. (2006) Mashour (2024)
Consciousness is simultaneously a phenomenon to characterize and an explanatory problem. A vivid report of seeing an object, an electrophysiological response to the object, and a successful recognition response are different observations. Their coincidence can establish correlations; their dissociation reveals constraints on proposed mechanisms. Researchers therefore investigate necessary enabling conditions, candidate sufficient neural mechanisms, content-specific representations, and broader consequences for thought and behavior. Distinguishing those tasks prevents the search for consciousness from collapsing into the search for attention, verbal output, or general intelligence. The contemporary program is particularly shaped by proposals for precise neural correlates and by methodological challenges concerning reports and task demands. Its strongest conclusions emerge when the same interpretation is supported by phenomenological assessment, carefully matched behavioral contrasts, temporal measurements, and causal intervention. Koch et al. (2016) Tsuchiya et al. (2015)
States, Contents, and Access
The state of consciousness describes a changing global condition of an organism, whereas conscious content denotes an experienced object, feeling, or thought within that condition. Wakefulness usually supports a large repertoire of contents, but the two variables can dissociate. A sleeping person may experience a dream, and a severely brain-injured patient may show preserved cognition despite limited movement or speech. Under anesthesia, responsiveness and report may be suppressed while specific neural signals remain, and their meaning requires careful interpretation. Arousal and awareness are therefore related dimensions, not interchangeable measurements. Clinical and experimental classifications benefit from recognizing intermediate states, transitions, and fluctuations rather than forcing every observation into a single awake–unconscious dichotomy. Siclari et al. (2017) Bodien et al. (2024) Mashour (2024)
Conscious access is commonly investigated through the influence of a stimulus on deliberate judgments, report, working memory, and flexible choice. The taxonomy developed by Dehaene and colleagues separates subliminal processing, preconscious processing, and consciously accessible processing in relation to stimulus strength and attentional amplification. This makes clear why a stimulus may affect subsequent responses even when a participant denies seeing it. Yet operational access remains a particular target: a participant can possess an experience that is difficult to describe, and some neural events associated with reporting may arise downstream of the experience under study. The question is accordingly whether the experiment measures experience, reportability, or a combination. Such differences shape comparisons between global-workspace, recurrent-processing, and higher-order accounts. Dehaene et al. (2006) Kouider & Dehaene (2007) Tsuchiya et al. (2015)
A third level concerns the structure of contents themselves. Visual scenes contain objects and spatial relationships; bodily awareness includes interoceptive and affective qualities; autobiographical thinking can combine imagery, memory, and a sense of oneself across time. These modalities differ in stimulus pathways and representational demands but share a requirement that observations be tied to a specified target experience. Researchers can compare a reported face with a reported object, dream imagery with waking perception, or remembered pain with current sensory input. Such comparisons should not assume that all conscious contents share identical local neural machinery. The methodological problem is to identify what remains constant across modality, task, and state while retaining the specificity of individual experiences. Koch et al. (2016) Seth & Bayne (2022) Siclari et al. (2017)
Historical Foundations of Consciousness Research
Scientific reflection on conscious experience preceded contemporary brain imaging. Nineteenth-century introspective psychology made experience a central subject but encountered fundamental problems in standardizing verbal descriptions and separating experience from expectations about it. The rise of behaviorally operational methods later emphasized observable responses and learning, establishing tools for reproducible measurement even as first-person phenomena remained difficult to operationalize. Cognitive psychology reopened the investigation of internal representations through experimental performance and computational explanation. Modern consciousness science developed where these approaches intersected with neurological lesions, psychophysics, sleep research, anesthesia, and recording methods able to track activity across distributed brain networks. The historical progression explains why awareness is studied through several partly independent operational traditions. Chalmers (1995) Dehaene et al. (2006)
In the 1990s, David Chalmers distinguished mechanistic questions about discrimination, integration, and report from the philosophical question of why physical processing is accompanied by experience. His formulation sharpened a debate about explanatory adequacy without providing a behavioral test that alone settles the issue. During the same period, workspace, recurrent-processing, and neural-correlate programs sought specific cognitive and neural signatures. Subsequent studies of visual masking, attentional blink, and perceptual rivalry supplied experimentally manipulable contrasts between seen and unseen stimuli. Their development was cumulative: theories increasingly made predictions about processing stages, temporal signatures, cortical participation, and the conditions under which a representation becomes accessible to broader cognitive systems. Chalmers (1995) Lamme & Roelfsema (2000) Sergent et al. (2005)
The subsequent history includes a transition from theory-specific demonstrations to direct comparative testing. Noninvasive imaging made whole-brain spatial mapping possible; magnetoencephalography and electroencephalography supplied finer temporal descriptions; intracranial recordings afforded detailed access to selected populations. Perturbational techniques introduced stronger causal and mechanistic tests. In 2025 an adversarial collaboration jointly preregistered predictions for global neuronal workspace and integrated information theories, and tested them with converging modalities. Historical significance here attaches to the development of methods capable of challenging favored explanations through shared data rather than merely accumulating theory-consistent observations. Continuing progress depends on how rigorously measurement assumptions, selection of contrasts, and test interpretation are specified in advance. Cogitate Consortium et al. (2025) Casali et al. (2013)
Experience, Report, and the Measurement Problem
The measurement problem begins with a basic asymmetry: an investigator can observe behavior and neural activity directly but normally receives evidence of another person's experience through communication or indirect indicators. Subjective reports are indispensable in many psychophysical paradigms because they identify the reported contents under comparison. Their interpretation nonetheless requires attention to demand characteristics, response criteria, memory, decisional confidence, and linguistic capacity. A visibility rating is a response generated by multiple processes; changes in rating may result from a change in experience, a changed criterion, or changed memory for an experience. Researchers therefore combine continuous and categorical scales, trial-level performance, confidence judgments, and models that estimate perceptual sensitivity separately from response bias. Fleming & Lau (2014) Tsuchiya et al. (2015)
No-report paradigms seek to isolate perceptual awareness from the neural operations required for explicit reporting. Eye movements, pupil responses, reflexive behavior, and spontaneous perceptual alternation can sometimes identify the content likely being experienced without a simultaneous verbal answer. The methodological gain is a reduction in task-related planning and motor signals. The cost is dependence on an independent validation of the indirect marker. A pupil change correlated with a percept is not a logically infallible certificate that the percept occurred. The most useful strategies compare reported and unreported conditions, quantify the marker's accuracy, and assess whether the inferred neural activity remains associated with consciousness after report-specific contributions are controlled. Tsuchiya et al. (2015) Boly et al. (2017)
Metacognition adds another layer because confidence in perception concerns how an individual evaluates a first-order judgment. Signal-detection approaches can estimate whether confidence tracks correctness while accounting for task performance. A person may respond accurately with low confidence, or express high confidence in an error; neither outcome by itself resolves whether perceptual experience was present. The distinction matters when studying blindsight, perceptual uncertainty, and states in which language is unavailable. No measure has universal validity across all tasks, ages, species, and clinical conditions. Construct validity comes from a network of converging evidence and explicit modeling of where measurement may fail. A theory of consciousness should therefore predict both the observable indicators and the conditions under which those indicators cease to be trustworthy. Fleming & Lau (2014) Lau & Rosenthal (2011)
Experimental Inference and Neural Correlates
A neural correlate of consciousness is a measurable neural event associated with a specified conscious condition or content. The influential formulation of minimal neural mechanisms jointly sufficient for particular experiences makes the target more precise than a general list of active brain areas. Researchers distinguish full neural correlates supporting conscious experience across contents from content-specific neural correlates and from enabling background conditions such as arousal. A study comparing conscious with unconscious perception must additionally consider prerequisites for perception and consequences of report. For example, frontal activity appearing after a participant recognizes a stimulus may participate in decision or communication instead of constituting the sensory experience. Anatomical location, timing, and task dependencies thus require separate investigation. Koch et al. (2016) Boly et al. (2017)
Experimental contrasts can be constructed by adjusting stimulus strength, masking, attentional allocation, rivalry, or state transitions. Matching sensory input while experience varies can control important confounds, but the resulting conditions may still differ in attention, vigilance, expectation, and decision criteria. Parametric designs estimate changes across graded visibility rather than selecting only extreme responses. Multivariate decoding tests whether recorded patterns contain information about content; decoding success alone does not establish causal participation in consciousness because decodable activity may belong to preparatory or downstream processing. Converging evidence becomes stronger when activity predicts a consciously reported content across independently replicated tasks and when perturbation produces a selective, interpretable alteration of the target experience. Sergent et al. (2005) Cogitate Consortium et al. (2025)
Causal inference remains difficult because intervening on a neural region may affect sensory capacity, attention, memory, arousal, and motor output together. Lesions may have widespread network consequences; electrical stimulation may spread beyond a proposed circuit; anesthetics change several aspects of brain dynamics simultaneously. The inference from an intervention to a specifically conscious mechanism therefore depends on the selectivity of the change and on the quality of the behavioral and phenomenological assessment. A rigorous research program records alternative causal pathways and distinguishes correlational localization from interventions that test mechanisms. The problem is particularly visible in clinical consciousness research, where failure to respond can reflect damaged output systems rather than absent awareness. Schiff et al. (2007) Owen et al. (2006) Mashour (2024)
Global Neuronal Workspace Theory
Global neuronal workspace theory explains conscious access through the selective amplification and widespread availability of information within a distributed cortical system. A stimulus may be analyzed locally without entering the workspace, while sufficient amplification produces a state in which its content can influence reasoning, flexible action, and report. Accounts associated with Dehaene, Changeux, and colleagues connect this transition to nonlinear ignition and long-range integration. The proposal is computationally explicit in a useful sense: it distinguishes local processing from the broader functional accessibility of information and expects detectable changes in activity and connectivity under matched sensory inputs. Its scope is especially clear for access consciousness, the capacity to make a stimulus available to several cognitive operations. Dehaene et al. (2006) Mashour et al. (2020)
An important empirical tradition employs backward masking and attentional blink. Sergent, Baillet, and Dehaene measured event-related potentials during an attentional-blink task in which identical second targets were sometimes reported and sometimes missed. Early responses could remain despite failure of conscious access, whereas later differences supported an account of a more extended processing transition. Such findings are consistent with the claim that elementary sensory activation is insufficient for reliable report. Yet they also depend on which processes contribute to detecting and reporting the target. A later frontal signal may reflect global availability, task monitoring, or a consequence of making a decision. Disentangling these possibilities requires no-report measures and independently specified temporal predictions. Sergent et al. (2005) Tsuchiya et al. (2015)
The theory's scientific strength lies in its operational predictions about broadcast, ignition, and flexible availability. Its central challenge is to identify exactly which components are necessary for subjective contents themselves and which implement access to or use of those contents. The preregistered Cogitate collaboration tested several predictions in a common protocol and observed evidence that qualified aspects of global-workspace and integrated-information formulations. The experiment did not eliminate the broad idea that global availability is relevant to conscious cognition, but it constrained specific predicted patterns of neural maintenance and interregional exchange. Treating the results as decisive proof or complete refutation would exceed the design's discriminating power. Theory development must specify how task context and content modality alter expected workspace signatures. Cogitate Consortium et al. (2025) Seth & Bayne (2022)
Integrated Information Theory
Integrated information theory begins with proposed structural properties of experience, including differentiation and unity, and relates them to the causal organization of a physical system. Tononi's 2004 formulation sought to explain why consciousness is associated with both a rich repertoire of possible states and their integration. Later formalizations, including the version presented by Oizumi, Albantakis, and Tononi, attempt to derive physical conditions from phenomenological axioms. This research program asks not merely where information is represented but whether the system possesses an irreducible intrinsic causal structure. Its core concepts therefore differ from ordinary measures of behavioral complexity or mutual information in recorded signals. Tononi (2004) Oizumi et al. (2014)
Empirical implications of integrated-information approaches have guided analyses of differentiated yet coordinated brain activity across wakefulness, sleep, and anesthesia. Perturbational complexity indices examine how a direct brain perturbation propagates and produces a distributed response. Casali and colleagues proposed an index derived from transcranial magnetic stimulation and electroencephalography with the aim of assessing consciousness independently of sensory inputs and overt behavior. The index is a practical indicator under studied conditions, not a direct calculation of the full theoretical quantity postulated by integrated information theory. Keeping that separation explicit is essential: an operational classifier may predict clinical states even if its proposed theoretical interpretation remains contested. Casali et al. (2013) Massimini et al. (2005)
The causal structure required by integrated information theory is difficult to estimate for large, realistic neural systems, and its theoretical commitments generate disagreements about what should count as a conscious substrate. Empirical data concerning posterior cortical content representations and long-distance interactions can constrain claims about particular anatomical implementations, while leaving formal assumptions open to further scrutiny. The 2025 adversarial comparison reported both support and challenges for selected predictions, including content localization and temporal maintenance, rather than a simple winner. Continued evaluation requires specifying which derived measures belong to a particular formal version, what observations would decisively contradict it, and how predictions travel from a simplified causal model to complex biological recording systems. Cogitate Consortium et al. (2025) Seth & Bayne (2022)
Recurrent Processing and Local Cortical Dynamics
Recurrent-processing theories emphasize interactions that return information to earlier processing stages rather than proceeding only through a feedforward sweep. Lamme and Roelfsema distinguished rapid initial feedforward processing from later feedback and horizontal interactions that modify representations in visual cortex. On this account, a sufficiently structured recurrent exchange can be critical to perceptual experience, while widespread prefrontal deployment may serve attention, report, or deliberate use. The framework helps explain why a stimulus can induce activity at multiple sensory processing levels without becoming visible to its observer. It generates questions about the timing, anatomical scale, and circuit specificity of feedback interactions under controlled viewing conditions. Lamme & Roelfsema (2000) Lamme (2006)
A strength of the recurrent approach is its close relation to visual physiology and perceptual tasks. Electrophysiological studies can distinguish initial evoked activity from later recurrent components; masking can disturb subsequent interactions while preserving parts of the initial response. Yet a local recurrent signal also participates in attention, prediction, object recognition, and task adaptation. Recurrent activity is therefore not automatically evidence of experience, and recurrence can occur when a participant denies seeing the stimulus. Rigorous tests compare variations of the same content across awareness conditions, examine the effect of perturbing feedback selectively, and evaluate whether recurrent mechanisms are sufficient for the experiences claimed by the theory. Kouider & Dehaene (2007) Tsuchiya et al. (2015)
The tension between local recurrence and global broadcast should be stated as a question of explanatory scale. A percept may depend on interactions within sensory circuits and later become accessible through broader networks, while different theories assign the crucial experiential transition to different parts of that sequence. Evidence for sensory recurrence does not alone exclude workspace participation, and evidence for frontal report signals does not establish a frontal substrate of phenomenal content. A productive experiment places theory-specific predictions at matched temporal and functional levels. It also distinguishes conditions in which report is required from those in which the conscious content can be inferred through independently validated markers. Boly et al. (2017) Cogitate Consortium et al. (2025)
Higher-Order Accounts and Metacognitive Representation
Higher-order theories locate an important condition of conscious awareness in representations of one's own mental states. On this approach, merely processing a stimulus can occur without awareness, while representing oneself as occupying the relevant perceptual state contributes to its conscious status. Lau and Rosenthal surveyed empirical findings interpreted as supporting higher-order accounts and highlighted the need to distinguish their predictions from those of first-order and global-workspace theories. Theories in this family vary in how they characterize higher-order representations, their neural implementation, and the relationship between awareness and reflective self-knowledge. It is consequently misleading to infer a single prediction from every version of the approach. Lau & Rosenthal (2011) Seth & Bayne (2022)
Metacognitive experiments can estimate the relation between task accuracy and confidence. Their results offer an indirect route to questions about internal monitoring, introspective access, and the transformations between perceptual evidence and judgments about experience. However, confidence sensitivity, metacognitive bias, and subjective awareness are distinct measurement targets. A participant may experience a faint stimulus but lack accurate information about the reliability of the resulting decision. Some higher-order accounts treat such dissociations as theoretically significant; other frameworks interpret them as downstream evaluation differences. The practical task is to formulate tests that would isolate a higher-order representation while preserving comparable first-order sensory evidence across conditions. Fleming & Lau (2014) Lau & Rosenthal (2011)
Anatomical debates about prefrontal cortex are especially relevant to these models. Prefrontal areas contribute to executive function, decision making, working memory, and metacognitive judgments, all of which frequently co-occur with reports of consciousness. The presence of those signals establishes participation in the task but does not settle whether they constitute experience. Lesion evidence, stimulation, no-report contrasts, and temporally resolved recording offer complementary tests, each with limitations. The research aim is to establish whether a particular higher-order operation is necessary for a particular form of awareness, while controlling how damage or intervention may independently change the capacity to express that awareness. Boly et al. (2017) Koch et al. (2016)
Predictive Processing and Theoretical Integration
Predictive-processing approaches treat perception as inference involving predictions, sensory evidence, and the updating of internal models. Such formulations offer candidates for understanding perceptual stability, uncertainty, and the relation between bodily conditions and experienced contents. They do not by themselves specify a universal threshold at which inference becomes conscious. Theoretical work must identify additional properties of prediction and error processing that distinguish conscious from unconscious representations. Seth and Bayne group predictive and re-entry approaches among major contemporary lines of investigation while emphasizing that the theories often address different explanatory targets and may be only partly comparable. Seth & Bayne (2022)
Integrative research asks how content-specific coding, recurrent exchange, global availability, internal monitoring, and causal differentiation relate across time. Several processes can occur in one conscious task without being identical mechanisms. A reported visual object, for example, may require early visual processing, recurrent stabilization, selection under attention, a decision criterion, and a motor response. Disentangling these contributions is an empirical matter and may reveal explanatory roles at several levels. The relative necessity of each component could depend on whether the target is seeing an object, remembering it, communicating it, or consciously reflecting on having seen it. Integration should therefore preserve distinct operational hypotheses rather than merging theories into a verbal supermodel. Mashour et al. (2020) Tsuchiya et al. (2015)
Theories are most informative when they identify observations that could count against their preferred claims. Examples include whether frontal activity contributes uniquely to decoding content, whether information persists throughout an experienced stimulus, whether particular connectivity patterns distinguish states, and whether a perturbation changes awareness without changing lower-level sensory performance. Pre-registration and adversarial collaboration provide a method for translating broad explanatory language into shared, risky predictions. Scientific synthesis then compares the actual inferential reach of each test with the conclusion drawn from it. The result is a structured account of what the evidence supports, where independent theories overlap, and which mechanisms remain underdetermined. Cogitate Consortium et al. (2025) Seth & Bayne (2022)
Visual Awareness, Masking, and Thresholds
Visual awareness is experimentally tractable because investigators can manipulate luminance, duration, contrast, and interruption of a stimulus while asking which aspects enter reported experience. Backward masking presents a target followed by another stimulus that interferes with subsequent processing. Different timing conditions can preserve early visual responses and certain priming effects while reducing explicit visibility. Kouider and Dehaene reviewed evidence that masked stimuli can influence lexical, semantic, and response-related operations, which demonstrates that substantial computation can occur without reliable report. The degree and significance of unconscious processing depend on the masking procedure, sensitivity of the awareness measure, and the possibility that a participant possesses partial information about a target. Kouider & Dehaene (2007) Dehaene et al. (2006)
Threshold experiments compare graded changes in visibility with neural responses and decision behavior. A sharp transition in report does not necessarily imply an equally sharp transition in the underlying experience; categorical responses can arise from decisional thresholds applied to continuous evidence. Likewise, a gradual neural response may coexist with a more categorical cognitive access event. The aim is to fit models of perception, sensitivity, and judgment rather than read phenomenology directly from a binary keypress. Temporal methods are particularly valuable because they can separate early sensory encoding from later recurrent and integrative activity. Experiments should also account for attention, stimulus expectation, and trial history, which may influence both performance and neural signatures. Sergent et al. (2005) Fleming & Lau (2014)
An informative contrast asks whether consciously seen and unseen trials differ while the external stimulus is physically identical. The strength of this design lies in matching input; its weakness is that internal attention or vigilance may still fluctuate between trials. The same event can produce a remembered report after several seconds or an immediate response, introducing different memory and decision requirements. No-report markers offer complementary evidence, but their validity must be established against conditions in which reports remain possible. Across protocols, the safest inference concerns a specified mode of visual experience in the studied population rather than consciousness in all its forms. The accumulation of converging visual studies supplies valuable constraints without settling every question about thought, affect, or self-awareness. Tsuchiya et al. (2015) Koch et al. (2016)
Attention, Conscious Access, and Perceptual Competition
Selective attention prioritizes information for processing under limited capacity, while consciousness concerns what enters experience. Their interaction is extensive because attended stimuli often become easier to report, but attentional selection and conscious perception are operationally separable. Masked stimuli can guide subsequent behavior without a positive awareness report, and visible stimuli can be experienced without being the object of deliberate focal attention. An experimental claim that attention is independent of consciousness must specify which form of attention, which sensory modality, and which awareness assessment it uses. Otherwise, a dissociation may involve changes in effort, report confidence, or the available time for integrating information. Dehaene et al. (2006) Lamme (2006)
Perceptual competition allows physically stable sensory input to support alternating experiences. During binocular rivalry, different images are presented to the two eyes, and participants may report changes in dominant perception. Neural activity correlated with perceptual alternation can be compared with constant input, providing a route to content-specific correlates. Yet voluntary reporting can also generate temporal signals at each switch. No-report approaches infer alternations from oculomotor and related indicators, reducing some reporting demands while introducing validation questions. The research problem is to identify neural signals that follow the experienced content regardless of the method by which a switch is registered. Tsuchiya et al. (2015) Koch et al. (2016)
Attention also affects the stability and resolution of conscious content. An observer may consciously experience the overall scene while failing to individuate a peripheral object, and the limits of later recall can be mistaken for limits of momentary experience. This distinction matters in change detection, attentional blink, and divided-attention tasks. Stronger designs compare immediate and delayed response conditions and explicitly model the role of working memory. They ask when attention changes the content itself, when it changes access to an existing content, and when it only changes the reliability of subsequent judgment. The conclusions are about measurable functional relations between attention and awareness, with no need to impose a single binary identity between them. Sergent et al. (2005) Mashour et al. (2020)
Sleep, Dreaming, and Conscious Experience
Sleep demonstrates that behavioral disengagement from the external environment can coexist with rich internal experience. Dream reports obtained after awakenings provide evidence of visual imagery, emotion, bodily experience, and narrative organization in sleeping participants. The relation between stage of sleep and dreaming is probabilistic rather than an absolute equivalence: experiences occur across several physiological conditions, and the absence of a later dream report may reflect forgetting. Research therefore benefits from timed awakenings, physiological monitoring before awakening, and standardized questions that distinguish having had an experience from remembering its details. A global measure of responsiveness is inadequate to characterize the full set of conscious contents associated with sleep. Siclari et al. (2017) Mashour (2024)
Massimini and colleagues used transcranial magnetic stimulation with electroencephalography to investigate changes in cortical effective connectivity during sleep. Their findings showed that an evoked response could be more local and less effectively integrated in certain sleep conditions than during wakefulness. Such experiments supplied a causal perturbational perspective that differs from recording spontaneous rhythms alone. They do not imply that every sleeping brain is devoid of consciousness, because dreams and sleep-stage transitions complicate any simple association. The findings instead identify a relationship between particular patterns of network responsiveness and the global conditions under which sustained external awareness is common. Massimini et al. (2005) Casali et al. (2013)
Siclari and colleagues studied neural activity preceding reports of dreaming and identified associations involving posterior cortical activity and the presence or contents of subsequent experience reports. The results connect a specific phenomenological question to spatially and temporally defined physiological measurements. Their interpretation remains subject to the fact that reports occur after awakening and to differences between dream production and dream memory. Nonetheless, sleep research is a crucial natural experiment: it shows how conscious experience can vary while sensory input, motor output, and environmental engagement are comparatively constrained. Studies of dreaming also help separate state changes from changes in what is experienced within a given state. Siclari et al. (2017) Koch et al. (2016)
Anesthesia and Controlled Transitions
General anesthesia provides experimentally and clinically important changes in responsiveness, memory formation, sensory processing, and global neural dynamics. Different anesthetic agents act through different molecular and network pathways, so consciousness cannot be explained by a single interchangeable drug effect. Researchers examine induction, maintenance, emergence, and possible dissociations between behavioral unresponsiveness and neural indicators of processing. Mashour's review relates these observations to arousal mechanisms, large-scale functional connectivity, and the organization of sensory systems. Interpretation requires special care because anesthetics can impair movement or memory independently of whether a particular experience occurs during the procedure. Mashour (2024)
Comparing anesthetized and waking states can reveal changes in spontaneous oscillations, effective connectivity, and responses to stimulation. Such patterns contribute to practical monitoring and theory development, but each marker must be validated against defined endpoints. A classifier that predicts behavioral responsiveness may be highly useful clinically while remaining an imperfect measure of subjective experience. Perturbational complexity, developed in work by Casali and colleagues, illustrates a strategy less dependent on sensory stimulation or an overt answer. Its empirical significance must be evaluated by testing across agents, patient groups, and states rather than by assuming theoretical specificity from the index's name. Casali et al. (2013) Mashour (2024)
Experimental anesthesia also tests whether background enabling conditions and specific content-generating mechanisms can be distinguished. Arousal systems can maintain the excitability that permits complex cortical interactions; changes in those systems may alter the overall probability of experience. Local sensory responses may persist without wide accessibility, while widespread activity can be modified by drugs that change the effective coupling of neural networks. No single observation proves that all components of consciousness have disappeared. Future evidence must combine accurate pharmacological descriptions, monitoring of sleep–wake-like transitions, independent neural measures, and where ethically and technically feasible post-state reports. The resulting model is of interacting pathways that affect different components of consciousness under controlled conditions. Mashour (2024) Koch et al. (2016)
Neural Timing, Connectivity, and Information Dynamics
Conscious perception develops over time, and its neural study requires careful alignment of stimulation, processing, report, and subsequent memory. Electroencephalography and magnetoencephalography provide fine temporal resolution and can identify changes in event-related activity, oscillations, and distributed synchrony. Functional magnetic resonance imaging maps spatially distributed blood-oxygenation responses with different temporal constraints, while intracranial recordings directly sample selected neural populations with a precision unavailable to whole-brain noninvasive methods. Each modality has characteristic limits concerning coverage, temporal inference, noise, and the relationship between measured signals and underlying computation. Multimodal convergence is therefore more persuasive than treating any one recording method as a transparent image of subjective experience. Cogitate Consortium et al. (2025) Koch et al. (2016)
Long-range synchronization and connectivity have been proposed as signatures of conscious integration. Correlation between distant neural signals shows patterned co-variation; it does not automatically reveal the direction of influence or whether the relationship is causally necessary. Effective-connectivity methods and perturbation add information about interactions, but their interpretations depend on model assumptions and intervention selectivity. Findings during sleep suggest that the spatiotemporal propagation of responses changes with state, while perceptual studies show that content-related information can occupy distinct cortical territories. A careful account distinguishes distributed representation, communication, recurrent influence, and large-scale causal integration rather than treating them as equivalent measurements. Massimini et al. (2005) Casali et al. (2013)
The challenge of neural timing became especially concrete in the preregistered Cogitate project. Participants encountered visible stimuli with systematically varied presentation durations while investigators recorded fMRI, MEG, and intracranial EEG. The consortium evaluated predicted content representations, activity patterns, and connectivity under a common set of contrasts. Results revealed content information in visual, ventrotemporal, and inferior frontal regions, but the detailed support for predicted localization, maintenance, and synchronization was mixed. The study is important as a demanding test of particular registered predictions and of the feasibility of theory-neutral collaboration. Its findings require interpretation at the level of those predictions rather than a universal classification of every conceivable model of consciousness. Cogitate Consortium et al. (2025)
Thalamocortical and Brainstem Conditions
Brainstem and thalamic systems are central to the regulation of arousal and to the conditions that enable sustained cortical processing. Their contribution must be characterized at several levels, including ascending activating pathways, thalamocortical exchange, cortical recurrence, and large-scale state transitions. The distinction between background conditions and the neural implementation of a particular conscious content is useful but not absolute: thalamic circuits may participate in more than state regulation. Contemporary work on the thalamus reviews both its role in changing overall conscious state and possible contributions to the specificity of contents. An adequate account considers anatomical nuclei, their targets, and the differences between perturbing arousal and perturbing an individual representation. Koch et al. (2016) Whyte et al. (2024)
Thalamic lesions and stimulation studies have provided valuable causal evidence, although their interpretation is complicated by distributed consequences of intervention. Schiff and colleagues reported behavioral improvements after central thalamic stimulation in a patient with severe traumatic brain injury. The result demonstrates that altering a selected neural circuit can influence clinically meaningful behavior in a particular context, but it does not establish that the same intervention will benefit every disorder of consciousness. Translating such observations into a general neural account requires replication, carefully specified patient selection, longitudinal assessment, and mapping of how stimulation influences wider networks. Schiff et al. (2007) Mashour (2024)
The relation between cortex and subcortical structures also cautions against anatomical exclusivity. Consciousness research may find content-discriminating activity in posterior cortex while identifying enabling, coordinating, or modulatory processes in thalamic and brainstem circuits. Distinguishing those contributions is a question about causal organization across levels rather than about choosing a solitary anatomical headquarters. Converging lesion, stimulation, sleep, anesthesia, and recording data are particularly important because different interventions expose different vulnerabilities. The resulting evidence suggests an organized system with differentiated causal roles, while leaving open the exact mechanisms sufficient for each kind of experience. Whyte et al. (2024) Koch et al. (2016)
Disorders of Consciousness and Covert Cognition
Severe brain injury can dissociate awareness from the motor and language abilities commonly used to assess it. Traditional bedside examination relies heavily on the patient's capacity to demonstrate responses to commands, track objects, or communicate. When movement, attention, hearing, comprehension, or sustained vigilance is impaired, a negative observed response can be ambiguous. Disorders of consciousness are accordingly studied through repeated standardized behavioral assessment and, when suitable, supplementary physiological or imaging methods. The fundamental clinical-scientific question is what can be established about residual awareness under conditions in which the usual communication pathways are unreliable. Owen et al. (2006) Bodien et al. (2024)
Owen and colleagues reported in 2006 that a patient behaviorally diagnosed as being in a vegetative state generated task-dependent fMRI patterns during instructed mental imagery. The importance of this finding lay in the possibility that deliberate modulation of neural activity could reveal command following when overt movement was absent. Monti and colleagues extended the paradigm in a 2010 study of 54 patients with disorders of consciousness and demonstrated that a subset could generate task-specific responses. These paradigms require intact comprehension, sustained attention, and the ability to maintain mental imagery; a negative result does not prove absence of awareness. Positive findings in appropriate controls and conditions provide evidence of residual volitional cognitive capacity. Owen et al. (2006) Monti et al. (2010)
More recent multicenter research has broadened the investigation of cognitive motor dissociation. Bodien and colleagues examined a convenience sample of 353 adults with disorders of consciousness using clinical, behavioral, and task-based fMRI or EEG measures. The study supports the existence of covert command following in some behaviorally unresponsive patients while also showing the need to account for heterogeneous injuries, examination protocols, and test availability. Diagnostic inference depends on the properties of each measure and the population studied. These findings have consequences for scientific models of consciousness and for how uncertainty is communicated in clinical care; they do not justify treating one experimental signal as a substitute for comprehensive specialist assessment. Bodien et al. (2024) Claassen et al. (2019)
Perturbation, Complexity, and Clinical Indicators
Clinical consciousness research needs indicators that remain informative when the patient cannot describe an experience. Electroencephalographic rhythms, responses to sound, network connectivity, and perturbational measures provide complementary views of neural function. The perturbational complexity index proposed by Casali and colleagues measures properties of the brain's response after transcranial magnetic stimulation. It was designed to identify differentiated and integrated response patterns without requiring verbal report or an external sensory input. This offers an important methodological advance because a signal elicited under controlled perturbation can reveal properties that spontaneous recordings may not expose. Its reliability and diagnostic applicability still depend on the investigated groups, recording conditions, and validation procedures. Casali et al. (2013) Massimini et al. (2005)
Imaging-based command following and spontaneous measures answer different questions. A mental-imagery protocol asks whether an individual can intentionally follow instructions, thereby supporting an inference about covert cognitive capacity when a robust response is detected. A resting-state or perturbation measure assesses features of brain organization that can be associated with different states, but it need not demonstrate deliberate comprehension. Combining these techniques may improve the scope of assessment, provided that their evidentiary roles remain distinct. Apparent agreement between several markers can reflect a shared underlying arousal change rather than independent validation of subjective experience. Investigators therefore need prespecified thresholds, control conditions, and transparent reporting of indeterminate findings. Monti et al. (2010) Casali et al. (2013)
For therapeutic and prognostic claims, observational associations must be separated from causal evidence of benefit. A case report of stimulation-induced behavioral change raises a mechanistic possibility and may motivate controlled trials; it is not an estimate of average treatment effect. Diagnostic classification and outcome prediction are equally susceptible to selection bias, time-dependent recovery, and variation in rehabilitation. Reviews and prospective cohorts should be assessed for external validity and confidence in their measurement procedures. The present monograph discusses these studies as scientific evidence concerning consciousness and its indicators; intervention decisions require individualized specialist evaluation beyond the scope of an encyclopedic research dossier. Schiff et al. (2007) Bodien et al. (2024)
Developmental and Comparative Questions
Consciousness is investigated across development through changes in sensory responsiveness, learning, integrated behavior, and the development of reliable reporting capacities. Infants and young children cannot always describe experiences in the terms available to adults, so developmental experiments use eye movements, looking behavior, physiological responses, and age-appropriate behavioral choices. These are indicators of processing and possible experiential capacities, with their interpretation depending on converging evidence. The absence of sophisticated language cannot independently establish the absence of subjective experience. Developmental comparisons instead ask which mechanisms support identifiable experiences at different stages and how the requirements for report and reflective self-knowledge change over time. Seth & Bayne (2022) Koch et al. (2016)
Comparative animal research presents a related problem because communicative abilities, sensory ecology, and neural organization differ among species. Behavioral flexibility, perception, pain-related responses, and complex learning provide important evidence about capacities that may accompany experience, but none of these observations supplies a universally accepted single test. Neurobiological homologies and analogies can strengthen interpretation when the functions of the relevant circuits are well characterized. Scientific standards require specifying the inferred capacity, evaluating alternative explanations for behavior, and distinguishing the presence of a general arousal system from evidence about particular conscious contents. A responsible comparative research program accommodates uncertainty without equating human-style report with the definition of consciousness. Seth & Bayne (2022)
Comparative conclusions must also avoid premature extensions of a theory calibrated on adult human visual tasks. A proposed frontal ignition signature may be difficult to assess in animals with different cortical organization; a phenomenology-based causal model may apply beyond familiar anatomy but require new empirical mappings. Developmental and cross-species comparisons therefore test the generality of theoretical principles and expose assumptions hidden in standardized human protocols. These comparisons are valuable precisely because no single model is guaranteed to remain invariant across differences in anatomy, behavior, and environmental niche. Independent behavioral ecology, neurophysiology, and perturbation studies provide the strongest basis for cautious generalization. Lamme & Roelfsema (2000) Seth & Bayne (2022)
Artificial Systems and the Boundaries of Consciousness Attribution
Artificial systems extend the scientific question of consciousness beyond familiar biological examples. An artificial agent can exhibit fluent language, flexible planning, memory-like persistence, or metacognitive statements; these functions can be measured without concluding that the system has subjective experience. Functional performance and phenomenal attribution are separate claims that require different evidentiary arguments. The major theories of consciousness offer distinct candidate requirements, including global availability, recurrent integration, higher-order representation, and intrinsic causal organization. Applying them to artificial architectures requires identifying which operations or causal relations are present, whether their implementation matches the theory's commitments, and what observations could discriminate alternative explanations. Seth & Bayne (2022) Oizumi et al. (2014)
The relevant method is comparative, operational, and explicit about inference. A network's internal signals can be examined for information accessibility, persistence, functional feedback, causal interventions, and internally generated monitoring. Those measurements may establish capacities useful for artificial cognition while leaving the character or existence of subjective experience unresolved. The same discipline applies in the opposite direction: biological consciousness cannot be defined solely by human conversational behavior because clinical evidence already documents cases where intentional cognition is decoupled from ordinary output. In both cases, researchers must distinguish a lack of report, a failure of a particular instrument, and a justified conclusion about an experiential state. Owen et al. (2006) Bodien et al. (2024)
Within the Artificial Era, the central philosophical and social stakes include how humans interpret artificial agency and how artificial systems affect human awareness, attention, identity, and dependence. These research questions interact with consciousness science but do not collapse into a verdict about artificial experience. Psychology can study anthropomorphism, trust, cognitive partnership, and attribution using observable behavior and human reports, while theoretical work defines the additional claims necessary for an account of sentience. The result is a research program that preserves the independent scientific subject of human consciousness and allows carefully delimited comparisons with artificial processes. Philosophical proposals about Artificial are treated as proposals when they extend beyond currently established neurobiological findings. Seth & Bayne (2022) Fleming & Lau (2014)
The 2025 Adversarial Comparison
The 2025 Cogitate Consortium study is a methodological landmark because rival theoretical groups and a theory-neutral consortium preregistered predictions, experimental contrasts, and interpretation criteria before collecting the decisive data. Across 256 human participants, the program used functional MRI, magnetoencephalography, and intracranial electroencephalography during visual tasks with suprathreshold stimuli of different durations. Data were collected across independent laboratories and analyzed using separated development and test phases. The research design reduced several forms of post hoc theoretical accommodation and combined modalities with different temporal and spatial strengths. Its empirical claims concern a specified set of visual content and maintenance predictions, not every aspect of conscious experience or the entire philosophical meaning of consciousness. Cogitate Consortium et al. (2025)
The consortium reported information about experienced contents in visual, ventrotemporal, and inferior frontal cortex. Sustained responses in occipital and lateral temporal regions tracked aspects of stimulus duration, and content-specific synchronization occurred between frontal and early visual areas. At the same time, some previously articulated expectations of both theories were challenged. A reader should not translate the presence of content information in a region into the claim that this region alone produces the experience. Nor does the absence of an anticipated effect establish that a whole class of mathematical or computational models is untenable. The important empirical advance lies in describing exactly which registered prediction was supported, qualified, or disconfirmed by which measurement. Cogitate Consortium et al. (2025)
Adversarial collaboration clarifies the distinction between broad theoretical families and specific experimental instantiations. One theory may make a general claim about information availability while a particular implementation predicts a specific timing or cortical focus. Evidence against the narrow implementation may motivate a revision but should be documented as a revision rather than silently absorbed. Pre-registration strengthens accountability, yet the outcome still depends on the agreed task and the sensitivity of available measures. A mature research field needs repeated cross-laboratory comparisons that include experiences beyond a particular set of visual categories and protocols that separate conscious content from action demands. The Cogitate work provides both an evidence base and a reproducible model for designing these future comparisons. Cogitate Consortium et al. (2025) Seth & Bayne (2022)
Evidence Triangulation and Reproducibility
Triangulation means testing the same proposed explanatory relationship through genuinely different sources of evidence. Behavioral visibility, fMRI decoding, EEG dynamics, intracranial recordings, stimulation, lesions, and clinical command-following tasks each impose distinct assumptions and measurement errors. Their agreement is more compelling when errors and confounds are independent. For instance, a frontal marker present during reports can be evaluated with no-report designs, and a candidate network mechanism can be compared across wakefulness, sleep, and anesthesia. The objective is not merely to accumulate measures but to build inferential tests that could expose contradictory interpretations. Tsuchiya et al. (2015) Mashour (2024)
Reproducibility begins with a precise definition of the effect and its boundaries. A laboratory should specify the content under investigation, participant characteristics, awareness measures, stimulus timing, statistical comparisons, and predetermined exclusion criteria. The distinction between exploratory and confirmatory analyses is crucial when a study contains many regions, frequency bands, time windows, and possible contrasts. Multivariate methods can produce strong predictive accuracy while reflecting subtle differences in attention or motor preparation. A replication should therefore preserve critical experimental features and also vary them deliberately to test generalizability. Collaborative design and preregistration can reduce selective interpretation and establish an interpretable record of prediction failures. Cogitate Consortium et al. (2025) Fleming & Lau (2014)
Clinical evidence requires additional safeguards because patients may differ in sensory loss, sedation exposure, brain injury, and ability to follow instructions. A study reporting covert awareness may select participants with enough preserved cognitive capacity to complete its mental-imagery task; the resulting prevalence cannot automatically be applied to every patient who lacks motor responses. In a 2024 multicenter study, a convenience sample was explicitly described, and this sampling feature matters for external validity. In clinical and basic research alike, uncertainty should be quantified at the level of the actual sample and measured construct. This principle supports independent verification without reducing the scientific importance of unexpected dissociations between behavior and cognition. Bodien et al. (2024) Claassen et al. (2019)
Conceptual and Methodological Limits
Consciousness theories sometimes use the same word to designate different targets: phenomenality, cognitive accessibility, introspective confidence, arousal, or a particular computational organization. Apparent disagreement may be caused partly by these definitional differences; apparent agreement may conceal distinct commitments about causation. A disciplined comparison first translates a theory into observable predictions and identifies the level at which the prediction is intended to operate. For example, content decoding, cortical ignition, and perturbational complexity are related to different explananda, and success on one task does not validate the others automatically. Theoretical clarity is a condition for meaningful critical testing. Seth & Bayne (2022) Koch et al. (2016)
The principal empirical limits include measurement dependence, observer-related report effects, incomplete neural coverage, nonselective interventions, and failures of cross-task generalization. Functional imaging can localize statistical associations while leaving neural causation unresolved; scalp recordings can identify temporal signals while obscuring sources; intracranial recordings provide detailed but anatomically restricted samples. Investigations of no-report markers reduce some confounds while introducing others. Consequently, scientific synthesis should avoid ranking techniques by a single scale of authority. It should evaluate whether each method addresses the target construct with acceptable specificity, and whether alternative processes could generate the measured signal. Tsuchiya et al. (2015) Cogitate Consortium et al. (2025)
The philosophical explanatory problem remains distinct from the empirical assignment of correlates. Chalmers's hard-problem formulation asks why particular physical processes are accompanied by experience; modern laboratory studies identify structural, causal, and functional relationships that any successful explanation must accommodate. Philosophical positions disagree about whether those findings can close the explanatory gap, and no singular experiment has resolved that disagreement. The useful boundary is methodological: experiments support claims about measured participants, neural systems, and reported or inferred experiences, while broader metaphysical conclusions require independent argument. Keeping the levels explicit allows consciousness science to advance without treating explanatory pluralism as a failure of empirical research. Chalmers (1995) Seth & Bayne (2022)
Outstanding Problems and Research Directions
A central open problem concerns how to distinguish the neural substrates of experience from the prerequisites and consequences of experiencing. Progress requires experiments that control or model attention, decision, memory, motor output, and global arousal separately. This is particularly important when frontal and posterior neural signals are compared. The most productive designs make temporally and anatomically discriminating predictions, use independent data to set analysis decisions, and test the same theoretical claim in both report and validated no-report conditions. Research should also broaden its studied contents beyond restricted visual categories to include audition, bodily experience, emotion, and forms of conscious thought. Boly et al. (2017) Cogitate Consortium et al. (2025)
Another open problem is whether state-level measures can transfer reliably across sleep, anesthesia, injury, and ordinary wakefulness. Physiological signatures may change for several reasons, and their relation to subjective experience can depend on mechanism and diagnosis. Large longitudinal and multicenter cohorts, repeated behavioral evaluations, and mechanistically targeted interventions can test the stability of proposed markers. Clinical applications require evidence of diagnostic accuracy and benefit in relevant populations, including documented cases where physiological indicators and observable behavior diverge. Research ethics is inseparable from the interpretation of uncertainty because mistaken attribution of awareness can have serious consequences for patients and families. Bodien et al. (2024) Mashour (2024)
Long-term theory development also requires mathematically explicit bridges among psychological constructs, neural mechanisms, and causal models. Global broadcasting, recurrent interactions, higher-order monitoring, and integrated causal organization may be investigated as separable computational or physiological dimensions. Experiments should determine when each dimension is necessary, when it is insufficient, and how its contribution changes with task or organism. Comparative development and artificial systems present tests of whether theory commitments depend on human anatomy or a more general organization. The immediate scientific obligation is to make claims falsifiable and to preserve the distinction between empirically observed capacities and hypotheses concerning the existence or qualities of experience. Oizumi et al. (2014) Mashour et al. (2020)
Conclusion
Consciousness research investigates the conditions and mechanisms of subjective experience through coordinated psychological and neuroscientific methods. Its fundamental distinctions concern state, content, access, and phenomenological character. Across psychophysics, sleep, anesthesia, neural recording, and clinical brain injury, the evidence shows that external responsiveness, cognitive processing, and reported awareness can dissociate. These findings make the accurate definition of measured constructs a central scientific requirement. They also explain why no isolated brain signal or behavioral task supplies a universal measure of experience. The most stable conclusions are those supported by convergent methods and clearly bounded to the populations and phenomena actually studied. Koch et al. (2016) Seth & Bayne (2022)
Theoretical programs contribute different explanations. Global neuronal workspace theory emphasizes amplification and flexible broadcasting; integrated information theory models intrinsic causal organization; recurrent-processing theories emphasize feedback interactions; higher-order theories investigate representations of mental states; predictive approaches address inference and expectations. Their explanatory targets partly overlap and partly differ. The 2025 preregistered adversarial comparison provides a significant empirical constraint on selected claims from two major programs, while leaving fundamental questions open. Critical synthesis is therefore best expressed through explicitly delimited results, preserved alternative hypotheses, and proposals for experiments that could differentiate competing mechanisms. Cogitate Consortium et al. (2025) Mashour et al. (2020)
Future progress depends on integrating direct experience reports, computational modeling, behavioral inference, brain-wide observation, selective perturbation, and clinical validation. The study of consciousness continues to inform research on perception, cognition, self-monitoring, clinical recovery, comparative minds, and artificial systems. Its scientific integrity rests on distinguishing established empirical results from interpretations and wider philosophical commitments. Within Psychology by Bogdanova, this dossier serves as the foundational research object for consciousness itself, linked to broader Science areas and to specialized dossiers that investigate particular states, measurement techniques, or theories. The resulting synthesis places a complex phenomenon within a coherent, verifiable body of knowledge while preserving the questions that the current evidence has yet to resolve. Seth & Bayne (2022) Bodien et al. (2024)
References
- Bodien, Y. G., Allanson, J., Cardone, P., et al. (2024). Cognitive motor dissociation in disorders of consciousness. New England Journal of Medicine, 391(7), 598–608. https://doi.org/10.1056/NEJMoa2400645
- Boly, M., Massimini, M., Tsuchiya, N., Postle, B. R., Koch, C., & Tononi, G. (2017). Are the neural correlates of consciousness in the front or in the back of the cerebral cortex? Clinical and neuroimaging evidence. Journal of Neuroscience, 37(40), 9603–9613. https://doi.org/10.1523/JNEUROSCI.3218-16.2017
- Casali, A. G., Gosseries, O., Rosanova, M., et al. (2013). A theoretically based index of consciousness independent of sensory processing and behavior. Science Translational Medicine, 5(198), 198ra105. https://doi.org/10.1126/scitranslmed.3006294
- Chalmers, D. J. (1995). Facing up to the problem of consciousness. Journal of Consciousness Studies, 2(3), 200–219. https://consc.net/papers/facing.html
- Claassen, J., Doyle, K., Matory, A., et al. (2019). Detection of brain activation in unresponsive patients with acute brain injury. New England Journal of Medicine, 380(26), 2497–2505. https://doi.org/10.1056/NEJMoa1812757
- Cogitate Consortium, Ferrante, O., Gorska-Klimowska, U., et al. (2025). Adversarial testing of global neuronal workspace and integrated information theories of consciousness. Nature, 642, 133–142. https://doi.org/10.1038/s41586-025-08888-1
- Dehaene, S., Changeux, J.-P., Naccache, L., Sackur, J., & Sergent, C. (2006). Conscious, preconscious, and subliminal processing: A testable taxonomy. Trends in Cognitive Sciences, 10(5), 204–211. https://doi.org/10.1016/j.tics.2006.03.007
- Fleming, S. M., & Lau, H. C. (2014). How to measure metacognition. Frontiers in Human Neuroscience, 8, Article 443. https://doi.org/10.3389/fnhum.2014.00443
- Koch, C., Massimini, M., Boly, M., & Tononi, G. (2016). Neural correlates of consciousness: Progress and problems. Nature Reviews Neuroscience, 17, 307–321. https://doi.org/10.1038/nrn.2016.22
- Kouider, S., & Dehaene, S. (2007). Levels of processing during non-conscious perception: A critical review of visual masking. Philosophical Transactions of the Royal Society B: Biological Sciences, 362(1481), 857–875. https://doi.org/10.1098/rstb.2007.2093
- Lamme, V. A. F., & Roelfsema, P. R. (2000). The distinct modes of vision offered by feedforward and recurrent processing. Trends in Neurosciences, 23(11), 571–579. https://doi.org/10.1016/S0166-2236(00)01657-X
- Lamme, V. A. F. (2006). Towards a true neural stance on consciousness. Trends in Cognitive Sciences, 10(11), 494–501. https://doi.org/10.1016/j.tics.2006.09.001
- Lau, H., & Rosenthal, D. (2011). Empirical support for higher-order theories of conscious awareness. Trends in Cognitive Sciences, 15(8), 365–373. https://doi.org/10.1016/j.tics.2011.05.009
- Mashour, G. A., Roelfsema, P., Changeux, J.-P., & Dehaene, S. (2020). Conscious processing and the global neuronal workspace hypothesis. Neuron, 105(5), 776–798. https://doi.org/10.1016/j.neuron.2020.01.026
- Mashour, G. A. (2024). Anesthesia and the neurobiology of consciousness. Neuron, 112(10), 1553–1567. https://doi.org/10.1016/j.neuron.2024.03.002
- Massimini, M., Ferrarelli, F., Huber, R., Esser, S. K., Singh, H., & Tononi, G. (2005). Breakdown of cortical effective connectivity during sleep. Science, 309(5744), 2228–2232. https://doi.org/10.1126/science.1117256
- Monti, M. M., Vanhaudenhuyse, A., Coleman, M. R., et al. (2010). Willful modulation of brain activity in disorders of consciousness. New England Journal of Medicine, 362(7), 579–589. https://doi.org/10.1056/NEJMoa0905370
- Oizumi, M., Albantakis, L., & Tononi, G. (2014). From the phenomenology to the mechanisms of consciousness: Integrated information theory 3.0. PLOS Computational Biology, 10(5), e1003588. https://doi.org/10.1371/journal.pcbi.1003588
- Owen, A. M., Coleman, M. R., Boly, M., Davis, M. H., Laureys, S., & Pickard, J. D. (2006). Detecting awareness in the vegetative state. Science, 313(5792), 1402. https://doi.org/10.1126/science.1130197
- Schiff, N. D., Giacino, J. T., Kalmar, K., et al. (2007). Behavioural improvements with thalamic stimulation after severe traumatic brain injury. Nature, 448, 600–603. https://doi.org/10.1038/nature06041
- Sergent, C., Baillet, S., & Dehaene, S. (2005). Timing of the brain events underlying access to consciousness during the attentional blink. Nature Neuroscience, 8, 1391–1400. https://doi.org/10.1038/nn1549
- Seth, A. K., & Bayne, T. (2022). Theories of consciousness. Nature Reviews Neuroscience, 23, 439–452. https://doi.org/10.1038/s41583-022-00587-4
- Siclari, F., Baird, B., Perogamvros, L., et al. (2017). The neural correlates of dreaming. Nature Neuroscience, 20, 872–878. https://doi.org/10.1038/nn.4545
- Tononi, G. (2004). An information integration theory of consciousness. BMC Neuroscience, 5, Article 42. https://doi.org/10.1186/1471-2202-5-42
- Tsuchiya, N., Wilke, M., Frässle, S., & Lamme, V. A. F. (2015). No-report paradigms: Extracting the true neural correlates of consciousness. Trends in Cognitive Sciences, 19(12), 757–770. https://doi.org/10.1016/j.tics.2015.10.002
- Whyte, C. J., Redinbaugh, M. J., Shine, J. M., & Saalmann, Y. B. (2024). Thalamic contributions to the state and contents of consciousness. Neuron, 112(10), 1611–1625. https://doi.org/10.1016/j.neuron.2024.04.019