How can differences in levels or states of consciousness be explained?

IIT makes an explicit, intrinsic identity claim that the LEVEL/quantity of consciousness IS Phi, the integrated information of a complex computed from the system's own cause-effect structure. This is not merely an external correlation: Phi is a property the substrate has from its own intrinsic perspective, and IIT specifies WHY levels differ via an internal mechanism, namely the degree of irreducible integration-plus-differentiation among the units. Distinct states map onto distinct intrinsic causal regimes: NREM sleep and anesthesia drop Phi because cortical dynamics become either bistable/stereotyped (loss of differentiation) or functionally fragmented (loss of integration), while wakefulness and REM sustain high Phi. This addresses the asked-for HOW for state/level differences rather than only WHEN/WHERE, which is what distinguishes this question from others where IIT fails.

Key evidence: IIT's central identity (Tononi & Koch; IIT 3.0/4.0): the quantity of consciousness equals Phi, the maximally irreducible intrinsic cause-effect power of a complex, so reductions in level (NREM, anesthesia, coma) ARE reductions in intrinsic integration of the substrate, operationalized empirically by the integration+differentiation-sensitive PCI (Casali et al. 2013).

DIT's central commitment is that the STATE of consciousness is governed by an intrinsic control variable: the strength of non-specific (matrix/high-order) thalamic gating of apical-somatic coupling in layer-5 pyramidal neurons, mediated by metabotropic-receptor effects on apical calcium dynamics. This is a mechanistic internal-control account, not a passive external correlate — it specifies WHY states differ (coupling vs. decoupling) and how the same local perturbation yields different effects depending on coupling state. When apical and somatic compartments are coupled, reverberant corticocortical/thalamocortical dynamics sustain the wakeful state; thalamic decoupling under anesthesia or deep sleep suppresses that reverberation and collapses the level of consciousness. The theory is explicitly subtitled a theory of 'state and content' and uses this gating mechanism to explain wakefulness vs. anesthesia vs. sleep, which is precisely what the yes-bar demands.

Key evidence: Bachmann, Suzuki & Aru (2020) frame DIT as a thalamo-cortical theory of state and content in which non-specific thalamic gating of L5p apical-somatic coupling controls the global state — strong coupling = wakeful reverberant dynamics, decoupling = unconsciousness (anesthesia/deep sleep).

The yes-bar asks for an INTRINSIC internal mechanism that explains how differences in level/state of consciousness arise, with anesthesia vs wakefulness as the paradigm example. Orch OR supplies exactly such a mechanism rather than a mere external correlate: consciousness is identified with the occurrence of orchestrated OR events sustained by microtubule quantum coherence, and the LEVEL of consciousness is tied to the presence and scale of that coherence (intensity proportional to gravitational self-energy E_G of larger, more widely orchestrated, faster-collapsing superpositions). The flagship anesthesia account is genuinely internal and graded in principle: anesthetic gases bind tubulin and quench dipole oscillations/the ~613 THz resonance, dampening microtubule quantum effects and thereby abolishing OR events (consciousness) while leaving non-conscious neural activity intact. This is a stated causal control variable internal to the substrate (degree of coherence/OR), not a wakefulness-EEG correlate standing in for an explanation, so it clears the specific bar as written. I withhold high confidence because the account is admittedly thin on graded states beyond the anesthesia/wakefulness dichotomy (sleep, sedation, minimal consciousness) and rests on a coherence claim challenged by decoherence calculations.

Key evidence: Hameroff & Penrose (2014) and Hameroff's anesthesia work: anesthetics selectively erase consciousness by binding tubulin and quenching microtubule dipole/quantum resonance (disrupting OR), with conscious intensity scaling with E_G of the coherent state — an intrinsic substrate-level control variable for level of consciousness.

GNW makes an explicit architectural distinction between two dimensions: a global vigilance/arousal enabling state and content-specific conscious access. Crucially, it offers an internal control mechanism rather than mere correlation: the all-or-none nonlinear "ignition" that constitutes global access requires a sufficient level of global activation, which is gated by thalamocortical loops and ascending neuromodulatory systems. When that gating tone falls (non-REM sleep, anesthesia, coma), the ignition threshold cannot be reached, so workspace broadcasting collapses and the conscious level drops; preserved-but-disconnected workspace activity is invoked for dreaming/covert awareness. This is a genuine intrinsic dynamical account of how state differences arise (a threshold/ignition mechanism modulated by arousal), not just a map of where states show up, and it has been operationalized clinically to index residual consciousness.

Key evidence: Dehaene, Changeux & Naccache (2011) "The Global Neuronal Workspace Model of Conscious Access" applies the ignition-vs-vigilance distinction to disorders of consciousness, treating arousal-gated workspace ignition as the internal determinant of conscious level across wakefulness, sleep, anesthesia, and coma.

NPS as actually formulated by its proponents (Lyre 2022, Fink 2024) deliberately brackets the level/state distinction — it addresses only the structure of content within already-conscious states and is explicitly silent on when or how strongly a state is conscious. Any answer to the State question must be imported from Northoff's TTC, a distinct framework that the seed conflates with NPS. Even granting that import, TTC's account ("temporo-spatial nestedness" of spontaneous activity tracks wakefulness/sleep/anesthesia) is, by the theory's own listed limitations, correlational and at risk of being "a redescription rather than an explanation" — it identifies a graded internal property that co-varies with level but does not supply a genuine intrinsic control mechanism for how the system moves between levels. Under the strict, hard-problem-honest standard, correlating a state with degree of nestedness without an explanatory internal control mechanism falls short of the yes-bar, and the actual theory under judgment is silent.

Key evidence: "Lyre/Fink's NPS proper deliberately avoids the conscious/unconscious (level) distinction... This is a major silence. The gap is filled by Northoff's TTC" — i.e., NPS itself does not answer, and the substitute (TTC nestedness) is conceded to be correlational and possibly redescriptive.

The question asks how differences in LEVEL/STATE of consciousness (wakefulness vs non-REM sleep vs anesthesia) arise via intrinsic internal mechanisms. PCT's graded structures address something different: the control hierarchy gives a gradient between automatic and attended control, and Mansell's tripartite scheme distinguishes primary/secondary/tertiary consciousness as TYPES or aspects (phenomenal vs access vs symbolic self-reference), driven by rate of integrating novel information and where awareness is directed. These are not accounts of global arousal states. The theory's own stated limitation is explicit: "The base theory says little about states like sleep, anesthesia, or dreaming" and it is "comparatively underdeveloped on states like sleep/anesthesia." Even the partial graded account rests on Mansell's stipulated emergence of consciousness from reorganization, an identity claim rather than a mechanism, so there is no intrinsic control mechanism that would explain transitions in level of consciousness.

Key evidence: PCT's own stated limitation (drawn from Mansell 2022 and Powers 1973): "The base theory says little about states like sleep, anesthesia, or dreaming" and the theory is "comparatively underdeveloped on states like sleep/anesthesia."

HOT theory's core mechanism is metacognitive: a first-order state becomes conscious when targeted by a non-inferential higher-order thought. This machinery addresses STATE consciousness (whether a particular content is conscious) and graduates introspective access via yet-higher-order thoughts, but it offers no intrinsic mechanism that generates or regulates the GLOBAL level of consciousness (wakefulness vs non-REM sleep vs anesthesia). By the proponents' own admission, the account of global states is "less developed" and treats sleep/anesthesia merely as background conditions that enable or disable HOT formation — i.e., external preconditions, not an internal control mechanism that explains how the level itself is set. The Lau signal-detection work explains content access (blindsight, degraded metacognitive sensitivity from PFC lesions/TMS), which is a different axis from global arousal level. This fails the yes-bar's explicit requirement of an intrinsic internal control mechanism.

Key evidence: The theory's own characterization: "Global states (sleep, anesthesia) are less developed but treated as conditions enabling or disabling HOT formation" — HOTs are gated by these states rather than explaining how the states/levels themselves arise.

AST's core commitment is an account of what awareness IS (the brain's attention schema, the S+A+V model), not of why tonic arousal varies. By its proponents' own framing, the theory is "thin on graded arousal/state changes" and explicitly delegates global tonic arousal to "underlying brainstem/thalamocortical mechanisms it does not model in detail." Its only implied account of level — that awareness reports cease when the S+A+V model is not being constructed/updated (deep sleep, anesthesia) — is a correlation/gating observation, not an intrinsic control mechanism explaining HOW the transitions between wakefulness, non-REM sleep, and anesthesia are generated. In the proposed "standard model," level/access aspects are handed off to Global Workspace Theory, confirming AST itself supplies no internal state-control mechanism. This fails the yes-bar's demand for an intrinsic internal mechanism over mere external correlation.

Key evidence: Graziano & Webb (2015) and the standard-model paper (Graziano et al. 2020) frame AST as a theory of subjective awareness as a model of attention, with level/arousal left to brainstem/thalamocortical systems and access/level functions delegated to GWT — i.e., AST offers no intrinsic mechanism for graded states.

Question 4 demands an account of how differences in LEVEL/STATE of consciousness arise via INTRINSIC internal control mechanisms. Irruption Theory's treatment of states is, by the description's own admission, "addressed mainly through entropy correlations rather than a developed theory of levels" — it associates lower neural entropy with reduced awareness (dreamless sleep), baseline entropy with waking, and elevated entropy with psychedelic states. These are correlations between conscious level and a measured physiological quantity (entropy), not an internal mechanism that controls or generates the transitions between states. The absorption/irruption balance offers a taxonomy (flow vs. self-monitoring) but the theory explicitly declares the underlying mind-matter influence "unintelligible" / a black box quantifiable only indirectly, so it provides no mechanistic account of how the level is regulated. Under the strict standard, correlating states with an entropy measure without an intrinsic control mechanism is precisely the insufficiency the yes-bar excludes.

Key evidence: The theory's own characterization: state differences are "addressed mainly through entropy correlations rather than a developed theory of levels," and it "does not systematically map graded levels of consciousness onto graded motivational involvement beyond these entropy correlations" (Froese 2023/2024).

RPT identifies a discriminator for global states — effective recurrent processing and, in Lamme's 2018 amendment, recurrence-induced network plasticity (NMDA/calcium-dependent Hebbian change) — and uses presence/absence of these to mark conscious vs unconscious states like anesthesia. But the yes-bar requires an INTRINSIC internal CONTROL mechanism that explains how differences in LEVEL arise, not merely a marker whose presence/absence labels a state. RPT tells you what is missing in anesthesia (effective recurrence, plasticity) but does not specify the internal regulatory dynamics that generate or titrate transitions across wakefulness, sleep, and anesthesia; the gradation of levels is asserted via "depth and reach of recurrence" rather than mechanistically derived, and what enables vs suppresses recurrence across states is left to exogenous neuropharmacology. This is a where/what-distinguishes account standing in for the asked-for how-levels-arise, which the strict standard treats as insufficient.

Key evidence: Lamme (2018, Phil. Trans. R. Soc. B 373:20170344) concedes that recurrence persists under anesthesia without consciousness, forcing the post-hoc plasticity amendment — showing RPT supplies a marker to flag the unconscious state rather than an intrinsic mechanism that controls the level transition itself.

Unlike most other questions where PP only correlates or gestures, on conscious LEVEL/STATE the theory does commit to an explicit INTRINSIC control mechanism: precision-weighting, which is the brain's internal estimate of the inverse variance of priors versus sensory evidence. Seth's computational simulations model altered states (psychedelic, Parkinsonian, dementia-related) as shifts in this internal balance of prior vs sensory precision, and Friston ties richer/deeper conscious processing to the depth of "temporally thick" generative models — both are internal architectural/dynamical parameters, not mere external correlates. This satisfies the yes-bar's demand for an internal control mechanism generating state differences rather than just locating where states occur. The account is genuinely about a tunable internal gain/precision parameter that grades the contribution of top-down models, which is a mechanistic story for how level changes (e.g., loss of sensory precision under anesthesia, runaway prior precision in hallucination).

Key evidence: Seth's computational simulations of predictive perception model altered states as shifts in the balance of prior precision against sensory precision (Seth 2021, Being You; Seth & Bayne 2022), an intrinsic precision-weighting mechanism rather than an external brain-activity correlate.