Empirical Refutation and Correction of Thermodynamic Markers in Sleep and Psychedelics

Empirical Refutation of a Universal Thermodynamic Marker

The recent multi-cohort study published on September 13, 2026, rigorously refutes the universality of the R = E/|O_info| ratio—a thermodynamic marker proposed as a candidate biomarker for consciousness level—across sleep, anesthesia, and psychedelic states. By analyzing eight pre-registered datasets (including Sleep-EDF, DOSE-I propofol, and DMT+harmine trials), the researchers found that the original R metric fails as a scalar predictor and does not serve as a universal frontier between conscious states. Specifically, the ratio's denominator (O_info, a measure of integrated information) was often near-null, leading to mathematical singularities and unreliable predictions. The study’s outcome correlation for R as a consciousness marker was weak (r = -0.36), and separability between wake and non-REM (NREM) sleep was dominated by spectral energy alone, not by the ratio or its informational component.

Mechanistic Insights: Corrected Metrics and Distinct Neural Dynamics

By correcting the informational denominator with a non-parametric estimator sensitive to non-Gaussian synergy, the study revealed robust gradients within sleep but not across pharmacological or psychedelic states. In sleep, the corrected metric (R_np) monotonically tracked consciousness level, with a descriptive cost minimum at lighter sleep stages (N1/N2 and REM) and a maximum at deep sleep (N4). Notably, the sleep field exhibited consistent negative O_info (synergy) in nearly all epochs, indicating a locked informational structure unique to natural sleep. In contrast, psychedelic states (e.g., DMT) showed a shift from marginal redundancy toward synergy without locking, and the invariants that characterized sleep failed to generalize. This suggests that the neural informational architecture of sleep and psychedelic states are fundamentally distinct, challenging the notion of a universal biomarker for consciousness level based on thermodynamic principles.

Research and Methodological Implications

The study’s methodological rigor—pre-registration, frozen criteria, and mandatory marginal decomposition—sets a new standard for biomarker research in neuroscience. The authors introduce a transferable framework for “killing” an observable honestly, resurrecting it with improved estimators, and mapping the boundaries of its validity. This approach is particularly valuable in the field of psychedelic research, where robust, quantitative biomarkers for consciousness level are in high demand but often lack empirical validation across diverse altered states. The finding that sleep invariants do not generalize to psychedelics or anesthesia underscores the need for state-specific markers and cautions against overgeneralization from sleep research to pharmacological interventions.

• Concrete example: The study demonstrates that the corrected synergy metric (R_np) can distinguish between sleep stages with high accuracy (AUC 0.904), but fails to do so in propofol anesthesia or DMT-induced psychedelic states, where energy alone carries the predictive signal.
• Non-obvious implication: The methodological arc—especially the use of blindness-vs-absence discrimination and floor-declaring estimators—provides a blueprint for future work aiming to validate or falsify candidate biomarkers in complex brain states.

Risks, Unknowns, and Cautions

The study does not claim clinical or diagnostic utility for the corrected thermodynamic marker, nor does it address legal or policy ramifications directly. The risk of misapplying sleep-derived biomarkers to psychedelic or anesthetic states is highlighted: such generalizations may obscure fundamentally different neural mechanisms and lead to erroneous interpretations of consciousness level. Additionally, while the corrected metric offers improved descriptive power within sleep, its utility in real-time or clinical monitoring remains unproven. The study also acknowledges the limitations of current informational estimators and the need for further research to refine these tools for broader application.

Looking Forward: Toward State-Specific Biomarkers

The empirical refutation of a universal thermodynamic marker for consciousness marks a pivotal advance in the neuroscience of altered states. For researchers and clinicians, the findings emphasize the importance of state-specific approaches when developing and validating biomarkers for consciousness in sleep, anesthesia, and psychedelics. The methodological innovations introduced—such as mandatory marginal decomposition and pre-registered refutation branches—are likely to influence future studies seeking robust, generalizable neural metrics. As the field moves forward, the lesson is clear: rigorous empirical testing and methodological transparency are essential to distinguish genuine invariants from context-dependent observables in the complex landscape of brain states.

Reviewed by Dr. Alexei Markov, PhD (Neuroscience, University of Zurich). How we research: All findings are sourced directly from the original study (OpenAlex W7212458611), with methodological details and cohort information verified against the pre-registered protocols. Reviewed on 2026-09-15.