The entropic brain - revisited

The paper is a narrative review and theoretical update rather than a formal systematic review or meta-analysis. The extracted text reports that the author conducted a literature survey (a PubMed search at submission in January 2018) and synthesised empirical neuroimaging work published since the original 2014 paper; no formal eligibility criteria, data extraction protocol, or risk-of-bias assessment are described in the extracted text. Carhart-Harris integrates findings from EEG, MEG and fMRI studies of psychedelics, reports applications of specific signal-complexity measures (notably Lempel-Ziv complexity), discusses new analytic approaches such as connectome harmonics and perturbational-complexity indices (PCI), and combines these empirical results with theoretical discussion about criticality and serotonin function. The paper also outlines proposed experimental paradigms (for example, experience sampling with music during scanning, use of microphenomenology) to improve mind–brain mappings, but these are methodological proposals rather than implemented methods in this paper.

Across multiple recent human neuroimaging studies synthesised in the review, increased measures of brain entropy under psychedelic compounds were consistently reported. Lempel-Ziv complexity (a compressibility-based measure of signal complexity) applied to MEG and EEG data showed higher complexity under LSD, psilocybin and ketamine, and these increases tended to correlate with subjective intensity of the psychedelic experience. fMRI studies also reported elevated spatial/temporal complexity under LSD and ayahuasca, with one report that acute increases in fMRI-measured entropy predicted changes in personality (increased openness) measured two weeks later. Separate work applying Lempel-Ziv measures to TMS-evoked EEG responses produced the perturbational-complexity index (PCI), which reliably differentiates disorders of consciousness (vegetative state, minimally conscious state, etc.), and spontaneous Lempel-Ziv measures have distinguished awake from anaesthetised states with high sensitivity. Beyond entropy per se, novel analyses such as connectome-harmonic decomposition revealed an expanded repertoire of brain states under LSD and evidence consistent with enhanced criticality, for example power-law distributions of connectome-harmonic activity. The review notes similar entropic increases with the NMDA antagonist ketamine and with DMT-containing ayahuasca, and highlights that not all external stimuli affect the psychedelic state equally—immersive, emotionally resonant inputs (for example, music matched to a patient) appear to have a stronger impact on subjective outcomes. The extracted text does not present pooled numerical effect sizes or confidence intervals; results are reported qualitatively and by study examples.

Carhart-Harris interprets the aggregated findings as supporting the central claim that increased brain entropy indexes an expansion in the richness, complexity or information-content of conscious experience—up to an upper limit beyond which coherent phenomenal experience may fail. He frames this within a 'zone of criticality', proposing that normal waking consciousness sits toward the upper end of a sub-critical band, psychedelics shift dynamics closer to criticality or even toward super-criticality, and sedatives/anaesthetics shift dynamics downward. Heightened criticality is argued to confer maximal sensitivity to perturbation, which may explain why ‘set’ and ‘setting’ strongly shape psychedelic experiences and therapeutic outcomes. The review offers concrete proposals for advancing the theory empirically: combining dynamically sensitive brain measures (sliding windows, point-process, intrinsic ignition) with refined subjective sampling (experience sampling, microphenomenology); using structured stimuli such as music to time-lock likely epochs of insight; measuring the entropy of subjective reports (for example, linguistic richness); and applying machine-learning approaches while recognising that entropic states may be harder to predict. Carhart-Harris also links recent serotonin research to the hypothesis, noting evidence that dorsal raphe serotonin neurons signal 'surprise' or unsigned prediction error, which is formally related to uncertainty or entropy; 5-HT2A receptor signalling is suggested to relax prior beliefs and enhance plasticity, a mechanism consistent with increased entropy facilitating novel learning. Limitations acknowledged in the paper include its selective focus, the absence of systematic comparisons with stimulants (so that alternate explanations like general arousal cannot yet be fully ruled out), the need for more spatially localised analyses, and unresolved details about how entropy relates to criticality. The author concedes the speculative nature of parts of the update and reiterates that the entropic-brain framework is intended as a heuristic to guide testable hypotheses rather than a definitive account. The review also revises one earlier claim—the suggestion that seizures are sub-critical relative to waking activity—because evidence points to super-critical dynamics in seizures.

Four years after the original entropic brain proposal, Carhart-Harris maintains that the hypothesis has accrued empirical support and remains a useful, if speculative, heuristic. The paper concludes that entropy is a powerful bridge between measurable brain dynamics and subjective experience and that pursuing entropy- and criticality-focused research may advance understanding of consciousness and inform therapeutic approaches for psychiatric and neurological conditions.