Order, Chaos, and Consciousness: Examining the Entropic Brain Hypothesis
How does the coordinated activity of billions of neurons give rise to the richness of subjective experience?
Among the many frameworks proposed to tackle this mystery, the Entropic Brain Hypothesis (EBH) has emerged as one of the most ambitious and controversial. First articulated by Robin Carhart-Harris and colleagues in 2014, EBH argues that the quality of consciousness can be indexed by the entropy – the randomness or diversity – of brain activity patterns.
In this model, ordinary waking life is a relatively ordered, low-entropy state, while dreams, psychosis, and especially psychedelic states represent higher-entropy forms of awareness, marked by looser cognitive constraints and more diverse neural activity.
EBH gained traction through a striking paradox in early neuroimaging work with psilocybin. Instead of simply “activating” the brain, psychedelics suppressed activity in certain regions – most notably the Default Mode Network (DMN) – while simultaneously increasing the diversity of connectivity motifs across the cortex. This suggested that psychedelic consciousness was not a matter of heightened activation, but of more unpredictable and flexible dynamics.
Carhart-Harris drew a bold analogy to Freudian metapsychology: primary process thinking (dreamlike, free-associative, ego-dissolving) maps onto high-entropy brain states, whereas secondary process thinking (rational, reality-oriented) corresponds to low-entropy order.
EBH also draws on complex systems theory, a field that studies how networks of interacting components give rise to emergent properties that cannot be predicted from the individual parts alone. This framework emphasises concepts such as emergence and self-organisation, where simple interactions between elements can generate complex, system-level behaviours through nonlinear dynamics. Carhart-Harris proposed that consciousness operates optimally when the brain hovers near criticality – a boundary between rigid order and chaotic disorder. In such states, complex systems display their richest diversity of patterns, enabling adaptability without total collapse. Normal waking consciousness, EBH suggests, lies just below this critical point, preserving coherence and stability. Psychedelics, by disrupting high-level priors and relaxing hierarchical constraints, push the brain closer to (or beyond) this threshold, producing both heightened flexibility and vulnerability to disorganisation.
This claim is provocative: it recasts psychedelic consciousness as not merely “drug-induced distortion” but as the brain revealing its latent capacity for greater dynamical freedom. In this framing, altered states may be revealing the underlying principles of conscious organisation.
The Default Mode Network: The Seat of the Self?
Central to EBH is the Default Mode Network (DMN), a set of interconnected regions (including the medial prefrontal cortex and posterior cingulate cortex) that is reportedly most active when the mind is engaged in introspection, autobiographical memory, and self-referential thought. Carhart-Harris suggested that the DMN is the neural correlate of the Freudian ego, enforcing top-down order by constraining other brain systems. Under psychedelics, this network becomes less coherent, its internal connectivity dissolving while cross-talk with other networks increases. The result is the subjective phenomenon of ego dissolution, in which the boundary between self and world breaks down.
This mechanistic link between DMN disruption and altered selfhood has become one of the most widely cited aspects of EBH, even though critics note that large-scale imaging studies have failed to consistently tie DMN activity to measures of ego strength or consciousness. Still, the appeal of mapping a psychoanalytic construct onto a neural network has made EBH influential across neuroscience, psychiatry, and philosophy of mind.
The Freudian Connection
While RCH later distanced EBH from its connection to more explicit Freudian concepts, it’s worth examining the framework’s psychoanalytic origins. Freud’s 1923 structural model of the psyche – id, ego, and superego – cast the ego as a fragile mediator between instinctual drives and external reality. For Freud, mystical experiences of “oceanic boundlessness” represented a regression to an infantile state of undifferentiated ego, a longing for the lost narcissism of infancy rather than a genuine higher consciousness. Later psychoanalysts, however, moved in divergent directions. Ego psychologists such as Anna Freud and Heinz Hartmann championed the strengthening of ego capacities as central to mental health, while Jung and existential-humanist thinkers reframed ego dissolution as a potential path to individuation or authentic being. By the 1960s, the Freudian influence had gone elsewhere: figures like Timothy Leary and Stanislav Grof recast ego death as a liberatory, even mystical process. Leary’s Psychedelic Experience mapped LSD trips onto the Tibetan Book of the Dead, while Grof interpreted ego disintegration as a gateway to transpersonal insight. Psychedelic culture thus embraced ego dissolution as positive, aligning it with spiritual awakening and countercultural rebellion against conformity.
Together, these conceptual pillars – Freud’s metapsychology, complexity and criticality, the DMN as ego, and links to contemporary theories – constitute the foundation of EBH. Yet, as the next section will show, the hypothesis stands or falls on whether empirical data substantiate its bold claims.
What Does the Evidence Say?
The earliest functional MRI studies with psilocybin revealed results that completely defied expectations. Instead of heightened global brain activity – which researchers initially expected – participants actually showed reduced blood flow and oscillatory power in key brain regions, especially the Default Mode Network. At the same time, analyses showed that connectivity patterns across the cortex became far more diverse and dynamic. Rather than locking into stable, predictable patterns, brain networks continually formed and dissolved in new combinations. This greater "repertoire" of brain states was interpreted as evidence that psilocybin pushes the brain into a more entropic (unpredictable and flexible) regime, reflecting the free-associative quality of psychedelic thought.
Similar patterns emerged in MEG studies, which measure magnetic fields produced by brain activity. Psilocybin and LSD were both shown to suppress dominant rhythms in the alpha frequency band – a neural signature typically associated with top-down control and mental stability. The loss of this oscillatory order coincided with increases in broadband signal diversity, again pointing toward a shift to a higher-entropy condition. Importantly, these physiological changes matched participants' reports of vivid imagery, ego dissolution, and a breakdown in the ordinary sense of self.
Beyond conventional fMRI and MEG analyses, researchers have applied information-theoretic tools such as Lempel–Ziv complexity to capture the diversity of neural signals – essentially mathematical ways to measure how unpredictable and varied brain activity becomes. Across multiple studies, psilocybin, LSD, and even ketamine produced reproducible increases in these entropy-like measures. What makes these findings compelling is not only their consistency across different drugs with distinct pharmacologies but also the correlation with subjective ratings of intensity. The stronger the signal diversity, the more powerful the reported psychedelic experience.
In one landmark LSD study, participants who showed the largest entropy increases also exhibited lasting psychological changes weeks later, particularly in the personality trait openness – essentially becoming more curious, creative, and open to new experiences. This suggested that transient excursions into entropic states might leave enduring imprints on cognition and personality. The possibility that psychedelics could "unlock" long-term flexibility by temporarily destabilising entrenched patterns has become a cornerstone of the therapeutic narrative surrounding these drugs.
Functional connectivity analyses have further reinforced EBH by tying DMN disintegration to the subjective phenomenon of ego dissolution. In a 2016 LSD trial, the degree of decoupling within the DMN directly predicted the intensity of participants' reported loss of self. As DMN coherence diminished, participants described boundaries between self and world collapsing, often accompanied by feelings of unity, boundlessness, or transcendence. This mechanistic link between a neural network and a core feature of psychedelic phenomenology has become one of the most cited pieces of evidence in support of EBH.
Other Theories of Psychedelic Consciousness
Carhart-Harris himself has evolved beyond simple entropy-based explanations, however. He co-developed the REBUS (Relaxed Beliefs Under Psychedelics) model with Karl Friston. REBUS attempts to provide mechanistic specificity to entropy increases by proposing that psychedelics disrupt hierarchical predictive processing – the brain's system for making predictions about sensory input based on prior experience. According to REBUS, psychedelics reduce the “precision” or confidence assigned to high-level beliefs, allowing bottom-up sensory information to drive perception more directly.
The REBUS model has faced strong criticism, especially from researchers like Manoj Doss. He argues that the theory is too vague and hard to test. According to Doss, REBUS explains almost everything people report on psychedelics – whether it’s surprise, fear, creativity, or deep insight – by calling it a “prediction error.” But it doesn’t spell out exactly how or why these things happen in a way that could be tested. On top of that, some evidence doesn’t fit the model’s expectations. REBUS suggests that higher-level thinking skills (like reasoning or self-reflection) should break down more than basic sensory processes. Yet in reality, psychedelics often disrupt simple reflexes, such as how the brain filters sudden sounds, while sometimes leaving complex functions intact – or even improving them, such as making words feel more familiar or meaningful.
One important theory is Integrated Information Theory (IIT), which tries to measure consciousness using a mathematical score called "phi" (Φ). While EBH focuses on entropy – meaning how varied and unpredictable brain activity becomes – IIT emphasises unity and how well different brain regions work together. Some researchers think these two approaches might work together rather than compete: entropy captures the brain's dynamic richness and flexibility, while phi measures how coherently everything fits together.
Another key theory is Global Neuronal Workspace Theory (GNWT), which sees consciousness as information being “broadcast” widely across brain networks – like news being shared across different departments in a company. This idea aligns well with the increased global connectivity researchers observe when people take psychedelics.
Finally, predictive processing frameworks dovetail perfectly with EBH's focus on “loosened priors” – essentially, psychedelics may make the brain less rigid about its expectations and assumptions about reality. This idea was later refined into the REBUS model (Relaxed Beliefs Under Psychedelics), which explains how psychedelics might temporarily "relax" our deeply held beliefs and assumptions, creating opportunities for new perspectives and therapeutic breakthroughs.
Why Most Neuroscience Fails to Replicate
A recurring issue for all psychedelic neuroscience is the reliance on very small samples. Early fMRI studies with psilocybin or LSD often involved fewer than 20 participants. While understandable given the regulatory and logistical challenges of psychedelic research, such sample sizes are insufficient to reliably detect brain-behaviour relationships. Brain-wide association studies suggest that hundreds, if not thousands, of participants are necessary to achieve robust and generalisable findings. So-called ‘translational psychiatry’, or the discipline of converting the findings of neurobiology to clinical practice, has largely failed to deliver any clinically meaningful biomarkers. Even large-scale studies of depression yield only minute and inconsistent changes in cortical thickness or connectivity. And since most cognitive processes involve multiple brain regions, and most regions participate in multiple processes, inference to particular brain regions is rarely justified and may constitute a ‘neuro-phrenology’. Another thorny methodological issue is the difficulty of maintaining blinding in psychedelic trials. The subjective effects of substances like LSD and psilocybin are unmistakable, making it nearly impossible for participants – or researchers – to remain unaware of who has received the active drug. This introduces powerful expectancy effects: participants who know they have ingested a psychedelic may report ego dissolution, therapeutic insights, or heightened openness simply because they expect such outcomes. Carhart-Harris has offered some interesting solutions. He advocates the use of active placebos (e.g. DXM) to improve experimental control. He also suggests replacing boring, repetitive ‘button push’ tasks that frustrate the tripper subjects with ‘experience sampling’ through highly calibrated questions about their phenomenologies, as well as continuous brain scanning during affectively immersive activities – like listening to music – to ‘draw out’ and deepen the richness of the experience.
That said, EBH encounters another obstacle: “entropy” is not a single, unified measure, but rather a collection of different metrics – Shannon entropy, Lempel–Ziv complexity, sample entropy, fractal dimensionality, and more – each emphasising distinct aspects of signal diversity. In practice, these metrics often diverge in outcomes. In one recent study assessing 12 different brain-entropy measures following psilocybin, only a subset showed significant effects, and many others did not. Moreover, the inter-metric correlations were weak: different metrics often did not line up with one another as one might expect. That suggests they may not be tapping into a single underlying property.
Another issue arises from the use of the Default Mode Network (DMN) as a neural linchpin in EBH: some critics argue there is no persuasive evidence that the DMN exists as a stable, coherent functional unit. The DMN is typically inferred from resting-state fMRI data as a set of brain regions whose activity fluctuates together when a person is not engaged in a task. But critics point out that these observed correlations may arise from methodological artefacts (such as vascular coupling or scanner noise) or from arbitrary choices in data processing, rather than reflecting a genuine, functionally unified network.
Empirically, the constitution of the DMN is unstable across individuals, sessions, and analysis methods, and its boundaries shift depending on how the data are filtered, aligned, or decomposed. Moreover, the brain is never truly “at rest” – even in so-called resting-state scans, people’s minds wander, think, remember, plan, and adjust breathing. This means that what is labelled as “default mode” may simply be the accumulation of uncontrolled cognitive content, not a specially privileged network.
Critics such as John Horgan have elsewhere emphasised the vagueness of complexity science altogether. The very terms the field relies on – entropy, emergence, criticality, and, of course, complexity – have always carried dozens of overlapping and often incompatible definitions that, when applied to all disciplines, soon lose application to any one in particular. Showing that the brain actually reaches this delicate balance of criticality is challenging. The markers of true criticality – such as patterns of neural activity that repeat across different scales – are hard to capture with tools like fMRI, which have limited precision in time.
In his revisitation of EBH, Carhart-Harris describes a more complex picture of criticality. Psychedelics may ‘normalise’ pathologically subcritical states of consciousness, nudging them back toward criticality. This is proposed as one mechanism behind their therapeutic benefits. He also stresses context dependence: music, for instance, can enhance criticality and shape the direction of psychedelic effects. But psychedelics may instead overshoot, tipping it into a supercritical regime: a state where activity becomes excessively excitable and unstable, similar to hypomania and psychosis, during which criticality may increase during onset but decrease again as rigid delusions consolidate. Moreover, until psychedelics are directly compared to stimulants, we cannot know whether “entropy increases” are not simply the by-products of arousal or alertness (caffeine, for instance, boosts entropy).
This complexified interpretation would make space for reports of impaired working memory, disorganised thought, and reduced reality testing that often accompany high-dose psychedelic experiences. Here, two researchers instead propose a multidimensional model, in which different domains – perception, cognition, selfhood – can be modulated independently. Psychedelics, for example, may intensify visual perception while simultaneously impairing working memory, and labelling such a state as “higher” consciousness obscures these trade-offs.
As for studies showing that caffeine can elevate entropy, the researcher Gaige Clark proposes that high entropy may simply reflect the brain’s chaotic search for solutions when faced with novel or unresolved problems, in contrast to low-entropy spaces, in which efficient, reflexive patterns are formed through learning. From this perspective, entropy may index arousal or motivation rather than the distinctive phenomenology of psychedelics. For example, caffeine might increase entropy by heightening motivation, prompting problem-oriented thinking and activity, while ADHD’s reduced entropy could reflect hypo-arousal and reduced problem engagement.
Why Putting Things in Terms of Brains Feels More True
The issues that face EBH are more fundamental still. Functional neuroimaging, particularly fMRI, relies on correlational data. As Vidal and Ortega note in Being Brains: Making the Cerebral Subject, this limitation leads to recurrent “slippage” from correlation to causation in both academic and public discourse. RCH presents ‘neurophenomenology’ – or, extremely close associations of neural data with subjective events – as a bridge between neural dynamics and subjective experience, but the hard problem of consciousness remains untouched. We may add that the reliability of self-report is dubious under psychedelics, given their tendency to induce false perceptions, memories, and beliefs.
The allure of neuroimages as “windows into the mind” has contributed to both the commercialisation of brain imaging and the broader “neurologisation” of human research. Sometimes, this has shaped policy – such as using neuroimaging evidence to promote the U.S. Mental Health Parity Act of 2008 – and has birthed a range of ‘neuro’ disciplines, like ‘neuro-law’, ‘neuro-ethics’, and ‘neuro-politics’ in an inflationary process known as ‘neuromania’. As documented in Neuropsychedelia by Nicolas Langlitz, numerous scientists in the early phases of the psychedelic renaissance “used the neuroscience hype of the 1990s strategically to re-legitimise human research with hallucinogenic drugs”.
Illes et al. (2008) found that both health providers and patients with major depressive disorder expressed strong interest in brain scans for tailoring treatment, improving understanding of illness, and reducing stigma. Yet, the authors acknowledged that neuroimaging research has no current clinical application in depression and remains “a long way” from individual-level use. They nevertheless justified continued research on the basis of “rapid innovation” and the anticipated progression “from discovery to implementation.”
The critic Chaim Wigder has charged that psychedelic research more often mirrors longstanding and unchallenged notions from folk understanding rather than undertaking real science. Psychedelics have long been described as breaking down ego structures, reconnecting users to nature, and inducing sacred or mystical unity. Carhart-Harris suggested that psychedelics may re-align human consciousness with the ‘criticality’ found across natural systems (ecosystems, avalanches, forest fires), which he interprets as the basis of their reportedly sacred qualities.
EBH risks simply translating those claims into complexity-science jargon – “entropy,” “criticality,” “hierarchical relaxation” – which conveniently align with 1960s countercultural ideals of anti-hierarchy, authenticity, and harmony with nature. Indeed, if psychedelics increase entropy and unpredictability, why do they reliably produce predictable outcomes (ego dissolution, nature-relatedness, religiosity) shaped by set and setting? Cults have made considerable use of psychedelic drugs to assist in programming their members. The researcher Gul Dolen explicitly called for such techniques of suggestion to be operationalised in research, through introducing climate action concepts during a psychedelic trip to create long-lasting activism. Such outcomes illustrate how predictable and Pavlovian the apparently “entropic” state can be.
It is not clear that EBH succeeds as a definitive theory of consciousness. It has catalysed important research into brain network dynamics, forced us to reconsider the relationship between neural activity and subjective experience, and highlighted the poverty of our current models of consciousness. Perhaps it has done as much as any hypothesis can in being a productive framework for generating questions.