ResearchPod Summary
The REBUS (Relaxed Beliefs Under pSychedelics) model integrates the free-energy principle—which posits that the brain functions to minimize uncertainty by updating internal models—with the entropic brain hypothesis, which suggests psychedelics increase the entropy of brain activity. The authors argue that psychedelics, through their action on 5-HT2AR receptors, specifically target the brain's high-level hierarchical structures. By reducing the 'precision weighting' (or felt confidence) of these high-level priors, the drug effectively 'flattens' the brain's energy landscape. This process releases the brain from the rigid, top-down constraints that normally govern perception and cognition, leading to a state of increased entropy and flexibility.
The 'anarchic brain' refers to the consequence of relaxing these high-level priors. In a normal state, the brain's highest levels (such as the Default Mode Network) exert a compressive, top-down influence that suppresses lower-level information. Under the influence of psychedelics, this hierarchy is disrupted. As top-down control weakens, bottom-up information—particularly from intrinsic sources like the limbic system—is liberated. This allows for a more chaotic, or 'anarchic,' state of information processing, which the authors argue is the neurobiological basis for the vivid, diverse, and often unitive experiences reported during a psychedelic trip.
The authors propose that many psychiatric disorders, such as depression and obsessive-compulsive disorder, are characterized by pathologically rigid, overweighted high-level priors that lock the individual into negative or restrictive patterns of thought. By temporarily relaxing these priors, psychedelic therapy creates a 'hot state' of heightened plasticity. This window of opportunity allows for the revision of these entrenched beliefs. When combined with proper psychological support, this process can lead to long-term therapeutic benefits, as the individual is able to recalibrate their internal models to be more flexible and better aligned with reality.
Alex: Welcome to another episode of ResearchPod. Today we're looking at the REBUS model — a theoretical framework that integrates the free-energy principle and the entropic brain hypothesis to explain how psychedelics alter consciousness.
Sam: The framing that caught my attention: why do psychedelics produce such profound shifts in high-level cognition and personality while leaving basic motor function largely intact?
Alex: That asymmetry is exactly what the model is built to explain. The central claim is that psychedelics don't simply induce neural chaos — they specifically target the precision weighting of high-level priors, the brain's most entrenched top-down expectations, and flatten the hierarchical free-energy landscape at that level, while leaving lower-level sensorimotor processing relatively undisturbed.
Sam: So the selectivity is the thing to explain. The brain runs on predictive processing — higher levels constantly suppress what bubbles up from below. If that suppression becomes too rigid, the whole system locks in.
Alex: Right. Think of it as a bureaucracy. The CEO — your high-level priors — routinely suppresses reports from the factory floor. Under normal conditions, that's efficient. But if the CEO has entrenched a pathological model of the world, nothing from the ground level can challenge it. Psychedelics, on this account, temporarily fire the CEO. Bottom-up sensory and affective data reaches the board directly, without being filtered through those overweighted expectations.
Sam: Which is why the therapeutic framing makes sense for something like treatment-resistant depression. The prior has locked in a rigid, negative self-model, and no amount of talk therapy can dislodge it because it's too precisely weighted to update.
Alex: Exactly. By relaxing the precision of those overweighted high-level priors, the drug allows suppressed limbic inputs — emotional signals that were being systematically discounted — to reach consciousness. The authors frame this as a potential reset: the belief system becomes temporarily plastic, and a new, less pathological attractor state becomes accessible.
Sam: So what's the proposed mechanism at the neural level? How does 5-HT2A receptor agonism actually produce this flattening?
Alex: The key is where those receptors are densely expressed: deep-layer pyramidal neurons in association cortex — precisely the cells thought to carry high-level predictions downward through the hierarchy. Agonism at those receptors induces what the authors call spike-field decoherence. Neuronal firing becomes decoupled from local field potentials, which effectively reduces the signal-to-noise ratio of those descending predictions. The high-level prior loses its grip — its confidence, in the computational sense — and the sharp energy minima that normally constrain the system get flattened out.
AI-generated third-party summary by ResearchPod. Not official content or an endorsement by the paper authors or affiliated organizations.
Sam: So the system moves away from a deep, narrow attractor toward something much broader and shallower. It becomes hypersensitive to incoming data rather than filtering everything through a fixed model.
Alex: The authors use the analogy of simulated annealing. You heat the system — raise its entropy — so it can escape local minima and explore the broader energy landscape. Then as the drug clears, the system cools back down and, ideally, settles into a more adaptive configuration. That's the mechanistic story behind the afterglow effect clinicians observe: the days or weeks of increased psychological flexibility after a session aren't just residual pharmacology — they reflect a genuine structural recalibration of the generative model.
Sam: That's a tidy story. Where does it strain?
Alex: The authors are candid about the central limitation: the "high-level" attribution is still largely abstract. The framework predicts that psychedelics selectively flatten precision at the top of the hierarchy, but mapping that claim onto specific neuronal populations with sufficient resolution to test it rigorously is still out of reach. Ego dissolution is a well-characterized phenomenological state, but pinning it to a precise computational signature in identified circuits — rather than gesturing at association cortex broadly — remains an open problem. The model is generative and internally consistent, but it's running ahead of what the empirical methods can currently resolve.
Sam: So it's a framework that organizes existing findings and generates predictions, rather than a model that's been directly confirmed at the mechanistic level.
Alex: That's the right read. What it does well is unify two previously separate accounts — the free-energy principle and the entropic brain hypothesis — into a single computational story that makes contact with both the pharmacology and the phenomenology. It gives researchers a shared vocabulary and a set of testable predictions. If precision weighting of high-level priors is the operative variable, you'd expect the therapeutic signal to correlate with measures of prior relaxation, not just with the intensity of the acute experience. Some early data points in that direction, but the causal chain hasn't been closed.
Sam: It also reframes what psychedelics are doing at a conceptual level — not disrupting the brain indiscriminately, but selectively modulating the confidence the system places in its most entrenched beliefs.
Alex: Which is what makes the model worth taking seriously as a research scaffold, even before the mechanistic details are nailed down. The question it poses — how do you therapeutically update a prior that's too precisely weighted to revise through ordinary experience? — is a genuinely important one, and REBUS offers a principled computational answer. Whether that answer survives contact with the next generation of circuit-level data is the experiment worth watching.
Sam: Thanks for walking through that, Alex.
Alex: Thanks for listening to ResearchPod.