#deepresearch for the [[psychedelic aesthetics]] paper # Psychedelics, [[Predictive Processing]], and the Relaxation of Priors in [[Perceptual Experience]] ## Introduction Psychedelic drugs have long been noted for their profound effects on perception and consciousness. Recent theoretical models in cognitive science suggest that these effects can be fruitfully understood through the lens of **[[predictive processing]]** – a framework in which the brain is viewed as a prediction engine, constantly generating and updating hypotheses about incoming sensory information. In [[predictive processing]] terms, perception is not a passive reflection of the world, but an active **inference** guided by prior expectations (or “priors”) about what is likely to be out there. Psychedelics, according to emerging theories, **relax or attenuate these priors**, allowing unusual sensory experiences and cognitive flexibility. This discussion will examine how **perceptual priors** function in normal cognition and how psychedelics alter these processes. We will integrate perspectives from philosophy and cognitive science – drawing on thinkers like [[Jakob Hohwy]], [[Karl Friston]], Anil Seth, [[Thomas Metzinger]] and others – and reference relevant empirical findings. Key questions include the **[[philosophical implications]]** of altered [[predictive processing]] (for example, what it means for our notion of reality and veridical perception) and comparisons with other altered states (such as meditation or psychosis) that may similarly affect perceptual priors. We will also consider objections and alternative viewpoints, ensuring a rigorous and scholarly treatment of the topic. ## [[Predictive Processing]] and [[the Role]] of Priors in Perception [[Predictive processing]] (PP) models propose that the brain continuously generates predictions about sensory inputs at multiple hierarchical levels, using prior knowledge to interpret ambiguous stimuli. **Priors** are the brain’s expectations – built from past experience – about what it will encounter. These top-down expectations are constantly matched against bottom-up sensory signals. When sensory input deviates from a prediction, a **prediction error** is generated, which prompts the brain to update its beliefs or adjust the weighting of its predictions. In this Bayesian framework, perception results from a **combination of priors and sensory evidence**, weighted by their precision or certainty. As [[Andy Clark]] concisely puts it, our “perceptual world is a construct that emerges at the intersection between sensory information and priors”. Anil Seth similarly explains that the brain uses “*informed guesswork*,” integrating sensory signals with prior expectations to form its **best guess** of the external causes of those signals. In short, normal perception is **actively constructed** by what the brain expects, as much as by what the senses deliver. ### Priors, Precision, and Perceptual Inference Not all predictions are treated equally: the brain assigns a **precision weighting** to both priors and sensory inputs, effectively tuning how much weight to give each. Precision is akin to confidence or certainty. A highly precise prior will strongly shape perception, potentially overruling ambiguous sensory data; conversely, precise sensory evidence can override a weak prior. This dynamic weighting is crucial for balancing stability with adaptability in perception. For example, our strong prior that faces are convex causes us to see a concave mask as a normal convex face (the hollow-mask illusion), illustrating how robust high-level priors can dominate sensory ambiguity. In general, **priors help constrain perception**, allowing us to resolve noisy or incomplete data quickly by relying on what usually holds true in our environment. Under typical conditions, this is highly adaptive: it leads to perceptual **stability and efficiency**, as we quickly recognize patterns (e.g., identifying a moving shape as a person at a glance, based on prior experience). However, an overly strong prior can also lead to **illusory perceptions** – hearing a phone ring in random noise because you expect it, or seeing patterns that aren’t there – while overly weak priors can leave one captive to erratic sensation with little continuity. Thus, healthy perception maintains a balance: **enough prior constraint to make sense of the world**, but flexible enough to update when predictions prove wrong. ### Hierarchical Prediction and the Self-Model [[Predictive processing]] is inherently **hierarchical**. Lower levels of the cortex make predictions about simple features (edges, tones, etc.), while higher levels predict more abstract or invariant properties (objects, context, narrative). At the highest levels might be very abstract priors about “self” and world – deep assumptions of identity, reality, and continuity. Jakob Hohwy emphasizes that this hierarchy of predictions essentially **encapsulates us in a model of the world** generated from within: the brain is an “inference machine” locked in the skull, only able to test its predictions against the sensory signals it receives. One’s sense of being an independent self, for example, can be seen as the brain’s high-level **model of the organism** – a prior that unifies myriad sensory inputs and memories into the experience of a single, continuous identity. This **self-model** is ordinarily held with very high precision; we feel strongly that we *are* a self that persists over time, distinct from the world. Philosophers like Thomas Metzinger and others have argued that this feeling of selfhood is a kind of internally generated construction – what Metzinger calls the “phenomenal self-model” – which is usually **transparent** (we don’t see it as a model, we just see ourselves as we are). The predictive processing account aligns with this: the unified self may be understood as a **top-down prediction** that binds experiences together, rather than an irreducible essence. This has important implications when we consider altered states: if the brain’s highest-level priors (like the self-model) are changed or relaxed, the very fabric of our subjective reality can shift dramatically. ## Psychedelic Drugs and the Relaxation of Priors Classic psychedelic substances (such as LSD, psilocybin, DMT, and mescaline) interact primarily with serotonin 5-HT_2A receptors in the brain, triggering a cascade of neurochemical effects. Users report vivid sensory distortions, hallucinations, novel thoughts, and even mystical-type experiences. How might these phenomena be explained in predictive processing terms? A converging view among researchers is that **psychedelics relax the brain’s prior assumptions** or reduce the precision weighting of high-level predictions. In essence, psychedelics are thought to **lower the confidence of the brain’s top-down models**, allowing bottom-up sensory information (and previously suppressed signals or random neural noise) to play a larger role in shaping perception. This section explores this idea in depth. ### The REBUS Model: Relaxed Beliefs Under Psychedelics A prominent formulation of how psychedelics alter predictive processing is the **REBUS model** proposed by Robin Carhart-Harris and Karl Friston. REBUS stands for “**Relaxed Beliefs Under Psychedelics**.” According to this model, psychedelics have an **“entropic” effect** on brain activity, increasing neural randomness or complexity, which biologically corresponds to degrading the normal patterns of rhythmic oscillatory activity that encode confident predictions. Functional networks like the **default mode network (DMN)** – normally associated with self-referential thinking and high-level integration – become disintegrated or “decentered” under psychedelics. The net result is that the **precision of high-level priors is relaxed**: the brain’s top-down influences lose some of their grip. Carhart-Harris and Friston describe this as a **“lightening” of deeply held beliefs or assumptions**, making them more permeable to new information. With high-level priors (like one’s ego or worldview) temporarily weakened, **prediction errors from lower levels ascend with less inhibition**, “finding freer register in conscious experience”. In plain terms, sensory signals and latent neural dynamics that would usually be filtered out by expectations can now flood perception and cognition. This explains the **rich sensory hallucinations** – for example, the geometric visuals and novel shapes often reported are thought to arise from intrinsic cortical patterns (especially in visual cortex) being released from top-down constraints. At low psychedelic doses, these effects may remain confined to perception (e.g., visual distortions, enhanced colors), but at higher doses the **disruption of high-level priors** becomes “profound,” accounting for experiences like **ego-dissolution** (a breakdown of the self-model) and deep cognitive or existential shifts. In the REBUS framework, relaxing priors is not just a random loss of structure but can have a **functional role**. By reducing the weight of entrenched priors, the brain enters a more plastic, fluid mode that Carhart-Harris describes as an “**anarchic brain**” state. In this anarchic, high-entropy state, previously rigid beliefs (for instance, pathological beliefs in depression or rigid self-concepts) may become amenable to change. The model draws an analogy to **simulated annealing** in optimization: a system can escape deep, suboptimal energy wells (rigid beliefs) if given a dose of randomness or “heat,” after which it might settle into a new configuration once it cools. Empirically, this aligns with studies showing that psychedelic therapy can lead to **lasting revisions of outlook and beliefs** – for example, reductions in depressive rumination, changes in political or religious perspectives, and increased openness to new ideas. In one study, psilocybin-assisted therapy in treatment-resistant depression not only alleviated depressive symptoms but also led to patients endorsing more optimistic and “realistically optimistic” outlooks on life afterwards. Lyons and Carhart-Harris (2018) reported increases in “nature relatedness” and decreases in authoritarian political views after psilocybin therapy, suggesting a **softening of rigid high-level attitudes**. In summary, psychedelics appear to **“relax the precision of high-level priors or beliefs, thereby liberating bottom-up information flow”**. The result is a state of consciousness where perception becomes less constrained by what is normally expected and more sensitive to the raw (and sometimes chaotic) data of experience. It is as if the brain’s **filters have been lifted**: some longstanding predictive models dissolve, and one experiences both sensory impressions and emergent thoughts with a freshness or unpredictability that can be astonishing. ### Perceptual Changes and Predictive Processing under Psychedelics How do these theoretical changes manifest in concrete perceptual experiences? Many classic phenomena of the psychedelic state can be interpreted within the predictive processing framework: - **Visual Hallucinations and Distortions:** Psychedelics famously cause vivid visual effects – from shimmering colors and fractal patterns to seeing objects breathe or morph. Under PP theory, as high-level visual priors lose precision, **lower-level visual areas (rich in 5-HT2A receptors) start propagating prediction errors** freely. The brain, in trying to make sense of this activity, may settle on novel intermediate predictions, resulting in geometric hallucinations or unusual interpretations of shapes. For example, instead of using the strong prior “walls are static and straight,” the brain confronted with noisy signals might infer “the wall is moving in waves,” leading to the perception of breathing walls. The **increased entropy** in neural activity means the visual cortex is exploring a larger repertoire of patterns than usual, which correlates with seeing a rich variety of spontaneous form and color. Indeed, neuroimaging studies show that under psilocybin, functional connectivity in the visual cortex becomes more diverse and less tied to normal sensory input, paralleling the subjectively reported visuals. - **Reduced Top-Down Stabilization of Ambiguity:** Under normal conditions, strong priors resolve ambiguity – e.g., the hollow-mask illusion where we insist a hollow face must be normal due to prior knowledge of faces. Psychedelics, by weakening priors, might allow one to perceive the raw sensory truth of such ambiguous stimuli. While formal studies are sparse, it is hypothesized that someone on a psychedelic might **see through certain optical illusions** more readily, because the brain is less compelled to impose the usual interpretation. (Interestingly, people with psychosis or schizophrenia, who may also have aberrant prior weighting, are known to be less fooled by the hollow-mask illusion, lending credence to this idea by analogy.) - **Emergence of Novel or Idiosyncratic Percepts:** With loosened priors, the brain may temporarily form **new predictive models** on the fly. Users often report seeing faces or figures in random patterns (enhanced pareidolia), or attributing profound meaning to ordinary stimuli. The predictive mind abhors a vacuum – if top-down models are weakened, it may try creative “guesses” to interpret what it senses. This can yield **inventive but transient constructs**: e.g., seeing a deity in the clouds, or perceiving that sounds have colors (synesthesia is commonly reported under psychedelics). Such phenomena underscore that relaxing priors increases **both sensitivity and susceptibility**: the mind becomes more sensitive to subtleties (picking up patterns normally ignored) but also more prone to false positives (detecting illusory patterns). In predictive coding terms, with reduced prior constraint, **prediction errors are over-weighted**, and the brain will infer explanations even for what is essentially noise. The result can be **visionary experiences** that feel revelatory, though they might be understood as the brain’s heightened creativity in model-building when usual models destabilize. - **Ego-Dissolution and Altered Self-perception:** One of the most impactful psychedelic experiences is “ego dissolution” – a loss of the usual sense of self. From a PP perspective, this corresponds to a **dramatic relaxation of the high-level priors that constitute the self-model**. The feeling that “I exist as an independent self” is normally a deeply entrenched prediction that binds experiences (the brain constantly predicts incoming sensations as belonging to “me,” providing a stable first-person perspective). Psychedelics can disrupt the neural substrates of this self-model – for instance, by desynchronizing the default mode network (involved in self-referential processing). As the prior of self loses precision, users report a dissolution of boundaries: the distinction between self and world blurs, sometimes yielding a mystical sense of unity or oceanic boundlessness. Letheby and Gerrans describe this phenomenology as the **“unbinding” of the self-model** – the normally coherent narrative of self fragments, revealing that our sense of being a unified self was a construction that can come apart. They note that even in intense psychedelic experiences the self-model isn’t destroyed so much as **“decohered”**, allowing one to momentarily realize that the feeling of being a separate, unified ego is generated by the brain’s predictive processes. In other words, ego dissolution can **“disclose the nature of self-awareness”** by showing how contingent it is on active modelling. This aspect of the psychedelic state has fascinated philosophers and cognitive scientists alike, as it provides a window into how the brain constructs the self and what happens when that construction is loosened. ### Cognitive Flexibility, Learning, and Therapeutic Implications Beyond perceptual distortions, the relaxation of priors under psychedelics is thought to engender a state of **cognitive flexibility** that has potential therapeutic benefits. If many psychiatric disorders involve overly rigid predictive models or **entrenched priors** – for example, deeply held negative beliefs in depression, or fixed fear expectations in PTSD – then loosening those priors might allow the brain to update and revise maladaptive predictions. This is a major motivation behind contemporary psychedelic therapy research. Carhart-Harris et al. have suggested that disorders like depression or addiction are marked by **pathologically strong priors** (e.g., “I am worthless and nothing can change that” in depression) encoded with high precision in the brain’s networks. Under psychedelics, those high-level priors may temporarily relax, offering a window during which new perspectives can be introduced and integrated. A recent computational account by Letheby and colleagues argues that psychedelics make high-level self-related priors “*more revisable by bottom-up input*”. In therapy, this bottom-up input could come from positive environmental signals or therapeutic guidance – for instance, reassurances of safety, or suggestions of new ways of viewing one’s life. The combination of a **pharmacologically induced plasticity** (via relaxed priors) and supportive cognitive input is hypothesized to enable profound, lasting change, essentially **rewiring the predictive brain** toward healthier beliefs. Empirical research supports the idea of enhanced flexibility and learning under psychedelics: - **Emotional Breakthroughs:** Many patients in psychedelic therapy report that the experience allowed them to confront emotions or memories in a new way, suggesting a break from avoidance priors (e.g., dropping the habitual prediction of intolerable fear when recalling trauma, thus allowing reprocessing). Such breakthroughs may reflect the brain’s ability to update high-level priors about safety, self-worth, or the meaning of traumatic memories when they are briefly unlatched from their usual constraints. - **Personality Changes:** A striking study by MacLean et al. found that a single high-dose psilocybin session, especially one inducing a “mystical experience,” led to **lasting increases in the personality domain of Openness** (one of the Big Five traits) more than a year later. Openness is associated with imagination, aesthetic appreciation, and broad-mindedness. That such a core personality trait shifted suggests a lasting enhancement in cognitive flexibility and willingness to revise one’s mental models of the world – consistent with the idea that certain priors (perhaps related to how one approaches novel ideas or experiences) were relaxed and recalibrated during the psychedelic experience. Participants self-reported becoming more creative and open-minded, which aligns with the notion of an **expanded repertoire of predictions** the mind is willing to entertain (versus a previously narrow set of expectations). - **Therapeutic Outcomes:** Clinical trials of psilocybin for depression and anxiety (especially in life-threatening illness) have shown not only symptom reductions but enduring changes in outlook. For example, patients often describe a shift from rigid despair to a renewed sense of possibility and meaning after psychedelic therapy. One analysis linked psilocybin sessions to increases in psychological insight and **“realistic optimism”** – patients became less pessimistic in a manner that was not delusional but rather a more balanced appraisal of life circumstances. This reflects a recalibration of high-level priors about life and self: pessimistic priors were rolled back, allowing evidence (like personal strengths or support from others) that had been down-weighted to be acknowledged, resulting in more optimistic yet *realistic* beliefs. In Bayesian terms, previously overweighted negative priors lost some precision, so that new data could update the posterior belief to a healthier outlook. It is important to note that **context (“set and setting”)** plays a crucial role here. Relaxing priors can make the mind more malleable not only to positive inputs but also potentially to **suggestion or external influences**. In a supportive setting, this can be harnessed for therapeutic good – guiding the malleable predictive mind toward useful insights – but in a negative or chaotic environment, it could just as well lead to distress or the reinforcement of unhelpful ideas. Thus, while psychedelics pharmacologically induce a state of enhanced learning and openness, **the content of what is learned depends on environmental feedback**. This underscores that predictive processing under psychedelics doesn’t occur in a vacuum: the **brain still seeks explanations** for its altered signals, and it may grab onto cues from surroundings or the therapist to form those explanations – yielding very different experiences based on context. ## Philosophical Implications: Realism, Veridicality, and Mind-Construction The alteration of predictive processing by psychedelics raises profound philosophical questions. If our normal perception is a kind of controlled hallucination shaped by priors, then psychedelic perception is an *even less constrained hallucination*. What does this imply about the **nature of reality and our access to it**? Are psychedelic experiences mere distortions – departures from veridical perception – or might they reveal something about the mind and world that is normally hidden by our cognitive habits? ### Perception, Hallucination, and Veridicality Philosophers of mind often use hallucinations as test cases for theories of perception. The predictive processing view blurs the line between veridical perception and hallucination: both are constructed interpretations of sensory data, differing mostly in the **degree of top-down influence** and whether there is a fitting external stimulus. Under psychedelics, one could say the brain’s modeling process is laid bare – it **hallucinates more wildly**, confirming the adage that *“we are hallucinating all the time; when we agree about our hallucinations, we call it reality.”* Seth and Friston have popularized this notion, suggesting that even our sense of self and world in normal life is a kind of fantasy grounded by sensory constraints. Psychedelics then can be seen as **altering the balance**: reality is experienced with a new bias, less guided by prior beliefs and more by unconstrained sensory exploration or spontaneous neural dynamics. One implication is a form of **philosophical skepticism or idealism**: if both ordinary and psychedelic perceptions are constructions, what does “real” mean? Some users feel that psychedelics opened their eyes to a *“deeper reality”* or a “spiritual dimension” normally filtered out. Aldous Huxley famously called the brain a “reducing valve” that limits consciousness, and suggested that psychedelics open the valve, letting in an influx of reality from which we are normally protected. Predictive processing can reinterpret this: the brain normally filters (via priors) to keep perception tuned to practical needs, whereas psychedelics remove some filters, letting in not so much a mystical reality as **more raw data and entropy**. Whether this counts as “deeper reality” is debatable. It may simply be a *less structured* one. The **entropic brain hypothesis** explicitly posits that psychedelic state is akin to a **“primitive” state of consciousness**: more chaotic, akin to infant cognition or dream states, lacking the refined order that evolution and development have imposed for good functional reasons. From this perspective, normal waking consciousness is *less veridical in some ways* (it’s a tightly constrained model, not an image of the world as it is), but also *more useful*, having sacrificed some breadth of information for stability and accuracy in relevant features (like object permanence, logical consistency, etc.). Psychedelic consciousness, by contrast, has **broader information (more entropy)** at the expense of coherence and pragmatic accuracy. Thus, one philosophical view is that psychedelic perception is **not straightforwardly “seeing reality” unfiltered**, but a trade-off: it reveals the *constructedness* of experience by temporarily suspending some constructions. It highlights aspects of mind and world (e.g., the ubiquity of uncertainty or the latent richness of stimuli) that are usually hidden by automatic predictions. In doing so, it can produce **insights about the mind** – for instance, recognizing that what we normally take as reality is just one interpretation. As Letheby notes, a benefit of psychedelic experience is that it can make users appreciate that **“our normal experience of unity [of self and world] depends on a modeling process”** – an insight into the mind’s workings. However, this does not require us to believe that the content of every hallucination is *literally true*. We can maintain an epistemically cautious stance: the psychedelic traveller may *feel* they have encountered divine beings or ultimate truths, but a naturalistic interpretation would say these are mind-generated experiences that might be symbolically meaningful without corresponding to external entities. The **Comforting Delusion Objection (CDO)** in the philosophy of psychedelics addresses this tension: it asks whether the therapeutic or spiritual gains from psychedelics come at the cost of believing comforting falsehoods – essentially, are psychedelics just **“foisting a comforting delusion”** on people?. For example, a terminally ill patient might lose fear of death after a psilocybin session because they become convinced of an afterlife or some cosmic purpose – beliefs that skeptics might view as unwarranted, albeit comforting. Philosopher Chris Letheby tackles this objection by arguing that **psychedelic benefits need not depend on false beliefs**. He suggests that many psychological insights from psychedelics are about oneself and one’s patterns (e.g., realizing one has been stuck in a rigid narrative, or rediscovering a sense of connection) and do not require taking the metaphoric or hallucinatory content at face value. In predictive processing terms, the *form* of the experience (temporarily suspending certain priors) can be invaluable for reconfiguring the mind, even if the *content* (say, a vision of a guiding spirit) is interpreted symbolically rather than literally. Letheby’s stance is a kind of pragmatic realism: the experience is **psychologically real** and can reveal *personal truths* (how one feels, what one values, etc.), even if it doesn’t provide new *metaphysical truths*. This position tries to preserve the idea that psychedelics can be **epistemically beneficial** (they can lead to new knowledge or understanding) without endorsing magical thinking. In fact, some argue that by exposing the brain’s propensity to fabricate experiences, psychedelics can foster a form of **epistemic humility** – one sees how easily the mind can be fooled or how one’s perspective can radically shift, which may lead to a more cautious, open-minded approach to what one considers true. ### Realism vs. Constructivism: Are Psychedelic Insights “Real”? A deeper philosophical issue is the debate between **naïve realism** and **constructivism** about perception. Naïve realism holds that we perceive the world directly as it is (with perhaps occasional errors), whereas constructivism (especially under the PP framework) holds that perception is always a *constructed guess* about reality. Psychedelic experiences, which can be extremely bizarre or “impossible,” challenge a naïve realist view because they show how drastically perception can depart from consensus reality while still feeling unquestionably real to the experiencer. They underscore the **constructive nature of perception**. Cognitive scientist Anil Seth notes that when the brain’s constraints are loosened, we get a taste of how our normal sense of an external, stable world is something the brain actively creates. However, even within a constructivist view, one can ask: *are there times when the brain’s guesses under psychedelics might be **more veridical** in some respects?* For instance, some users report an enhanced perception of actual sensory details (e.g., noticing tiny patterns or textures they normally overlook). This could be seen as **perception becoming more sensitive to raw input** – with top-down suppression reduced, subtle signals might breakthrough. A mundane example: you might notice intricate features in a piece of art or music on LSD that you normally miss; the drug can make the familiar seem novel. If those features were truly there (just usually filtered out), one could argue psychedelics provided a kind of **epistemic gain** – a veridical perception that everyday priors usually mask. Another domain is **emotional or introspective truth**. Many people describe confronting “the truth about myself” under psychedelics – e.g., recognizing they have been in denial about a relationship, or suddenly seeing the extent of their addiction. These realizations can be considered **veridical insights** into psychological reality. Predictive processing again provides a framework: normally, defense mechanisms might serve as **protective priors** (the brain predicts away certain uncomfortable truths, keeping them unconscious or reinterpreting evidence to fit a preferred self-image). Psychedelics, by relaxing those self-related priors, might allow repressed prediction errors (signals that something is wrong in one’s life) to surface into awareness. The “revealed truth” in such cases might be entirely real (for example, an alcoholic sees clearly the harm they are doing to their family). So, while the **sensory content** of psychedelic hallucinations may often be fantastical, the **cognitive content** (what one learns about oneself or reality) can be accurate. This aligns with Letheby’s argument that psychedelic therapy’s efficacy does not require false beliefs: patients can come away with genuine knowledge, such as a new understanding of their relationships or values, without having to accept metaphysical delusions. Philosophically, one might conclude that psychedelics are a tool for exploring the **space of possible minds** – they highlight that our ordinary mind is just one mode of constructing reality. This raises a form of **meta-philosophical reflection**: if the brain can have such radically different experiential states, any claims to have a final grip on “reality” should be tentative. Some theorists even suggest that comparing normal, psychedelic, dream, and pathological states can help **triangulate what aspects of experience are invariant and which are variable**, informing theories of consciousness and the mind-brain relationship. In summary, the relaxation of priors via psychedelics has dual epistemic implications: it can produce *misleading appearances* (hallucinations or delusions if taken literally), yet it can also *illuminate the process of appearance-making itself* and allow certain truths to emerge (by breaking mental gridlock or allowing new data into one’s models). The challenge and excitement for both philosophy and cognitive science is to separate the wheat from the chaff – to understand which aspects of psychedelic experience might constitute **genuine learning or insight** and which are best treated as imaginative phenomena or artifacts of a disordered predictive system. ## Comparisons with Other Altered States of Consciousness Psychedelic states are not the only conditions that involve changes to the brain’s predictive machinery. It is illuminating to compare psychedelics with other altered states – such as **meditative states and psychosis** – which also seem to affect the balance of priors and sensory input, albeit in different ways. These comparisons can deepen our understanding of what is unique to psychedelics and what is part of a broader spectrum of mind/brain dynamics. ### Meditation and the Trained Modulation of Priors **Meditation**, particularly forms of mindfulness or contemplative practice, shares with psychedelics an interest in altering habitual patterns of mind. Advanced meditation practitioners often report changes in perception: a heightened present-moment awareness, reduction in the sense of a separate self, and a lessening of automatic cognitive biases. Unlike psychedelics, which chemically induce changes, meditation is a **trained, gradual reconfiguring of predictive processing** through sustained practice. From a PP perspective, meditation can be seen as a way to **intentionally adjust precision weightings** and the engagement of priors. Some recent models (e.g. work by Ruben Laukkonen and others) describe meditation in predictive processing terms. In focused-attention meditation (such as concentrating on the breath), one **increases the precision weighting of a chosen sensory anchor (the breath)** while down-weighting other predictions and inputs. This has the effect of **tuning out distractions** – by assigning low precision to stray thoughts or sensations, they don’t capture attention. This is essentially the inverse of normal perception: instead of automatically letting the brain predict and react to every stimulus, the meditator practices **selective prediction suppression**, maintaining focus. In open-monitoring or choiceless awareness meditation, the practice shifts: one learns to **“release precision” from any particular content**. All experiences (sounds, thoughts, sensations) are observed without attaching high precision (importance) to them. This can lead to a state where **no prediction or input completely dominates** – a finely balanced state of equanimity. Meditation texts call this “bare attention” or “non-judgmental awareness,” which maps to a mind that **neither clings to priors nor is drowned in sensory chaos**, but simply witnesses phenomena arising and passing. At deeper levels, long-term meditation can result in experiences of **“non-dual awareness”** or unity, which sound reminiscent of the ego-dissolution under psychedelics. Laukkonen describes nondual meditation as “creating conditions that enable the mind to let go of its habitual, predictive tendencies”. Essentially, as meditation training progresses, practitioners learn to **quiet the predictive mind** – the constant commenting, judging, anticipating begins to fade. In PP terms, one interpretation is that the **brain’s model-making apparatus is stilled or seen through**, and thus consciousness can encounter a state of minimized mental construction (sometimes described as pure awareness or emptiness). Interestingly, this is like a **controlled relaxation of priors**, achieved not by a drug but by mental discipline. One difference is that meditation tends to produce clarity and calm (precision on nothing in particular, but awareness of everything), whereas psychedelics produce a flood of content (many things become salient at once). This is captured by an insight from a conversation with Laukkonen: “Where psychedelics loosen cognitive patterns and allow for new kinds of predictions, meditation is more about dampening the prediction system altogether”. Psychedelics add noise and novelty; meditation reduces bias and eventually even habitual model-making. Despite these differences, there are intriguing overlaps: both states can lead to diminished **self-prioritization**. Meditators cultivate non-attachment to the self (realizing thoughts like “mine” or “about me” are just thoughts), which weakens self-related priors over time. Psychedelics can do this suddenly, as discussed, by temporarily knocking out the self-model. Moreover, both have been associated with increased **cognitive flexibility and creativity**. Meditation practice has been found to reduce cognitive rigidity and increase openness to experience in some studies, aligning with the idea that by “loosening the grip” of habitual predictions, one can respond more freely and creatively to situations. In fact, as one quote from Laukkonen suggests, meditation’s long-term trait effect may be to **“reduce the sway of biases, allowing for a more flexible, creative, rich counterfactual capacity”**. In everyday terms, this means a meditator might become less bound by their past conditioning (priors) and better able to see things as novel or from multiple perspectives – not unlike how psychedelics can break someone out of their usual mental grooves. However, meditation typically lacks the *sensory fireworks* of psychedelics. The key difference is **control and integration**: meditation is a slow re-tuning of predictions with continuous self-observation, making it less overwhelming and more integrated into normal functioning. Psychedelics are like a sudden perturbation – extremely effective at shaking the system, but also chaotic and less directed. Interestingly, some research suggests that **experienced meditators under psychedelics** often navigate the experience differently, possibly because they have skills to remain equanimous amid the deluge of prediction errors. This might result in especially powerful experiences of insight or non-dual awareness with less anxiety. In summary, meditation and psychedelics both demonstrate that the predictive mind is *modifiable*. Meditation shows a *top-down* route (using the mind to change itself gradually), while psychedelics provide a *bottom-up* biochemical route (changing global brain function to force new states). Both can relax certain priors (particularly self-related and perceptual biases) and thereby alter conscious experience profoundly. They lie on a continuum: one could say a meditator systematically lowers the precision on the “self” prior over years of practice, occasionally experiencing ego-transcendence, whereas psychedelics knock it out in one afternoon. The results have similarities – such as feelings of unity or insight into one’s mental processes – but also differences in stability and context. This comparison highlights that **relaxing priors is a general principle** that can be achieved in multiple ways, not solely through drugs. ### Psychosis and Hallucination: Pathological Priors A very different comparison can be made with **psychosis**, as occurs in schizophrenia and related disorders. Psychedelic drugs are sometimes called “psychotomimetics” because they can mimic certain aspects of psychosis (especially hallucinations and unusual beliefs). Indeed, historically LSD was used in research to understand schizophrenia. Predictive processing provides a framework to compare these states: both psychedelics and psychosis involve **aberrant prediction error dynamics** and **mis-weighting of priors**. However, the patterns of aberration and the outcomes can differ. In some predictive coding models of schizophrenia, it’s proposed that patients have **diminished precision on high-level priors and/or inflated precision on sensory prediction errors**. Essentially, the brain in psychosis might not trust its priors enough, or it might assign too much significance to random sensory fluctuations. Adams, Stephan, and Friston (2013) describe psychosis as a state where the balance between top-down and bottom-up is “biased toward the likelihood (sensory input) and away from the prior,” resulting in **over-weighting of prediction errors**. This can explain why a person with schizophrenia might find meaning in noise (hearing voices in the wind, or feeling that trivial events have deep personal significance) – their brain is less constrained by what normally *should* be meaningful and so everything potentially is. It also aligns with the known observation of reduced susceptibility to some illusions in schizophrenia (since strong priors like the face convexity assumption may be weaker, the raw sensory truth – the mask is hollow – is seen). If that sounds very much like what we described under psychedelics, it’s because there is considerable overlap: psychedelics pharmacologically induce a similar reduction in high-level prior precision. One might say psychedelics produce a *controlled* (temporary and known) variant of certain psychosis-like computational changes. Both states feature **hallucinations** (percepts without external stimuli) and sometimes **delusions** (false beliefs or interpretations). The difference often lies in *degree and context*. Psychedelic hallucinations are usually recognized by the user (at least afterwards) as drug-induced and not “real” in a consensus sense, whereas psychotic hallucinations are woven into the individual’s reality without such insight. This could be partly because in psychosis the aberrant processing is chronic and the individual’s only frame of reference, whereas the drug state is acute and surrounded by an otherwise normal personality that can reflect on it later. There is also a hypothesis that psychosis may involve a combination of **weak low-level priors and overly strong high-level priors** in different ways. For instance, one theory is that early in schizophrenia, sensory chaos (from weak priors or high noise) leads the individual to develop *compensatory* strong high-level priors – essentially, the brain tries to make sense of the chaos by imposing a narrative (resulting in delusions). For example, if a person experiences inexplicable coincidences or percepts (due to aberrant error signals), they might formulate a strong prior like “I’m being persecuted by a conspiracy” to explain it, and then that prior itself becomes rigid and resistant to correction. In predictive terms, they oscillate between too little and too much prior influence, but not at the right levels: trivial sensations gain significance, and then broad incorrect theories solidify. Psychedelics, in contrast, tend to keep even high-level beliefs flexible *during* the state (one reason being the time-limited action and perhaps the retained metacognitive awareness that one is under a drug’s influence). In fact, some have used psychedelics as a model to study certain psychosis symptoms in healthy volunteers, with mixed success – classic psychedelics produce more visual hallucinations and less of the complex paranoid delusions typical of schizophrenia, likely because they affect sensory hierarchies strongly and don’t inherently implant specific fixed beliefs. Comparatively, one might say **psychosis is an *uncontrolled disordering*** of predictive processing, whereas a **psychedelic trip is a *controlled, temporary* disordering** of it. The “set and setting” principle – having intention, guidance, and a safe environment – can anchor a psychedelic user such that even as their perceptions go haywire, they can navigate it and later integrate the experience. A person undergoing psychosis lacks that context and often the changes are gradual and unnoticed by them until they are fully immersed in a delusional reality. Another altered state to mention is **dreaming (REM sleep)**, which Carhart-Harris’s entropic brain theory likens to the psychedelic state. Dreams are internally generated experiences where we accept bizarre scenarios as real until we wake. In dreams, the brain’s critical reflective faculties (high-level priors about what is possible) are reduced, allowing highly imaginative content. Psychedelics share some dreamlike qualities (intensified imagery, fluid identities), though with the key difference that one is awake and, in many cases, can remember and reflect on the content with greater lucidity. This makes psychedelics a unique hybrid of conscious insight and dreamlike looseness of priors. In summary, examining meditation, psychosis, and dreaming alongside psychedelics highlights that **perception and cognition can vary dramatically depending on the tuning of prior expectations**. Psychedelics stand out as a tool that, acutely and in a dose-dependent manner, can dial down the influence of priors across the perceptual and cognitive hierarchy. Meditation achieves some of the same end states but through effort and over longer durations. Psychosis and dreams show that the brain can also fall into altered predictive states pathologically or naturally. These comparisons reinforce confidence in the predictive processing framework itself: the fact that one explanatory scheme (priors vs. prediction errors) can illuminate such a range of phenomena is seen as evidence of its robustness. Jakob Hohwy and others have argued that this unifying power – explaining normal perception, illusions, hallucinations, self-awareness, etc., under one umbrella – is a strength of predictive processing (though, as we will address, not everyone is convinced it’s the whole story). ## Objections and Alternative Perspectives While the idea that psychedelics relax perceptual priors within a predictive coding framework is compelling and increasingly influential, it is not without challenges and alternatives. We should consider some objections and other angles: **1. Is Predictive Processing an Oversimplified Account of Psychedelics?** Critics like Matteo Colombo (2022) have questioned whether predictive processing offers a genuinely explanatory account of psychedelics’ effects or just a high-level description. One concern is that PP theory might be *too vague or unfalsifiable* in this context: saying “priors are relaxed” could be a catch-all that doesn’t uniquely predict specific effects. For example, can the PP model predict why LSD produces complex geometric visuals as opposed to random noise? Or why serotonin 2A agonists (psychedelics) differ phenomenologically from, say, dissociative drugs that also induce hallucinations via NMDA receptors? A PP advocate might respond that the specifics can be accounted for by noting where in the hierarchy the effects occur (5-HT2A receptors are densely expressed in visual cortex and high-level cortical layers, hence pronounced visual effects and broad cortical network disruption). However, the **devil is in the details**: PP is a broad conceptual framework and often needs to borrow from neuroscience details (e.g., the entropic brain, specific receptors) to fully explain psychedelic phenomenology. Some object that the theory can become circular – we identify that psychedelics cause increased neural entropy and then equate that to “lowered precision of priors,” but how do we independently verify this lowering of precision? Ongoing research attempts to link specific neural signatures (like changes in brain waves or connectivity) to the computational notion of precision, but it’s challenging. **2. Alternative Neural Models:** Others propose different (though not mutually exclusive) models for psychedelic action. For instance, **Reinforcement Learning (RL)** frameworks might emphasize how psychedelics affect neuromodulators (like serotonin and dopamine) that govern learning rates and exploration. Colombo’s commentary suggests that perhaps a RL model (with serotonin affecting reward/punishment prediction errors) could explain therapeutic outcomes without invoking hierarchical perception per se. Another approach is to focus on **neuroplasticity**: psychedelics (especially with repeated use or in animal studies) promote neural growth and synaptic changes. One might argue that the **psychological effects stem from increased plasticity** – essentially a window of rewiring – rather than Bayesian inference per se. Yet, from a broader view, these may be complementary: increased plasticity could underlie the brain’s ability to revise priors more readily after a psychedelic experience. **3. Are All Effects Due to Relaxed Priors, or Are Some New Priors Introduced?** The “relaxed priors” story might explain sensory distortions and flexibility, but what about the content of specific **visions or ideas** people have? Sometimes, psychedelic experiences seem **structured, not just free-form chaos**. For example, why do so many people see symmetrical patterns or mandalas? One could argue there are **common priors in the visual system** (perhaps symmetrical structure or certain frequency patterns) that become manifest when the system is disinhibited. Or consider the often-reported “encounter with an entity” on high-dose DMT – does the brain simply invent a supernormal agent when presented with incomprehensible input? Possibly, yes: the brain’s hyperactive pattern-recognition might default to personification (a high-level prior: “if there is complex meaningful information, assume an agent or voice”). In that case, rather than *relaxing* a prior, the psychedelic might be causing an *over-activation* of a certain archetypal prior (like the brain’s built-in models of faces or beings, which are normally constrained). This suggests that the relationship is complex: some priors may relax while others temporarily dominate in explaining the flood of errors. The current frameworks do acknowledge that **not all priors are equal** – the highest-level self and worldview priors relax, but mid-level priors can get weirdly recruited to make sense of novel signals (hence the brain might, say, use a “entity” hypothesis to explain an internal signal). **4. The Role of Context and Suggestion:** As noted, a psychedelic experience is heavily influenced by **set and setting** – the mindset and environment. This means that what the brain predicts or experiences under the drug can be guided externally. For example, in a supportive setting, a person might have a joyful transcendent experience; in a threatening setting, a terror-filled paranoid experience. This suggests that rather than simply “reducing all priors,” psychedelics might **increase the volatility or uncertainty**, allowing **contextual cues to shape new temporary priors** more strongly. In technical terms, one could say psychedelics increase the brain’s *learning rate* or the reliance on current inputs. If a therapist suggests “you may feel a sense of peace,” the individual might indeed feel that, because their brain, with loosened internal priors, leans into that suggestion as a guiding prediction. This could be seen as **enhanced suggestibility**. Hypnosis research parallels might be drawn – hypnosis is thought to involve focused attention and reduced top-down monitoring, making people more open to suggestion. Psychedelics are sometimes likened to a biochemical facilitation of a suggestible, plastic state. The predictive coding story can incorporate this (suggestions are external inputs that become priors in the moment), but it underlines that **psychedelic effects are not solely generated endogenously**; they are an interplay of brain, drug, and environment. A comprehensive view should integrate them, noting that **the priors that matter aren’t just internal “data priors,” but also beliefs about context**. Hohwy’s view of the brain as trying to explain inputs fits here: the context (therapeutic setting vs. noisy club) provides different inputs that the brain under psychedelic influence will attempt to integrate into its predictions – yielding very different experiences. **5. Not All Hallucinations Are the Same:** Comparing psychedelics to psychosis again, some critics might point out that certain features of schizophrenia (like chronic auditory hallucinations or systemic delusions) have aspects not typically seen in psychedelics. For example, schizophrenic hallucinations often have a repetitive, internally consistent character (e.g., the same voice haranguing the person), whereas psychedelic hallucinations are more transient and fantastical. This could indicate differences in predictive processes: in schizophrenia, some maladaptive priors become *too strong* (like a persistent voice model that doesn’t extinguish), whereas psychedelics rarely result in a single hallucinated narrative taking hold for long (the content shifts rapidly as attention shifts). This might be because psychedelics reduce the stability of any one high-level hypothesis – which is mostly true, but occasionally if someone on a psychedelic does latch onto a scary idea, it can feel very real for a time (a “bad trip” delusion, like thinking one is dying or time has stopped). Such temporary delusions under psychedelics show that **false high-level priors can still form** if circumstances push them (e.g., if one interprets a racing heart as “I am dying,” that might become a strong prediction leading to panic). The difference is that in a healthy person the drug wears off and normal prediction error learning reasserts that “No, I didn’t actually die; it was the drug.” In psychosis, there’s no such off-switch. **6. Free-Energy Principle: Too Abstract?** Karl Friston’s broader **Free-Energy Principle (FEP)**, of which predictive coding is one aspect, posits that brains minimize a quantity (free-energy) related to prediction error. Some philosophers and scientists find this principle *overly broad* or even unfalsifiable, since it can be used to explain nearly any organism behavior by re-describing it as free-energy minimization. In context, one could ask: does saying “the brain on psychedelics still tries to minimize prediction error, but now with changed precisions” actually explain *why* certain conscious experiences happen, or is it just relabeling? For a satisfying explanation, more specific mechanistic bridges are needed (like how 5-HT2A receptor modulation of cortical pyramidal neurons leads to desynchronized brain waves that correspond to reduced precision weighting). Fortunately, the field is moving in that direction, combining **neuropharmacology, neuroimaging, and computational modeling** to flesh out the story. For instance, one study might correlate levels of network entropy with self-reported ego dissolution intensity, lending quantitative support to the REBUS claim that higher entropy (lower precision) at high levels correlates with ego dissolution. So far, evidence broadly supports that psychedelics induce increased neural variability (entropy) and decrease the modularity of brain networks (networks become more integrated and less segregated), which is consistent with a breakdown of the usual predictive hierarchies. **7. Philosophical Objection – Is the Mind Just a Bayesian Machine?** Some philosophers might challenge the predictive processing paradigm itself as a totalizing explanation of mind. While it accounts for perception and action well, what about the qualitative depths of conscious experience? Does relaxing priors fully account for the **qualia** and the existential aspects of psychedelic trips? For example, the raw feeling of awe or the sense of profound significance that often accompanies psychedelic experiences might not be easily quantified in predictive terms. One could attempt an explanation: perhaps awe is what the mind feels when confronted with high uncertainty that is nonetheless positive or safe – a kind of reward for exploring new predictions. And a sense of significance could come from the brain’s tagging of an experience as learning-rich or prediction-error-rich (“pay attention, this is important”). But these are speculative. Another angle: Some argue that certain psychedelic insights border on **philosophy or metaphysics** (people report intuitions about consciousness, the universe, etc.). Can our Bayesian brain theory account for the *content* of those? Possibly it can, by seeing them as the brain’s best guesses to reconcile an utterly novel internal context. However, this edges into territory where **cognitive content** (like beliefs about life) intersects with **subjective experience**. Letheby’s work essentially tries to naturalize those insights – to say: yes, you felt a sense of cosmic unity, but that doesn’t mean you contacted a cosmic consciousness; rather, your brain underwent a transient breakdown of the self-other prior, giving a *feeling* of unity which you can interpret in naturalistic terms. In conclusion of the objections: the predictive processing account of psychedelics is a powerful integrative framework, but it is **not the only perspective**. It should be seen as complementary to neural, psychological, and phenomenological levels of analysis. As the science advances, this framework will be tested and refined. If certain predictions (no pun intended) fail – for instance, if we found that increasing some priors actually intensified certain psychedelic effects rather than dampening them – then revisions would be needed. So far, however, the **convergent evidence from computational theory, neuroimaging, and subjective reports** has lent weight to the notion that an essential part of what psychedelics do is indeed reduce the weight of prior assumptions in the brain. The remaining debates revolve around how complete this explanation is and how to integrate it with other levels of understanding. ## Conclusion The predictive processing framework provides a rich theoretical lens through which to understand the transformative effects of psychedelic drugs on perceptual experience. By viewing the brain as a prediction-bound organ – one that normally negotiates reality by imposing learned expectations – we can see psychedelic states as a profound experiment in **what happens when those expectations are loosened**. Psychedelics like LSD or psilocybin relax the precision of high-level priors and beliefs, plunging the brain into a mode of heightened entropy and surprise. In this mode, **perception becomes less anchored**: the familiar world may dissolve into unprecedented patterns, and the self that anchors experience may even melt away. Through this temporary “deweighting” of priors, the **brain’s model of reality becomes labile**, allowing new forms of experience – some wondrous, some terrifying, and often indescribably novel. From a cognitive science perspective, this underscores the delicate balance our brains normally maintain. The psychedelic experience, in revealing the mind’s constructive nature, simultaneously **illuminates the mechanisms of normal perception**. The fact that a simple molecule can drastically alter the world we experience speaks to how much of that world is generated from within. Thinkers like Hohwy, Friston, and Seth have, in different ways, prepared us for this conclusion: the world we perceive is always partly a controlled hallucination. Psychedelics simply make the hallucination less controlled. In doing so, they have implications for philosophy – challenging our notions of reality and self – and for therapy, by allowing the revision of maladaptive predictive models in conditions like depression or PTSD. We also saw that psychedelics are not unique in modulating priors. Meditation can quiet them, psychosis can disorder them, and dreaming can suspend them. These comparisons place the psychedelic state in a broader context of brain-mind states, highlighting that **our subjective reality can vary dramatically with different tunings of prediction and error**. What makes psychedelics particularly remarkable (and useful, when handled properly) is the combination of depth and transientness: they can induce very deep alterations (touching even the core of self-model) yet are temporary, allowing one to return and integrate the experience. This makes them a powerful tool not just for potential clinical use but also for probing the **nature of consciousness** itself. In terms of realism and veridicality, a balanced view has emerged: Psychedelic experiences are **real as experiences** – they reveal the mind’s capabilities and hidden corners – but we should be careful in interpreting their content literally. The relaxation of priors means the brain entertains possibilities it normally wouldn’t; among those are profound insights (e.g., recognizing one’s unhealthy patterns) as well as illusions (e.g., perceiving cosmic telepathy). The task, both for individuals and the scientific/philosophical community, is to **differentiate insights from illusions**. This is where a dialogue between empirical research and philosophy is valuable. Empirical research can track changes in brain and behavior, indicating which aspects of the experience correspond to beneficial, reality-congruent changes (for instance, improved well-being, more accurate self-perception), whereas philosophy can help interpret the subjective reports and integrate them into a coherent understanding of mind and world. As our discussion shows, thinkers beyond the usual trio of Clark, Letheby, and Lyon(s) have enriched this discourse. Jakob Hohwy’s emphasis on the neurocomputational underpinnings of perception, Karl Friston’s formal models of brain self-organisation, and Anil Seth’s eloquent demonstrations of perception as hypothesis have all set the stage for applying these ideas to psychedelics. Meanwhile, philosophers like Thomas Metzinger and others have provided concepts (like the transparent self-model) that mesh remarkably well with observed psychedelic phenomena like ego dissolution. The convergence of these ideas is a testament to the **interdisciplinary nature** of the inquiry: neuroscience, psychology, and philosophy are all needed to grapple with the full picture. In closing, the predictive processing approach to psychedelics offers a **unifying narrative**: Under the influence of these substances, the brain’s inference machinery is perturbed – it relaxes its prior assumptions and opens itself to surprise. In that state of heightened uncertainty, new patterns of perception and thought emerge. Some are creative, insightful, and healing, others are confabulatory or overwhelming. The outcome depends on how this state is entered (mindset), guided (environment), and exited (integration of new experiences). As research advances, we are likely to see a more nuanced map of which priors (at which levels of the cortical hierarchy) are affected by different substances and how this links to specific phenomenological facets. Already, models like REBUS provide a compelling sketch, arguing that by **lightening the mind’s predictions** we can temporarily revisit a more unconstrained mode of cognition – one that holds both promise and peril. Ultimately, studying psychedelics through predictive processing not only helps us understand these unusual experiences, but also **reflects back on our understanding of normal cognition**. It reminds us that what we take to be reality is, in a very real sense, *a construct of our brains* – a construct that can be bent or broken under certain conditions. Psychedelics happen to show us another side of that apparatus by relaxing the rules by which it usually operates. In doing so, they invite continued exploration – empirical, therapeutic, and philosophical – into the mysteries of perception, mind, and reality as a whole.