Ch. IV — Systems neuroscience and mechanism

Chapter IV of Post-2010 Psychedelics: An Expert-Panel Review. For the executive summary and full table of contents, start there.

Abstract. This chapter reviews the post-2010 systems-neuroscience evidence for classic serotonergic psychedelics, organised along three interlocking axes: (i) the acute cortical state — heightened entropy, broadband desynchronisation, decreased modularity, increased global integration; (ii) circuit-level reorganisation, principally of the default mode network and thalamocortical, claustral, and limbic loops; and (iii) the post-acute neurobiology, dominated by dendritic-spine remodelling under the psychoplastogen framework and, in 2023–2026, by the intracellular 5-HT2A thesis from the Olson laboratory. The bridge from receptor binding (Chapter III) through acute network reconfiguration to lasting cellular and tract-level change is more clearly articulated in 2026 than at any previous point, but several joints in that chain remain hand-wavy, and the field’s reliance on small-sample human imaging and rodent surrogate behaviour is a continuing source of inferential strain.

4.1 The acute psychedelic brain state

The acute neurobiology of classic psychedelics, as characterised across roughly fifteen years of human neuroimaging, is consistent enough to be summarised as a syndrome rather than a list of findings. Five features recur across modalities and across compounds (psilocybin, LSD, ayahuasca/DMT): (1) cortical hyperexcitability at the cellular level, (2) broadband desynchronisation of low-frequency oscillatory power, (3) increased measures of signal complexity and entropy, (4) increased between-network and global functional connectivity, and (5) decreased within-network connectivity and modular segregation.

The cellular-to-systems translation is most cleanly demonstrated in the Muthukumaraswamy et al. 2013 magnetoencephalography study of intravenous psilocybin, which combined task-free MEG with dynamic causal modelling.¹ Spontaneous oscillatory power was reduced from 1–50 Hz in posterior association cortex and from 8–100 Hz in frontal association cortex, with the largest decrements over default-mode hubs. Dynamic causal modelling resolved this paradox — broadband desynchronisation despite increased 5-HT2A-mediated drive — as posterior cingulate desynchronisation explained by increased local excitability of deep-layer pyramidal neurons. The Carhart-Harris et al. 2016 multimodal LSD study replicated the broader pattern (decreased alpha power, decreased DMN integrity, increased global connectivity) and tied DMN disintegration specifically to ego dissolution.²

Increased signal-diversity measures — the Lempel-Ziv complexity family — have been the second pillar of the acute-state account. Schartner et al. 2017 (MEG, ketamine + LSD + psilocybin) reported that all three psychoactive doses produced higher spontaneous signal diversity than the normal awake baseline, with the largest effects in single-channel LZ complexity, and showed selective correlations with the intensity of phenomenological reports.³ This was the empirical anchor for the claim that the psychedelic state exceeds, rather than degrades, normal waking consciousness on at least one entropy-like dimension.

A third strand — enhanced repertoire of dynamic brain states — was provided by Tagliazucchi et al. 2014 (psilocybin, fMRI), which found wider variability of BOLD signal in hippocampus and anterior cingulate and a larger repertoire of transient connectivity configurations than placebo;⁴ and by Singleton et al. 2022 (LSD + psilocybin, network control theory), which formalised the same intuition as a “flattening” of the brain’s energy landscape — fewer energetic constraints on transitions between brain states, weighted by 5-HT2A receptor density derived from PET maps.⁵ The 2024 Nature paper by Siegel and colleagues, using precision functional mapping (~18 MRI visits per participant), brought this consensus to its most rigorously sampled form: a single 25 mg psilocybin dose produced more than three-fold larger acute changes in functional connectivity than methylphenidate, driven not by stronger correlations but by desynchronisation across spatial scales — within-network correlations and between-network anticorrelations both attenuated.⁶

The ayahuasca literature is more sparse but convergent. Palhano-Fontes et al. 2015 reported reduced activity and connectivity across DMN hubs (posterior cingulate/precuneus, medial PFC) after oral ayahuasca in ten experienced subjects;⁷ Schenberg’s group has extended this with EEG and serial-imaging work in religious-use populations.

4.2 The default mode network

The DMN — the canonical set of midline and lateral parietal regions identified by Raichle et al. 2001 as more active during task-free rest than during goal-directed cognition⁸ — became, after 2012, the centre of gravity for psychedelic neuroimaging. Carhart-Harris et al. 2012 (intravenous psilocybin, fMRI in 15 healthy volunteers) reported reduced within-DMN resting-state connectivity, decreased blood flow in mPFC and PCC, and increased between-network functional coupling.⁹ The Palhano-Fontes ayahuasca study replicated within-DMN deactivation and connectivity reduction;⁷ Smigielski et al. 2019 extended the picture to psilocybin-assisted mindfulness, showing that decoupling of medial prefrontal and posterior cingulate cortices during the acute session correlated with ego-dissolution ratings and predicted positive psychosocial change at four months.¹⁰

The interpretive frame that grew up around these findings — DMN as the neural substrate of narrative self, its acute disintegration as the substrate of “ego death” — has held up phenomenologically (within-DMN disintegration tracks ego-dissolution scales reasonably consistently across compounds) but should not be confused with a complete mechanistic account. Several caveats are now well-established and should temper any single-network reductionism.

First, the DMN-disintegration signature is not unique to psychedelics. Similar reductions appear under sleep onset, anaesthesia, meditation, and some pathological states. It is the combination of DMN disintegration with hyperexcitability, increased entropy, and broadband desynchronisation that is psychedelic-specific.

Second, Doss et al. 2022 (“Models of psychedelic drug action: modulation of cortical-subcortical circuits”) explicitly de-centred the DMN in their review, proposing alongside the DMN/REBUS account a cortico-claustro-cortical model in which acute network-level desynchronisation is downstream of disrupted claustrum gating of cortical state, and a cortico-striato-thalamo-cortical model linked to the older “thalamic filter” tradition of Vollenweider.¹¹ These are not mutually exclusive with the DMN account, but they emphasise that the DMN’s prominence in the human imaging literature may reflect what is easiest to measure with resting-state fMRI rather than what is causally upstream.

Third, the DMN-as-mechanism story is partly circular: the DMN is identified by its task-rest contrast; ego dissolution is the most consistently reported acute phenomenology; the two correlate, which is the prediction, not the explanation. As Doss and Barrett’s group have argued, ego dissolution and DMN disintegration can dissociate at the individual level, and acute network change does not consistently predict therapeutic response — a point with direct relevance to §4.7 and §4.8.

4.3 REBUS and the entropic brain

The most influential theoretical frame in this literature remains the entropic brain hypothesis (Carhart-Harris et al. 2014, Frontiers in Human Neuroscience),¹² later cast into Bayesian-predictive-processing terms as REBUS — Relaxed Beliefs Under pSychedelics and the Anarchic Brain (Carhart-Harris & Friston 2019, Pharmacological Reviews).¹³

The entropic brain hypothesis takes the cluster of acute findings in §4.1 — increased Lempel-Ziv complexity, decreased low-frequency oscillatory power, increased between-network connectivity, increased repertoire of dynamic states — as evidence that the psychedelic state is a state of elevated entropy in spontaneous brain activity. Carhart-Harris and Friston framed this in 2014 as occupying a “critical zone” between ordered and disordered regimes, and proposed that the informational richness of a conscious state correlates with this entropy index. REBUS then re-described the same picture in active-inference terms: the brain at rest implements a hierarchical generative model in which top-down priors constrain bottom-up sensory and interoceptive input. Under classic psychedelics, the precision (inverse variance) of high-level priors is reduced — beliefs about self, world, and body become less constraining — and bottom-up information is liberated. The therapeutic logic is straightforward: in disorders thought to involve rigid, over-weighted priors (treatment-resistant depression, PTSD, addiction, OCD), a temporary loosening of those priors creates a window in which lower-level evidence can revise them, supported (ideally) by structured psychotherapy.

REBUS has been productive precisely because it offers a single computational vocabulary that spans the molecular (5-HT2A on layer-5 pyramidal cells), the cellular (decreased gain on prediction-error suppression), and the clinical (depressive rumination as over-precise self-referential prior). But it has not gone unchallenged.

The most direct counter-proposal is Safron 2020’s ALBUS (Altered Beliefs Under Psychedelics), which argues that classic psychedelics can both relax and, in some regimes, strengthen priors (the SEBUS effect: Strengthened Beliefs Under Psychedelics) — for example, by inflating the perceived significance of imaginal content, which then becomes a new high-precision prior post-trip.¹⁴ This is not a small caveat: it implies that the same neurochemistry can produce ideologically rigid belief change as readily as flexible revision, which has implications for both clinical safety (psychotic decompensation, integration of trauma-confirmatory beliefs) and the framing of “spiritual” content as therapeutic.

Letheby and Gerrans’ “self-binding” account is a parallel philosophical alternative that retains REBUS-style predictive processing but locates the explanatory action specifically in the disruption of the self-model — the inferred representation of an enduring agentic subject — rather than in priors more generally.¹⁵ On this view, the therapeutic effect is not a generic prior-loosening but a specific revision of self-representations that have become maladaptive (e.g., “I am worthless,” “I am unsafe in my body”). Letheby’s Philosophy of Psychedelics (2021) develops this into a naturalistic account that explicitly rejects the metaphysical readings of ego death while preserving its psychological reality.

Methodological critiques of REBUS overlap with those of the entropic brain more broadly. Entropy as a concept is under-specified across the human imaging literature — Lempel-Ziv complexity of MEG signals, Shannon entropy of BOLD time-series distributions, sample entropy, and several other measures are routinely lumped under the label, with very different statistical properties and very different sample-size requirements. Small-sample neuroimaging combined with multiple analytic degrees of freedom raises a real concern that the entropic-brain literature is at least partly a garden of forking paths. Doss et al. 2022 made this case explicitly, and a critical-review pipeline that emerged in 2023–2025 has begun to disaggregate which entropy measures replicate (LZ-style temporal diversity is robust; many BOLD-entropy measures are not). The empirical core of REBUS is genuinely there; the precision of the theoretical claim has out-run the precision of the data.

4.4 Cortical–subcortical reorganisation

The DMN story sits inside a larger picture of acute reorganisation that includes the thalamus, the claustrum, the amygdala, and the cingulate.

Thalamus. Preller et al. 2018 (eLife) used a placebo-LSD-versus-ketanserin-plus-LSD crossover design and reported that LSD reduced associative thalamic connectivity while increasing sensory/somatomotor thalamic connectivity, with whole-brain effects matching the cortical gene-expression map of 5-HT2A and the entire signature blocked by ketanserin.¹⁶ Preller et al. 2020 (Biological Psychiatry) followed up with a time-resolved oral-psilocybin study that mapped these changes onto both 5-HT2A and 5-HT1A spatial gene-expression patterns and demonstrated that baseline connectivity predicted the magnitude of acute change — an early sketch of a predictive imaging biomarker.¹⁷ The 2022 NeuroImage study by Gaddis and colleagues (Barrett group) refined this to spatially constrained, sub-nucleus-specific alterations of thalamic functional organisation rather than a monolithic “thalamic gate” effect.¹⁸ Together these support a modified, sub-region-specific revision of Vollenweider’s cortico-striato-thalamo-cortical model.

Claustrum. Barrett, Krimmel, Griffiths and Mathur 2020 (NeuroImage) provided the first human evidence for psilocybin modulation of claustral activity, a structure densely expressing 5-HT2A and reciprocally connected to most of cortex.¹⁹ Psilocybin reduced low-frequency BOLD fluctuation amplitude and BOLD variance bilaterally in claustrum, reduced right-claustrum connectivity with auditory and default-mode networks, and increased right-claustrum connectivity with the fronto-parietal control network. This finding underwrites Doss et al.’s cortico-claustro-cortical model and points to a candidate “gain control” node upstream of the more familiar DMN reorganisation.

Amygdala and cingulate. Mueller et al. 2017 (Translational Psychiatry, 100 µg LSD, 20 healthy subjects) showed reduced left-amygdala and right-medial-PFC reactivity to fearful faces under LSD, with amygdala deactivation tracking subjective drug intensity.²⁰ This converged with earlier psilocybin work from the Vollenweider group and grounded the clinical intuition that acute amygdalar dampening might mediate trauma-processing windows in MDMA-AT and psilocybin-AT. The cingulate — particularly anterior cingulate — appears throughout the literature as a site of reduced low-frequency power, reduced within-DMN connectivity, and (in the Doss neurochemistry work) altered glutamate/GABA balance.

Sub-second dynamics. Singleton et al. 2022’s network-control-theory analysis of LSD and psilocybin formalised the intuition of “increased flexibility” as a flattening of the brain’s control-energy landscape — fewer energetic constraints on transitions between brain states, with the flattening explained by 5-HT2A receptor density.⁵ This is the most quantitative version yet of the claim that psychedelics open a temporary regime of cognitive and affective flexibility.

4.5 Neuroplasticity I — the psychoplastogen framework and dendritic-spine growth

The post-2018 plasticity literature has been the single most consequential development for any mechanistic bridge from the acute state to the lasting clinical effect. The foundational paper is Ly, Olson and colleagues 2018, Cell Reports: “Psychedelics Promote Structural and Functional Neural Plasticity.”²¹ Across rat cortical neurons in culture and Drosophila in vivo, classic 5-HT2A psychedelics (LSD, DOI, DMT, psilocin) increased neurite branching, spinogenesis, and synapse number, with structural changes accompanied by enhanced excitatory transmission. The plasticity effects depended on 5-HT2A, TrkB, and mTOR signalling — the same intracellular cascade implicated in the rapid antidepressant action of ketamine, recasting psychedelics as members of a broader “psychoplastogen” class.

The Ly 2018 paper is, however, an in vitro and invertebrate paper. The decisive in vivo mammalian evidence came from Shao, Liao, Kwan and colleagues 2021, Neuron: “Psilocybin induces rapid and persistent growth of dendritic spines in frontal cortex in vivo.”²² Using chronic two-photon imaging of layer-5 pyramidal apical dendrites in mouse medial frontal cortex, a single 1 mg/kg dose of psilocybin produced ~10% increases in dendritic spine size and density, driven by elevated spine formation rather than reduced loss. The structural remodelling appeared within 24 hours and was still present one month later, and was accompanied by amelioration of stress-related behavioural deficits and increased excitatory neurotransmission. Crucially, this combined the temporal profile (rapid onset, weeks-long persistence) of the clinical antidepressant signal with the cellular profile (single-dose, spine-density change) of ketamine plasticity — bridging the chronic-disease-clinical-trial timescale to a cellular event timescale for the first time.

The Shao/Kwan paper invited a second wave of in vivo plasticity work — preserved spine effects after DMT (Ly group), DOI (multiple), and 5-MeO-DMT (Vargas group); evidence of activity-dependent rewiring of large-scale cortical networks following psilocybin (Kwan group, 2025); and the recurring observation that the structural effect is concentrated in prefrontal cortex and is partially recapitulated in the claustrum-frontal projection. Several open questions remain. The mouse work uses doses (1 mg/kg) far above the human therapeutic dose normalised by body mass; the dendritic-spine effect, while persistent, has not been causally linked to the persistence of mood improvement in any mammal (correlational only); and the link between spine remodelling and the functional connectivity changes in §4.4 and §4.7 is still inferential. The plasticity literature is now the single strongest mechanistic argument that psychedelic effects extend beyond the acute trip — but the inferential chain to clinical antidepressant durability is several joints long.

4.6 Neuroplasticity II — the intracellular 5-HT2A thesis

Vargas, Olson and colleagues 2023, Science: “Psychedelics promote neuroplasticity through the activation of intracellular 5-HT2A receptors” — is the paper around which the 2024–2026 plasticity discussion is now organised.²³

The puzzle the paper resolved is plain in retrospect. The endogenous agonist of the 5-HT2A receptor is serotonin (5-HT). If 5-HT2A activation alone were sufficient for plasticity, ordinary serotonergic tone, or pharmacological 5-HT elevation by SSRIs, ought to produce the same effects. It does not — SSRIs, even at high doses with chronic dosing, do not produce the dendritic-spine growth of a single psychedelic dose, and 5-HT itself does not produce neuritogenesis or spinogenesis in the assays used by the Olson group. Vargas et al.’s explanation: 5-HT is a charged molecule and does not readily cross cell membranes. Classic psychedelics (LSD, DMT, psilocin) are lipophilic and freely permeate. The 5-HT2A receptors that mediate the plasticity response are intracellular — present on Golgi/ER membranes inside the cell, accessible to lipophilic ligands but not to extracellular serotonin.

Empirical pillars of the paper. (1) Compounds were ranked by experimentally measured membrane permeability and shown to correlate with their plasticity-inducing potency. (2) Ketanserin (cell-permeant 5-HT2A antagonist) blocked psychedelic-induced plasticity; membrane-impermeable 5-HT2A antagonists did not, at concentrations sufficient to block surface receptors. (3) Engineering serotonin to be membrane-permeable made it plasticity-active. (4) Intracellular 5-HT2A receptors were directly visualised; their conditional knockout in defined neuronal populations selectively eliminated the plasticity effect.

The implications, while owned mechanistically by Chapter III, have systems-level relevance and warrant brief notice here. First, receptor compartmentalisation becomes a legitimate target for medicinal chemistry: the explicit “psychoplastogen” design programme assumes that permeability is as important a design parameter as binding affinity or functional selectivity. Second, the long-standing paradox that 5-HT2A agonism alone is insufficient to define a psychedelic (e.g. lisuride, certain ergolines) gains a candidate explanation: where the agonist acts (surface vs. intracellular pool) matters as much as whether it acts. Third — and this is the bridge to the non-hallucinogenic-psychedelic programme that Chapter III owns — it becomes coherent in principle to design compounds that engage the intracellular pool and induce plasticity without the cortical-state changes that produce subjective intoxication. Whether the Olson group’s 5-MeO-DMT analogues, the Roth group’s IHCH-7086 / IHCH-7113 β-arrestin-biased compounds (Chapter III), or other candidates ultimately decouple plasticity from psychoactivity is the empirical question of the next five years.

A caution. The intracellular thesis is published, replicated within the Olson group, and consistent with the broader pattern of which classic psychedelics induce plasticity in which assays. Independent replication outside Davis-affiliated groups is still partial as of mid-2026, and the inferred mechanism (intracellular 5-HT2A as the operative pool) is not the only possible explanation of the permeability-plasticity correlation — alternatives involving downstream lipophilic-only targets have not been definitively excluded. Treat the thesis as load-bearing, but acknowledge that the bedrock under it is still settling.

4.7 Long-term tract changes and connectivity persistence

A central post-2020 question has been whether the acute reorganisation reviewed in §4.1–4.4 leaves measurable durable traces, and whether those traces correlate with clinical outcome. The literature is still small but increasingly coherent.

Daws, Carhart-Harris and colleagues 2022, Nature Medicine: “Increased global integration in the brain after psilocybin therapy for depression,” pooled fMRI data from an open-label TRD trial and the psilocybin-versus-escitalopram MDD comparison.²⁴ One day after the 25 mg dose, brain network modularity was reduced (i.e., networks were less segregated), and this decrease correlated with antidepressant response at 6 months in TRD and at 6 weeks in the MDD comparison. The escitalopram arm did not produce comparable connectivity change. This was the first reasonably-powered demonstration that an acute connectivity signature persists at least into the early days post-dose and tracks therapeutic response. Doss et al. 2021 (Translational Psychiatry) showed parallel persistence: psilocybin therapy in MDD increased cognitive and neural flexibility for at least four weeks, although flexibility gains did not correlate with antidepressant response (an important null that argues against single-mechanism stories).²⁵

Siegel et al. 2024, Nature, extended the persistence finding to weeks with within-subject precision functional mapping: psilocybin’s acute connectivity changes mostly resolved within days, but a specific reduction in connectivity between the anterior hippocampus and the default mode network persisted across at least three weeks of post-dose imaging.⁶ This durable hippocampal–DMN decoupling — proposed by the authors as a candidate anatomical correlate of the plasticity and clinical effects — is currently the most spatially specific durable signature in the human literature.

The 2026 contribution from the Carhart-Harris group is Lyons, Spriggs, Kerkelä and colleagues 2026, Nature Communications: “Human brain changes after first psilocybin use.”²⁶ In 28 psychedelic-naive healthy participants given a single 25 mg dose, EEG within the first hour showed increased neural entropy that correlated with psychological insight the following day. Diffusion tensor imaging done at baseline and again one month later showed decreased axial diffusivity bilaterally in prefrontal-subcortical tracts — a microstructural change interpreted as increased fibre density/integrity, in the opposite direction of normal age-related decline — and this DTI change correlated with the one-month decrease in network modularity measured by resting-state fMRI. The acute entropy increase predicted both the next-day insight rating and the one-month well-being change. This is the first human evidence to date for coupled functional and white-matter changes at one month post a single psilocybin dose; it is consistent with, but does not by itself prove, the inference that the rodent dendritic-spine plasticity story has a tract-level counterpart in humans. (This citation was a known unresolved item from Phase 1 scoping — the Phase 1 DOI s41467-026-71962-3 resolves correctly to this Lyons et al. 2026 paper, but the first author is Lyons, not Carhart-Harris; Carhart-Harris is corresponding senior author.)

Read together, Daws 2022, Doss 2021, Siegel 2024 and Lyons 2026 establish three points relevant to the clinical bridge. First, the acute reorganisation does leave durable functional and now structural traces, at least at the 3- to 4-week timescale. Second, the connectivity-response correlation is real but not deterministic — global integration / modularity reduction tracks antidepressant response at the group level but is not a clean within-subject biomarker. Third, the persistence is not a maintained acute state but a reset to a different baseline — flatter modularity, weaker hippocampal-DMN coupling, modestly remodelled prefrontal-subcortical tracts.

4.8 Bridging molecule → systems → clinical

A defensible 2026 integrative model runs as follows. Lipophilic 5-HT2A agonists (psilocin, LSD, DMT, mescaline) cross the neuronal membrane and engage both surface and intracellular 5-HT2A pools. Surface 5-HT2A activation on layer-5 cortical pyramidal cells, particularly in association cortex and the claustrum, produces the acute cortical state (§4.1) — increased cellular excitability, broadband oscillatory desynchronisation, increased entropy and dynamic-state repertoire, reduced within-network modularity, increased between-network and global integration, and, at the network level, the characteristic disintegration of the default mode network with concomitant cortico-claustral, cortico-thalamic and cortico-amygdalar reorganisation. Phenomenologically this state is what REBUS describes as “relaxed beliefs” — reduced precision-weighting of high-level priors, including the self-model — though whether it is precisely that, as Letheby/Gerrans, Safron and Doss/Barrett variously dispute, is a still-live theoretical question.

In parallel, intracellular 5-HT2A activation triggers the BDNF/TrkB/mTOR cascade and structural plasticity (§4.5–4.6): dendritic-spine formation in prefrontal pyramidal neurons, beginning within 24 hours and persisting at the cellular level for at least a month in rodents. At the macroscale this appears in humans as decreased network modularity at one day (Daws 2022), persistent reduction in hippocampal-DMN coupling at three weeks (Siegel 2024), and at one month as coupled microstructural (DTI) and functional (modularity) change in prefrontal-subcortical tracts (Lyons 2026). The clinical antidepressant effect, on this model, depends on the combination of the acute window of cognitive-affective flexibility (during which maladaptive priors, including self-representations, can be revisited — ideally with structured psychotherapy) and the post-acute structural-plasticity window (during which revised representations can be consolidated into durable circuit change).

This bridge is firmer than it was in 2016, but several joints remain hand-wavy.

• The causal link from rodent dendritic-spine remodelling to mammalian behavioural antidepressant effect is correlational. The Olson and Kwan groups have demonstrated parallel time-courses; they have not yet selectively abolished one while preserving the other.
• The acute-state-to-clinical-response correlation is real at the group level (Daws 2022; Siegel 2024) but individual-level predictive accuracy is modest. Doss et al. 2021’s flexibility-without-response finding is the clearest counter-example.
• The intracellular 5-HT2A thesis is mechanistically clean in rodent cell culture but its specific translation into the human clinical effect — particularly the relative contribution of intracellular versus surface receptor pools in vivo — is not yet directly testable.
• Whether the non-hallucinogenic 5-HT2A psychoplastogens that Chapter III treats (IHCH-7086/7113; Olson group analogues; tabernanthalog) preserve clinical antidepressant efficacy without the acute REBUS-like state is the field’s central decoupling question. The answer will decide whether the systems-level reorganisation in §4.1–4.4 is mechanistically necessary, sufficient, or merely co-occurrent with the plasticity in §4.5–4.6.

These open questions are taken up in the integrative discussion (Chapter XIII).

References

Additional works referenced contextually

Vollenweider FX, Preller KH. Psychedelic drugs: neurobiology and potential for treatment of psychiatric disorders. Nat Rev Neurosci 2020;21(11):611–624. PMID: 32929261. doi:10.1038/s41583-020-0367-2.

Letheby C. Philosophy of Psychedelics. Oxford University Press; 2021. ISBN 978-0-19-884330-8.

← Ch. III · Overview · Ch. V →