Ch. IX — Cognitive enhancement and microdosing

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

Abstract. The placebo-controlled evidence base for psychedelic microdosing — Szigeti et al. 2021, Cavanna et al. 2022, de Wit/Molla 2022, and the Polito & Liknaitzky 2024 synthesis — has largely failed to dissociate active microdose from placebo on mood, cognition, or creativity, while showing measurable neurophysiological changes (EEG theta, amygdala connectivity, time perception) that do not translate to behavioural benefit. This is best read not as “microdosing does nothing” but as expectancy-dominated: the active-versus-placebo gap is small, the placebo-versus-no-treatment gap is substantial, and expectancy is itself a legitimate psychological mechanism that the literature has only recently begun to characterise rigorously. The chronic-dosing 5-HT2B/valvulopathy question remains the principal open safety concern.

9.1 The popular case and definitional terrain

The contemporary microdosing phenomenon traces to James Fadiman’s The Psychedelic Explorer’s Guide (2011), which popularised a four-day protocol (Day 1 dose, Days 2–4 off) at sub-perceptual doses — typically defined as one-tenth to one-twentieth of a recreational dose. The operational ranges that the literature has converged on are LSD 5–20 µg (versus ~100 µg recreational) and dried Psilocybe cubensis 0.1–0.3 g (versus 2–3.5 g recreational); 1P-LSD and 1cP-LSD analogues are dosed equimolar to LSD on the assumption of complete in-vivo conversion. The four-day schedule is a folk-pharmacological convention with no empirical optimisation; alternative regimens (every-third-day, twice-weekly, weekday-only) circulate without head-to-head comparison.

The early observational literature — Anderson et al. 2019 (n=909 microdosers versus controls, online survey),¹ Polito & Stevenson 2019 (n=98 daily diary, n=263 expectancy survey),² Lea et al. 2020 (Reddit-recruited prevalence and motivation study),³ Hutten et al. 2019 — established the demographic pattern (predominantly 25–45-year-old educated knowledge workers), the modal motives (depression and anxiety self-treatment, cognitive enhancement, creativity), and the modal substances (LSD/1P-LSD and psilocybin in roughly equal proportion). Self-reported benefits were broad and consistent: mood elevation, focus, energy, creativity, social connectedness. The early reception of these surveys in the popular press treated the self-reports as efficacy signals; the surveys themselves were considerably more cautious, noting the absence of any placebo control and the strong demand-characteristic context in which respondents were recruited. Polito & Stevenson explicitly compared participants’ expectations to actual reported effects and found a substantial gap in both directions — useful evidence that the field would need controlled designs to disentangle expectancy from pharmacology.²

9.2 Observational and open-label studies

The post-2018 wave of observational microdosing research consolidated three findings. First, microdosers as a self-selected population score modestly better than non-microdosers on mental-health questionnaires (lower depression, lower anxiety, lower dysfunctional attitudes; higher wisdom, open-mindedness, and unusual-uses creativity)¹ — a pattern that is causally ambiguous because the group selects into the practice on those traits. Second, daily within-subject diary measures show small positive effects on dosing days that fail to persist on the following days² — a pattern consistent with either a transient pharmacological effect or a daily expectancy boost. Third, motives and perceived benefits cluster around depression/anxiety self-treatment and cognitive performance more than recreational experience.³

These studies share methodological limitations that the panel should weight against any inference: anonymous online recruitment selects strongly for engaged users with positive prior experience; there is no placebo group; substance identity, purity, and dose are self-reported and unverified; outcome measures are self-report; and effect-size estimates are inflated by demand characteristics that scale with the prominence of the practice in the community discourse. Open-label microdosing trials (e.g., the LSDDEP1 open-label series from the Auckland group) reproduce the self-report pattern under more controlled supervision but do not address blinding. The observational literature was sufficient to motivate placebo-controlled work; it was never sufficient to establish efficacy.

9.3 Acute laboratory studies of single microdoses

Single-dose laboratory studies of LSD microdoses constitute the cleanest pharmacological evidence in the field, because they sidestep the question of cumulative chronic effects and isolate the immediate dose-response signal under blinded conditions.

Yanakieva et al. 2019⁴ administered placebo and 5, 10, and 20 µg oral LSD to older adults (mean age ~63) in a within-subject, double-blind, placebo-controlled crossover and measured temporal-reproduction performance over sub-second and supra-second intervals. The 10 µg dose reliably produced over-reproduction of supra-second intervals — a small but robust time-dilation effect — in the absence of any subjective-effects-questionnaire change. The interpretation is important: a microgram-range LSD dose engaged a measurable cognitive timing process at a dose participants could not distinguish from placebo subjectively. This is the kind of dissociation that motivates the rest of the literature: pharmacology without phenomenology, and (as later studies confirm) phenomenology without behavioural benefit.

Bershad et al. 2019⁵ administered 0, 6.5, 13, and 26 µg LSD to healthy young adults (n=20, within-subject crossover) at the University of Chicago Human Behavioral Pharmacology Laboratory. LSD produced orderly dose-related subjective effects at 13 and 26 µg, with the 26 µg dose increasing vigor ratings and slightly reducing positivity ratings of positive-valence images; other mood, cognition, and physiological measures were unaffected at any dose. The study established that 13 µg is a threshold dose at which subjective effects become reliably detectable, supporting subsequent repeated-dose protocols. Bershad et al. 2020⁶ reported the fMRI companion at 13 µg, showing altered amygdala resting-state connectivity to angular gyrus, middle frontal gyrus, and cerebellum, with positive-mood effects correlating with the amygdala–MFG connectivity change. Again: pharmacology at the network level, weak and inconsistent at the behavioural level.

Family et al. 2020⁷ reported the first registered modern Phase 1 trial of low-dose LSD in healthy older adults (mean age ~63), with 5, 10, and 20 µg administered every fourth day for six total doses, the only adverse-event signal of note being mild-to-moderate headache more common on active than placebo. The trial established a safety and pharmacokinetic envelope for repeated microdosing in older adults without claiming efficacy.

de Wit, Molla et al. 2022⁸ extended Bershad’s acute work to repeated low-dose LSD (13 or 26 µg every 3–4 days for four doses) versus placebo in healthy adults. The investigators’ summary was characteristically blunt: no dramatic improvements in mood or cognition, and no lasting changes on any measure examined. This is the closest thing in the literature to a clean negative result on the cognitive-enhancement hypothesis for repeated LSD microdosing in healthy volunteers, with the important caveat that “healthy adults” is not the population that drives observational benefit reports — which are dominated by users self-treating mild-to-moderate depression and anxiety.

9.4 Self-blinded citizen science: Szigeti 2021

Szigeti et al. 2021⁹ solved one part of the microdosing literature’s central design problem with an unusually creative methodology. The investigators developed an online protocol in which microdosers — already in possession of their own substance — prepared their own opaque-capsule active and placebo doses, randomised the capsule order with a barcode scheme, and self-administered without knowing which capsule they had taken on any given day. Outcome measures were daily online self-reports. The design preserved real-world heterogeneity of substance, dose, and intent while imposing genuine within-subject blinding; n=191 completed the four-week protocol, making it the largest placebo-controlled microdosing study to date and a small fraction of the cost of a conventional trial.

The headline finding was that active microdose and placebo produced statistically indistinguishable improvements on the principal psychological outcomes: well-being, life satisfaction, mindfulness, and self-rated paranoia all improved against baseline in both arms, with no significant active-versus-placebo difference. A post-hoc subset analysis examined participants’ guesses about which capsule they had received: participants who believed they had taken active microdose reported greater benefits, regardless of what they had actually taken. The placebo-versus-no-treatment delta was substantial; the active-versus-placebo delta was approximately zero.

Power calculation. Szigeti 2021’s n=191 active-vs-placebo within-subject design had approximately 80% power to detect a between-condition effect size of Cohen’s d ≈ 0.20 (a small effect) on the primary self-report outcome at α=0.05 two-sided. The observed point estimate was on the order of d=0.05 (essentially null), with 95% CI plausibly spanning roughly [−0.10, +0.20] on the well-being primary. This is a robust null against moderate effect sizes (d > 0.30); it does not exclude small effects (d < 0.20). Cavanna 2022 (n=34 within-subject, 4-week) had approximately 80% power to detect d ≈ 0.50; the observed effects were below threshold on behavioural outcomes. de Wit/Molla 2022 (n ≈ 20, within-subject, four-dose protocol) had approximately 80% power to detect d ≈ 0.65; the result was null on cognitive and mood primaries. The cross-study null is therefore robust against moderate-to-large effects on the outcomes most central to user reports; the literature does not exclude small effects (d < 0.20) on those outcomes, nor does it exclude clinical-population effects in mild-to-moderate depression/ADHD samples that the controlled corpus has not yet adequately covered (the Polito & Liknaitzky 2024 caveat). Polito & Liknaitzky’s caution about underpowered controlled studies applies most to Cavanna and de Wit/Molla, less to Szigeti 2021.

The strengths of the design were real: true within-subject blinding under field conditions, large sample by microdosing-trial standards, and ecological validity that conventional Phase 2 designs cannot achieve. The limitations are also real and have been discussed in the literature.¹⁰ Substance and dose were heterogeneous and self-reported; outcome measures were self-report only; the blinding-success rate, while better than most psychedelic-trial blinding, was imperfect (correct guessing was above chance on dosing days). The study’s principal contribution was not a definitive null but the demonstration that a citizen-science blinding protocol is feasible — and that, under such a protocol, the active-versus-placebo dissociation that the popular case depends on largely disappears.

9.5 Cavanna 2022 EEG findings

The Cavanna et al. 2022 Translational Psychiatry study¹¹ is the most informative single result in the controlled microdosing literature for the panel’s purposes, because it shows simultaneously that low-dose psilocybin has a real, measurable neural signature and that this signature does not translate to behavioural benefit.

The design was a within-subject double-blind placebo-controlled crossover in 34 individuals already starting to microdose with P. cubensis (0.5 g dried versus visually identical placebo), with outcome batteries on subjective experience, mood, divergent and convergent creativity, perception, cognition, and EEG. The pharmacological signal was clear: psilocybin reduced EEG power in the theta band and preserved Lempel-Ziv signal complexity — a profile consistent with sub-threshold engagement of the same neural state-space that full-dose psilocybin moves through. Subjective effects on the active arm exceeded placebo, but only among participants who correctly guessed which condition they were in — i.e., expectancy and unblinding co-vary, and the subjective effect collapses when you condition on participants who did not break blind.

The behavioural findings were null. There was no significant benefit of active microdose over placebo on creativity (divergent or convergent), cognition, or self-reported well-being. The authors concluded that expectancy underlies at least some of the anecdotal benefits attributed to psilocybin microdosing. For the panel: this dissociation — real neural pharmacology + no behavioural translation — is the key empirical fact that distinguishes the modern microdosing literature from credulous popular accounts. The drug is doing something to the brain; that something is not visible in cognitive or affective performance under controlled conditions.

A 2025 follow-on from the same group and from the Auckland MDLSD trial¹² reproduces this pattern at the cognitive level: repeated 10 µg LSD over six weeks in healthy adult males produced transient on-dose-day improvements in self-rated creativity, connectedness, energy, happiness, irritability, and wellness, but not sustained changes in overall mood or cognition. The on-dose-day effects survived statistical adjustment for pre-intervention expectancy — a methodological refinement that the panel should note is genuinely informative, though it does not overturn the broader pattern.

9.6 Polito & Liknaitzky 2024 review

The Polito & Liknaitzky 2024 synthesis¹³ is the field’s current best summary. The paper is a rapid systematic review of 19 placebo-controlled microdosing studies of LSD and psilocybin and the explicit question it asks is the one the panel needs answered: is the microdosing phenomenon a placebo effect?

Citation correction for the panel: the Phase 1 scoping that fed into this commission returned PMID 11311906 as the Polito & Liknaitzky reference. PMID 11311906 is a 2001 paper on an unrelated topic. The actual reference is PMID 38877715 (DOI 10.1177/02698811241254831), J Psychopharmacol 2024;38(8):701–711. The 11311906 confusion comes from the paper’s PMCID (PMC11311906) being mistaken for a PMID — a recurrent failure mode of LLM-mediated literature retrieval and one the field’s automated tooling needs to learn to handle.

The review’s conclusion is more nuanced than either “microdosing works” or “microdosing is placebo.” Polito & Liknaitzky catalogue eight methodological issues that should restrain a confident “placebo” verdict: small sample sizes in most controlled studies, limited dose-range coverage, heterogeneous substances and protocols, possible selection bias toward null-tolerant participants, outcome measures that may be insensitive to the most subjectively relevant changes, weak blinding integrity in some trials, short trial durations, and limited assessment of population subgroups who report the strongest naturalistic benefits (mild-to-moderate depression, ADHD). Their verdict is that available controlled data do not establish efficacy beyond placebo, and do not warrant a confident “purely placebo” verdict either — particularly given the consistency of mechanistic signals (Cavanna’s EEG theta, Bershad’s amygdala connectivity, Yanakieva’s time-dilation, the BDNF-rise signal from Hutten and Holze’s group).

The framing the panel should adopt — and the framing this chapter endorses — is the one Polito has articulated repeatedly: expectancy is not nothing. Expectancy effects are a legitimate, well-characterised psychological mechanism that mediates a substantial fraction of all psychiatric-treatment response (including antidepressant medication response). A treatment that works through expectancy is not a sham treatment; it is a treatment whose mechanism is psychological rather than pharmacological. The implication for research design is that microdosing trials need to be designed around expectancy as a variable to be characterised, not a confounder to be eliminated. The implication for clinical translation is more sceptical: regulatory pathways are built around dissociating pharmacological effect from placebo effect, and on that test the controlled microdosing literature does not currently support a regulatory case.

9.7 Cardiac safety: chronic 5-HT2B agonism

The principal unresolved safety question for chronic microdosing protocols is valvular heart disease via 5-HT2B receptor agonism. The mechanism is well-characterised from the fenfluramine/dexfenfluramine experience (Fen-Phen was withdrawn in 1997 after valvulopathy in long-term users) and from the dopamine-agonist literature (pergolide and cabergoline withdrawn or restricted in Parkinson’s disease for the same reason); chronic agonism of 5-HT2B receptors on cardiac valve fibroblasts drives proliferation, valve thickening, and regurgitation.

Tagen et al. 2023¹⁴ examined LSD, psilocybin, mescaline, DMT, and MDMA for valvulopathy risk under chronic-microdosing scenarios. Their in-silico exposure modelling concluded that LSD, psilocybin/psilocin, and several active metabolites bind 5-HT2B with potency comparable to or greater than their 5-HT2A engagement, and that maximum plasma concentrations achievable under typical microdosing schedules exceed the receptor-engagement thresholds for known valvulopathogens for some compounds and some protocols. The Tagen analysis is predictive, not observational — no human valvulopathy cohort has yet been studied in chronic microdosers — and the safety margins it derives are wide for some compounds (DMT, mescaline) and narrow for others (LSD, psilocin) when exposure is extrapolated to multi-year daily-equivalent dosing.

For Chapter III’s receptor-biology context: 5-HT2B engagement is intrinsic to the ergoline scaffold and is one of the targets that “differentiated LSD” patent programs (FP-LAD, FLUORETH-LAD; Chapter VIII) attempt to engineer away from. The clinical implication for the panel is that chronic microdosing is the dosing regimen for which the cardiotoxicity question is most live, and the question is at present unanswered by any prospective human data. The natural-experiment denominator — millions of self-reported chronic microdosers globally over the past decade — has not been studied with echocardiography. This is a tractable, unfunded research question that the field needs to resolve before clinical microdosing programs can responsibly extend beyond short-duration protocols.

9.8 Cognitive enhancement and creativity

The cognitive-enhancement and creativity claims have the longest history and the weakest empirical support. Harman, McKim, Mogar, Fadiman & Stolaroff 1966 conducted a small uncontrolled study (n=27 professional men, mescaline 200 mg or LSD 100 µg) in which participants worked on real-world technical problems they had brought to the session.¹⁵ The investigators reported facilitation of insight and persisting weeks-later productivity gains. The study was unblinded, uncontrolled, and used very high doses by contemporary microdosing standards — it is not a microdosing study at all, and its inclusion in the cognitive-enhancement lineage is more cultural than methodological.

Modern controlled work distinguishes divergent thinking (generating many candidate uses for an object — the alternative-uses task) from convergent thinking (identifying the single solution that connects three seemingly unrelated words — the remote-associates test). Some acute full-dose psilocybin and LSD studies report increases in divergent thinking on dose day with no effect on convergent thinking. The microdosing-specific picture is null. Cavanna et al. 2022 found no creativity benefit in either domain.¹¹ de Wit/Molla 2022 found no cognitive benefit in repeated LSD.⁸ The Anderson 2019 observational unusual-uses-task advantage¹ disappears under controlled conditions. The Murphy/Sumner 2025 6-week MDLSD trial reported on-dose-day self-rated creativity gains but not objective performance gains.¹²

The panel should note one structural reason for this null pattern: cognitive-performance gains are precisely the kind of outcome on which expectancy and demand characteristics exert the largest leverage. Self-report creativity scales are particularly vulnerable. Performance-based creativity tasks are less so but also less reliable as outcome measures. No rigorous, double-blind, performance-based demonstration of microdose-induced creativity enhancement exists as of the freeze date, despite the centrality of this claim to the popular case.

The cognitive-enhancement-as-nootropic framing — microdosing positioned alongside modafinil, methylphenidate, or stimulant off-label use — does not survive the controlled-trial data. There is no replicable microdose-induced gain on attention (Stroop, n-back), working memory, processing speed, or executive function in the controlled literature; some studies show small impairments on speeded tasks at 20 µg LSD or above. Microdosing as nootropic is a claim the controlled evidence does not support.

9.9 Where the evidence sits in 2026

The honest summary, suitable for an expert panel, has four parts:

One, the active-versus-placebo dissociation has largely failed across controlled studies of LSD and psilocybin microdosing on the outcomes that matter to users (mood, cognition, creativity, well-being). The clearest single-study demonstrations are Szigeti 2021 (self-blinded n=191, broad null) and Cavanna 2022 (in-lab placebo-controlled, real EEG effects + no behavioural benefit), and the Polito & Liknaitzky 2024 synthesis captures the field-wide pattern.

Two, neurophysiological signals are real: theta-band EEG changes, amygdala connectivity shifts, time-perception dilation, BDNF rises. These signals indicate that sub-perceptual doses engage 5-HT2A and adjacent receptor pharmacology meaningfully. The open mechanistic question is why this engagement does not translate to behavioural benefit — whether the dose is genuinely sub-therapeutic, whether the cognitive measures are insensitive to the relevant changes, or whether the behavioural effects are masked by floor effects in healthy populations and would emerge in clinical samples.

Three, the expectancy-dominated framing is the correct epistemic posture. Microdosing-as-placebo is not microdosing-as-sham; expectancy is a real psychological mechanism that produces real benefits in real users. The clinical-translation implication is sceptical (regulatory pathways disfavour placebo-mediated mechanisms), but the harm-reduction implication is permissive (millions of users self-reporting benefit are not all hallucinating their benefit — they are experiencing a placebo response that may genuinely improve their lives at low pharmacological cost).

Four, chronic-dosing safety, particularly 5-HT2B valvulopathy, is the principal unresolved risk. The literature does not yet have echocardiographic data on long-duration microdosers; the in-silico safety margins for LSD and psilocin under chronic protocols are not large; and the historical analogues (fenfluramine, pergolide) warrant prospective surveillance before any clinical microdosing program scales.

The panel should resist both the credulous framing (“microdosing has demonstrated benefits”) and the dismissive framing (“microdosing is placebo, therefore nothing”). The evidence supports a third reading: microdosing engages real pharmacology, produces real subjective effects via real expectancy mechanisms, has small-to-null active-versus-placebo dissociation under controlled conditions, and carries an unresolved chronic-exposure safety question. That is the verdict the controlled literature supports.

References

← Ch. VIII · Overview · Ch. X →