Is microdosing a placebo? A rapid review of low-dose LSD and psilocybin research

Microdosing—the regular ingestion of sub-hallucinogenic amounts of classic serotonergic psychedelics such as LSD or psilocybin—became widely discussed from about 2015, fuelled by anecdote and media reports claiming benefits for mood, cognition and wellbeing. Early academic work was dominated by surveys, qualitative interviews and observational prospective studies; more rigorous controlled research only began to appear from 2018. Importantly, some early controlled designs that attempted to blind participants, including a notable self-blinding prospective study, found little difference between placebo and active microdoses and suggested that participants' beliefs about what they had taken strongly influenced outcomes. In response to this mixed picture, Polito and colleagues carried out a rapid review that focused specifically on microdosing studies in which the quantity of psychedelic administered was known and a placebo comparison was included. The stated aim was to synthesise the controlled-dose literature in order to evaluate the extent to which observed effects might reflect pharmacology versus placebo or expectancy effects, and to identify methodological gaps that limit causal inference in this field.

The authors performed a rapid review searching for peer-reviewed human studies, published after 2018, that investigated controlled doses within a microdose range of classical serotonergic psychedelics and included a placebo comparator. They excluded papers where a microdose was only used as a comparator for higher doses and naturalistic self-report studies without known dosing. Search terms targeted psychedelic substances together with indicators of low-dose administration in titles or abstracts. Each author screened results independently and resolved disagreements by consensus. Inclusion criteria were: use of classic/serotonergic psychedelics, controlled microdose-range administration, presence of a placebo condition, reporting of primary empirical data, human subjects, and peer-reviewed publication. The initial search returned 131 items; after de-duplication 127 titles and abstracts were screened, 29 full texts were reviewed, and 19 papers met the inclusion criteria. Risk of bias was assessed using the revised Cochrane risk-of-bias tool for randomised trials. The review restricted attention to studies reporting mental health or cognitive enhancement outcomes and focused only on studies with an explicit placebo comparison.

Nineteen papers that met the inclusion criteria were synthesised, though these derived from only 10 independent dose-controlled experiments conducted by six different laboratories. Thirteen of the 19 papers (68%) reported trial registrations, but many registrations lacked pre-specified analysis plans and the authors note that 14/19 studies (74%) did not provide pre-registered analyses. Overall risk-of-bias assessments indicated low risk for randomisation, missing data and outcome measurement, but blinding was frequently imperfect. Neurobiological outcomes: Six neuroimaging or electrophysiological studies reported consistent neural changes associated with microdosing. Findings included altered effective connectivity in visual cortex from dynamic causal modelling, changes in amygdala–cerebellum connectivity in fMRI, reduced default mode network resting activity and increased neural complexity in EEG, reduced resting-state EEG power after psilocybin microdoses, altered error rates in visual oddball tasks, and increased ERP responses during reward processing under LSD. Physiology and biomarkers: Microdosing studies reported increased pain tolerance and higher brain-derived neurotrophic factor (BDNF) levels in some LSD studies. Measures of sleep quality, app-based physical activity and balance showed no effects. Cardiovascular findings were mixed: several LSD studies found no blood pressure or heart rate changes, while three reported increases. No long-term safety data were available and the authors highlight theoretical concerns about chronic activation of 5-HT2B receptors and cardiac risk. Phenomenology and subjective state: Across both LSD and psilocybin studies, visual analogue scale ratings reliably showed increased acute subjective effects on items such as feeling "under the influence", "good drug effects", subjective intensity, happiness and productivity. Standardised altered-states instruments (e.g. 5D-ASC) yielded inconsistent results; some subscales (oceanic boundlessness, insightfulness, visionary restructuralisation) showed changes in some studies but not others. Addiction Research Centre Inventory subscales were higher under microdosing in different studies, but without consistent patterns across studies. Affective outcomes: LSD microdosing consistently increased acute vigour; other mood dimensions showed inconsistent results across studies. Some trials reported immediate or 48-hour mood improvements in participants with higher baseline depressive symptoms, whereas most trials of healthy volunteers found no persisting mood benefits over weeks. Psilocybin microdoses increased perceived awe but did not change aesthetic experience or standard affect measures such as the Positive and Negative Affect Scale. Cognition and behaviour: Results were mixed. There were suggestions that LSD microdosing can alter time perception and reduce attentional lapses, and psilocybin microdosing was associated with increased verbosity and sentiment in language production. However, many studies failed to detect effects on standard cognitive batteries, creativity tasks, suggestibility or self-representation, and social-cognitive findings were inconsistent. Mental health endpoints: Only four studies assessed wellbeing or clinical symptom measures; one reported reductions in depression among participants with elevated baseline depressive symptoms after an LSD microdose, while the others found no differences in depression, anxiety, stress or wellbeing between microdose and placebo conditions. All clinical-like assessments were carried out in non-clinical, largely healthy samples. Blinding and expectancy: Twelve studies evaluated blinding and 11 of those reported participants guessed condition at rates greater than chance, indicating frequent breaking of naïve blinding. Some analyses showed that participants' beliefs about treatment influenced outcomes and in at least one study guesses had stronger associations with outcomes than actual condition. Nonetheless, when quantitatively assessed in other studies, expectancy explained only a modest proportion of outcome variance (for example 5%–8% in one observational study).

The authors interpret the controlled-dose literature as showing a mixture of signal and null findings. They highlight relatively consistent evidence for neurobiological effects, acute changes in conscious state, increased vigour and greater pain tolerance as among the more robust findings. By contrast, outcomes commonly touted in anecdotes and media—improvements in creativity and broad cognitive enhancement—are not well supported by the controlled studies reviewed. Methodological issues complicate causal inference. Small numbers of independent experiments, small sample sizes (average microdose-group N ≈ 31), heterogeneity of substances, doses and measures, and limited duration of exposure reduce sensitivity to detect subtle pharmacological effects. The authors note dose-dependent effects in several LSD studies, which argue for a pharmacological contribution, but they caution that many psilocybin studies may have used doses too low to reach meaningful plasma concentrations. All controlled studies reviewed used non-clinical participants and most recruited individuals with prior psychedelic experience, introducing selection bias and elevated expectations that could both influence results. Blinding failure is emphasised as a central concern: no reviewed study used an active placebo, and several investigations show that participants commonly identified their condition. While some studies report that expectations or guesses exerted substantial influence on outcomes, the authors conclude that the evidence for expectancy as the primary mechanism is mixed and in many analyses accounted for only a small fraction of variance. They therefore reject a definitive placebo-only interpretation on current evidence. In sum, Polito and colleagues argue that the field remains nascent and that current data are insufficient to conclude that microdosing effects are entirely placebo driven. They call for well-powered, longitudinal trials that include clinical populations, systematic dose–response work, longer-term dosing paradigms reflective of naturalistic use, better blinding strategies (for example active placebos), and stronger pre-registration and open science practices to settle whether observed effects are pharmacological, expectancy-driven, or some combination of both.

So, is microdosing a placebo? The authors conclude that there is not yet strong evidence for a purely placebo interpretation. Given the limited number of low-powered studies, selection biases, concerns about inadequate doses in some trials, reliance on non-clinical samples, and frequent blinding failures, definitive attribution of effects to drug versus expectation is premature. The review recommends further well-powered, longitudinal, and pre-registered research across healthy and clinical populations to determine the mechanisms and potential clinical utility of microdosing.