Microevidence for microdosing with psilocybin mushrooms: a double-blind placebo-controlled study of subjective effects, behavior, creativity, perception, cognition, and brain activity

Over the last decade, microdosing—taking sub-hallucinogenic amounts of psychedelics—has become a widely reported practice, with users claiming improvements in mood, cognition, creativity and relief from certain clinical symptoms. However, much of the evidence supporting these claims comes from self-selected, uncontrolled surveys and open-label studies, leaving the field vulnerable to expectancy effects, confirmation bias and variability in dosing schedules and materials. Prior controlled work on low doses of psychedelics (mostly LSD and some psilocybin) has produced mixed findings and has highlighted the importance of maintaining blinding and measuring physiological and neurobiological endpoints in addition to self-report. Cavanna and colleagues set out to test acute effects of a standardised microdose of Psilocybe cubensis in a pragmatic, double-blind, placebo-controlled design that preserved ecological validity. They recruited people already planning to begin a microdosing protocol with their own mushroom material, randomised each participant to one week of active capsules and one week of placebo capsules (order counterbalanced and blinded), and assessed subjective experience, behaviour, creativity, perception, cognition, physical activity and resting-state EEG. The study aimed to determine whether a 0.5 g dried mushroom dose (taken twice in a week) would produce measurable effects beyond placebo while controlling for unblinding and expectation.

Design and participants: The investigators recruited 34 Spanish‑speaking adults between December 2020 and August 2021 who were planning to start a psilocybin mushroom microdosing regimen. Participants reported prior psychedelic experience (mean 11 ± 14.9 experiences) but few had prior microdosing experience (n=6). Exclusion criteria, assessed via a non-diagnostic psychiatric interview, included psychotic or bipolar disorders (including first- and second-degree relatives), recent substance dependence (excluding nicotine), a range of mood and anxiety disorders, eating disorders, neurological disorders and current psychiatric medication. All participants provided informed consent and completed all scheduled visits. Blinding and dosing schedule: The experiment used a randomised double-blind within-subject design. Each participant completed two consecutive weeks: one week in which they consumed two 0.5 g gel capsules of finely ground dried Psilocybe cubensis (Wednesday and Friday) and one week in which they consumed visually identical capsules containing 0.5 g of dried edible mushroom as placebo. Capsule order was randomised by a third party and codes were revealed only after data collection and scoring. Participants visited the experimental premises on dosing days and completed tasks roughly 1.5 hours after ingestion. Baseline trait measures were taken at the start of each condition week. Assessments and tasks: The outcome battery included an acute subjective effects Visual Analogue Scale (VAS), a large set of psychometric instruments (Big Five Inventory, STAI state/trait, PANAS, Perceived Stress Scale, Tellegen Absorption Scale, suggestibility, flow, creative personality, mind-wandering, cognitive flexibility, empathy), creativity tests (Remote Associates Test, Alternative Uses Task, Wallach-Kogan), perception tasks (backward masking, binocular rivalry), cognitive tasks (attentional blink, Go/No-Go, Stroop, Trail Making Test), a Local‑Global auditory ERP paradigm, resting-state EEG (24 channels) with spectral and Lempel‑Ziv complexity analyses, and daily physical activity monitored by a Fitbit Charge 4 (steps and distance). EEG and preprocessing: EEG was recorded with 24 Ag/AgCl electrodes at standard 10-20 locations. Data were bandpass filtered (1–90 Hz) with a notch filter (47.5–52.5 Hz), artefactual channels and epochs were identified by automated criteria and visual inspection, and independent component analysis was used to remove ocular, cardiac and muscle artefacts. Broadband signal complexity was estimated using the Lempel‑Ziv algorithm applied to median-split binary sequences of the instantaneous signal envelope. Chemical characterisation: For each mushroom source, 150 mg samples were analysed by LC‑MS to quantify psilocybin, psilocin, baeocystin and norbaeocystin. Reported concentrations were psilocybin 640.2 μg/g, psilocin 950.7 μg/g, baeocystin 50.4 μg/g and norbaeocystin 12.5 μg/g. Based on these measurements, a single 0.5 g capsule was estimated to contain 0.32 mg psilocybin and 0.48 mg psilocin; the authors note the analysis occurred at least one month after the experiment and that degradation during storage could mean the reported potencies are underestimates. Three independent mushroom sources were used across participants. Statistical approach: Paired comparisons between active and placebo conditions were performed using non-parametric Wilcoxon signed-rank tests. Unless otherwise stated, p‑values were Bonferroni-corrected for multiple comparisons. The authors report both corrected and uncorrected outcomes and did not conduct an a priori power calculation because expected effect sizes were unknown.

Blinding and acute subjective effects: Participants correctly identified active versus placebo weeks in 49 of 64 measurement weeks (75%); the authors treated data as "unblinded" versus "blinded" subsets in secondary analyses. Summed VAS scores (indexing overall acute subjective intensity) were significantly higher for the active dose than for placebo on dosing days (Wednesday and Friday), but not on non-dosing days, indicating no clear carry-over. When analyses were restricted to instances in which participants remained blinded, the VAS total difference was absent. Individual VAS items showed some nominal differences (imagination, dreamlike quality, spatial distortions, mind‑wandering) but none survived Bonferroni correction, and the pattern of items differed across subjects rather than showing a consistent profile. Self-report questionnaires and psychometrics: Across the battery of trait and state questionnaires (normalized for comparison), no statistically significant differences were observed between active and placebo conditions at p<0.05 uncorrected. Trait measures were assessed at baseline only. Creativity measures: The convergent thinking RAT and the divergent thinking WK and AUT showed no significant differences between active and placebo weeks (p<0.05 uncorrected). Multiple component scores (fluency, originality, elaboration, repetitions) were analysed and none produced significant effects; results were similarly null when restricted to the unblinded subset. Perception tasks: Backward masking (objective and subjective visibility across stimulus-onset asynchronies) revealed the expected sigmoid subjective curve but no significant condition differences. Binocular rivalry alternation rates and mean dominance times averaged about 4 s for both conditions, with no statistical difference. Cognitive tasks: Across attentional blink, Go/No-Go, Stroop and Trail Making tests, few nominal differences emerged but none survived correction. Specifically, psilocybin reduced T2 visibility at one short lag in the attentional blink (did not survive correction), and increased reaction time for incongruent Stroop trials (also did not survive correction). Go/No-Go accuracy and reaction time, and Trail Making time/errors, showed no significant differences. Physical activity: Day-by-day step counts, distance travelled, resting and activity time from Fitbit data did not differ between active and placebo weeks. A midweek dip in activity corresponded with the measurement sessions requiring seated tasks. EEG and ERPs: Resting-state EEG showed reduced spectral power in the theta band under the psilocybin microdose relative to placebo. Lempel‑Ziv complexity measures did not differ between conditions. ERPs from the Local‑Global auditory paradigm (including P300 components for local and global deviants) did not show significant between-condition differences reported in the extracted text. Behavioural results from perception and cognitive EEG tasks were null and therefore not pursued further in EEG analyses. Analyses by blinding status: Repeating main comparisons within the "unblinded" and "blinded" data subsets did not reveal additional statistically significant findings beyond the acute VAS effects in the unblinded subset. Overall, the majority of measured domains showed null effects or trends toward impaired cognitive performance rather than enhancement.

Cavanna and colleagues interpret their findings as evidence that a standardised 0.5 g dried Psilocybe cubensis microdose (two such doses in a week) produced reliable acute subjective sensations but did not produce consistent improvements across a broad set of behavioural, perceptual, cognitive, creativity, mood or activity metrics. The significant increase in VAS total scores on dosing days was concentrated among participants who correctly identified the active condition, which the authors suggest is consistent with expectancy or unblinding effects driving at least part of reported subjective benefits in naturalistic microdosing. In terms of neurophysiology, the observed reduction in theta-band power under the microdose aligns with prior reports of broadband oscillatory power reductions at higher serotonergic psychedelic doses, suggesting that low doses may begin to elicit similar spectral changes. Conversely, the absence of a change in Lempel‑Ziv complexity leads the investigators to propose that increased signal entropy may be more characteristic of the altered conscious states produced by higher, overtly psychedelic doses rather than microdoses. The authors situate their null and modestly negative cognitive findings within recent controlled microdosing literature, noting comparable studies of low-dose LSD that reported limited or mixed effects on mood and cognition. They emphasise that many earlier positive reports derive from uncontrolled, open‑label designs vulnerable to expectation and positivity bias among self-selected users. The unblinding rate they observed (75%) is similar to that reported in other microdosing trials and may explain why subjective reports sometimes outpace objective measures. Several limitations are acknowledged. The chemical composition of mushroom material varies and the authors pooled samples from three sources; LC‑MS quantification (psilocybin 640.2 μg/g, psilocin 950.7 μg/g) implied a modest alkaloid load per capsule (0.32 mg psilocybin and 0.48 mg psilocin per 0.5 g), but the assay was performed at least one month after the study and degradation could lead to underestimation of potency. The design focused on acute effects and did not assess cumulative outcomes from repeated long-term microdosing; tolerance and schedule-dependent effects remain untested. Dose standardisation by body weight was not performed and may matter more at low doses. Finally, the study did not conduct an a priori power calculation because expected effects were unknown, which the authors note when interpreting null results. Implications and future directions articulated by the authors include the need for further rigorous, placebo-controlled work to disentangle expectancy from pharmacological effects, investigations of longer-term dosing schedules and cumulative effects, and assessment of physiological endpoints relevant to safety (for example, potential consequences of chronic 5-HT2A receptor stimulation on circulatory health). They conclude that, based on their data, it is premature to assert that microdosing reliably delivers the positive outcomes often claimed, and that safety and efficacy remain unresolved until more comprehensive research is completed.