Psychedelic Microdosing, Mindfulness, and Anxiety: A Cross-Sectional Mediation Study

Microdosing denotes the routine ingestion of sub-hallucinogenic amounts of classical psychedelics such as LSD or psilocybin, typically around one-tenth of a full dose and taken several times per week. Popular reports claim benefits for productivity, mood and anxiety, but empirical findings have been mixed: some observational studies report improved mental health among microdosers while others report elevated anxiety. By contrast, clinical research on moderate to large (“macrodose”) psychedelic sessions has produced promising results for anxiety and treatment-resistant depression and has also linked psychedelic use to changes in personality and increases in mindfulness. Neurobiologically and phenomenologically, psychedelic states and mindfulness meditation share overlaps (for example, effects on the default mode network and self-referential processing), motivating investigation of whether microdosing might relate to trait mindfulness and anxiety. Hartong and colleagues set out to test three hypotheses in a cross-sectional sample: (1) current microdosers report lower trait anxiety than microdosing‑naïve controls, (2) former microdosers also report lower trait anxiety than controls, and (3) trait mindfulness mediates any associations between microdosing (current or former) and trait anxiety. The study therefore combined self-report measures of trait anxiety and mindfulness with questions about microdosing and prior macrodose experience to explore associations and potential mediation effects.

Participants were recruited in October–November 2020 via Reddit posts and a university research advertisement; inclusion required being aged 16 or older and either having microdosing experience or being microdosing‑naïve. Those with microdosing experience were eligible if they were currently following or had previously followed a microdosing regimen using psilocybin, LSD, DMT or mescaline; typical regimen examples included "once every three days" or "four days on, three days off." Recruitment yielded 741 survey starts, 590 complete responses and, after excluding 93 participants who microdosed only irregularly, a final analytic sample of 497 participants. Of these, 263 (52.9%) reported some microdosing experience. Student participants recruited through the university received course credit; Reddit recruits were not remunerated. Ethical approval was obtained from the University of Amsterdam faculty review board. After informed consent, participants completed items on demographics, microdosing history (current/previous/none and substances used), previous moderate-to-large dose (macrodose) experience, and current mindfulness meditation practice. Trait anxiety was assessed with the State‑Trait Anxiety Inventory—Trait subscale (STAI‑T), a 20‑item measure scored 20–80 (higher = greater proneness to anxiety); internal consistency in this sample was Cronbach’s α = .94. Trait mindfulness was measured using the 15‑item Five‑Facet Mindfulness Questionnaire (FFMQ‑15), which yields a total score and five subscales (Observing, Describing, Acting with awareness, Non‑judging, Non‑reactivity); subscale αs ranged from .64 to .87 in this sample. Analytically, participants were grouped as Current MD, Former MD or No MD. Between‑group differences on demographics were tested with ANOVA for age and chi‑squared tests for categorical variables. Associations between microdosing status and STAI‑T scores were tested with linear regression, and mediation by FFMQ‑15 total and subscale scores was examined following a joint significance approach (testing path a: microdosing → FFMQ; path b: FFMQ → STAI‑T), with significance of indirect effects assessed by percentile bootstrapping using the PROCESS macro (5,000 samples). Exploratory analyses examined the role of previous macrodose experience by re-running regressions and mediation models excluding participants with such experience. The significance threshold was α = .05 and analyses were performed in SPSS v25. The authors report a priori power considerations for mediation (assuming a small path a and a medium path b), indicating a desirable sample size of approximately 400 for 80% power for joint significance testing.

After data cleaning, the analytic sample comprised 497 participants, of whom 263 (52.9%) had microdosing experience; the extracted text does not present a full table of demographic breakdowns but notes between‑group differences in age, gender and education. Regression analyses showed that Current MD status was associated with lower trait anxiety relative to No MD: the model reported r = .22, adjusted r 2 = .05, F(1,418) = 20.74, with an unstandardized coefficient B = -5.26 (SE = 1.16, p < .001). Former MD status was also associated with lower trait anxiety compared to No MD: r = .15, adjusted r 2 = .02, F(1,309) = 7.49, B = -4.33 (SE = 1.58, p = .007). Mediation tests indicated that the association between Current MD and STAI‑T scores was mediated by the FFMQ‑15 total score and by each FFMQ subscale. Path a (Current MD → FFMQ scores) and path b (FFMQ scores → STAI‑T) coefficients were all significant, and bootstrapped indirect effects (5,000 samples) produced 95% confidence intervals that did not include zero for the total and all subscale scores, indicating significant mediation. For Former MD, mediation was observed for FFMQ‑15 total as well as the Observing, Non‑judging and Non‑reactivity subscales; again, bootstrapped indirect effects were significant with confidence intervals excluding zero. In an exploratory set of analyses that excluded participants with previous macrodose experience (n = 386 excluded), associations between microdosing experience and STAI‑T scores became non‑significant: reported statistics were r = .05, adjusted r 2 < .01, F(1,202) = .57, with B = 49.99 (SE = 1.02, p = .453) as extracted text reports. Associations between microdosing and FFMQ‑15 total scores also became non‑significant (r = .10, adjusted r 2 < .01, F(1,181) = 1.78, B = 46.17, SE = .66, p = .184 as reported). Separately, previous macrodose experience was itself associated with lower STAI‑T scores compared to macrodose‑naïve participants: r = .15, adjusted r 2 = .02, F(1,495) = 11.84, B = -3.76 (SE = 1.09, p = .001). This macrodose–anxiety association was mediated by FFMQ‑15 total scores: macrodose experience predicted higher FFMQ total (path a: B = 3.73, SE = .78, p < .001) and higher FFMQ total in turn predicted lower STAI‑T (path b: B = -0.92, SE = .05, p < .001). The authors reported a standardized indirect effect approximated as .21 × .67 = .14.

Hartong and colleagues interpret the findings as showing that both current and former microdosers reported lower trait anxiety than microdosing‑naïve controls, and that these associations were statistically mediated by trait mindfulness, particularly the Non‑judging and Non‑reactivity facets, which produced small indirect effects. They emphasise, however, that exploratory analyses excluding participants with prior macrodose experience removed the associations between microdosing, mindfulness and anxiety, and that prior macrodose experience itself related to lower trait anxiety via higher trait mindfulness. From this pattern the authors suggest that previous macrodose psychedelic exposure may account for the anxiety differences observed between microdosers and controls in this sample. The paper situates this interpretation in prior work linking the subjective intensity of mystical experiences during macrodose sessions to longer‑term improvements in anxiety and depression, noting that microdoses are generally unable to induce such mystical experiences. Consequently, if mystical experiences are a necessary mechanism for therapeutic effects, that could explain why microdosing per se showed no association with anxiety once macrodose history was removed. The authors also acknowledge alternative explanations, especially expectancy and placebo effects, citing mixed evidence from prior placebo‑controlled and observational studies and concluding that more placebo‑controlled experimental work is required. Key limitations reported include the cross‑sectional design, which precludes causal inference; a broad and heterogeneous operationalisation of "microdosing" that left dose, frequency, regimen duration and time since cessation unspecified; and potential construct validity issues with the STAI‑T, which may reflect general negative affect rather than anxiety alone. The authors further note demographic imbalances between groups (for example, the control group included many female first‑year students), which could confound observed differences because young female students may report higher trait anxiety. Strengths highlighted were the relatively large sample for a microdosing study and the use of recommended mediation procedures and bootstrap testing. The authors conclude by calling for randomized, placebo‑controlled trials that control for prior macrodose experience to test causal hypotheses about microdosing and the potential mediating role of mindfulness.