In a randomised controlled trial of 80 healthy adult males who received 10 µg LSD or placebo every third day for six weeks, a multimodal battery of creativity tests (AUT, RAT, CAT, EPSQ) found no acute or lasting enhancement of creativity despite participants reporting feeling more creative on dose days.
Introduction
Enhanced creativity is often cited as an effect of microdosing (taking repeated low doses of a psychedelic drug). There have been recent efforts to validate the reported effects of microdosing, however creativity remains a difficult construct to quantify.
Objectives
The current study aimed to assess microdosing’s effects on creativity using a multimodal battery of tests as part of a randomised controlled trial of microdosing lysergic acid diethylamide (LSD).
Methods
Eighty healthy adult males were given 10 µg doses of LSD or placebo every third day for six weeks (14 total doses). Creativity tasks were administered at a drug-free baseline session, at a first dosing session during the acute phase of the drug’s effects, and in a drug-free final session following the six-week microdosing regimen. Creativity tasks were the Alternate Uses Test (AUT), Remote Associates Task (RAT), Consensual Assessment Technique (CAT), and an Everyday Problem-Solving Questionnaire (EPSQ).
Results
No effect of drug by time was found on the AUT, RAT, CAT, or EPSQ. Baseline vocabulary skill had a significant effect on AUT and RAT scores.
Conclusions
Despite participants reporting feeling more creative on dose days, objective measurement found no acute or durable effects of the microdosing protocol on creativity. Possible explanations of these null findings are that laboratory testing conditions may negatively affect ability to detect naturalistic differences in creative performance, the tests available do not capture the facets of creativity that are anecdotally affected by microdosing, or that reported enhancements of creativity are placebo effects.
Papers cited by this study that are also in Blossom
Creativity is commonly reported as a subjective benefit of psychedelic microdosing, yet controlled laboratory studies have rarely detected objective improvements on standard creativity measures. The literature reviewed by Murphy and colleagues highlights conceptual and methodological challenges: creativity is often defined as producing outputs that are both novel and useful, and this bipartite conception has motivated tests of divergent thinking (idea generation) and convergent thinking (idea evaluation). Previous high-dose psychedelic studies have suggested acute increases in associative or generative processes alongside reductions in evaluative control, and post-acute effects have been inconsistent across substances, contexts and measurement approaches. Microdosing is proposed as a way to obtain associative benefits while preserving cognitive control, but existing microdosing studies are mostly uncontrolled, rely on self-report, or have used creativity tasks whose construct validity has been questioned. This study set out to address the gap between subjective reports and laboratory measures by administering a multimodal battery of creativity assessments within a placebo-controlled microdosing trial (the MDLSD trial). The battery combined standard divergent and convergent tasks (Alternate Uses Test, AUT; Remote Associates Task, RAT), an applied visual-art Consensual Assessment Technique (CAT), and a bespoke Everyday Problem-Solving Questionnaire (EPSQ) to capture everyday idea generation and evaluation. Given the limited and mixed prior evidence, the investigators did not state directional hypotheses but aimed to test whether acute (on-dose) or durable (two days after a six-week regimen) changes in creativity could be detected across these modalities.
The MDLSD trial randomised 80 healthy male participants to receive either 10 µg LSD (n = 40) or placebo (n = 40). Creativity assessments were secondary measures within the trial. Testing occurred at three visits: a drug-free Baseline, a Treatment visit approximately one week later at which the first 10 µg dose was administered and creativity tasks were repeated at 240 minutes post-dose, and a Final visit two days after the final in-home dose (participants self-administered 13 further doses on an every-third-day schedule). Task order within the battery was AUT, CAT, RAT and was not counterbalanced. Some AUT and RAT administrations were moved to remote (home) completion in one study wave because of COVID-19 restrictions; Treatment visits remained on-site. The creativity battery comprised: (1) Remote Associates Task (RAT) — three counterbalanced testing versions of 20 trials each, with number correct and number attempted as outcomes; (2) Alternate Uses Test (AUT) — three counterbalanced versions, scored for frequency (fluency), flexibility, elaboration and originality; (3) Consensual Assessment Technique (CAT) — a 15-minute visual-art task (create a design conveying ‘silliness’) with photographs rated by 16 secondary school art teachers for creativity and technical goodness on 0–10 visual analogue scales; and (4) Everyday Problem-Solving Questionnaire (EPSQ) — a bespoke instrument in which participants nominated a real-life problem context and rated difficulty of idea generation (mean of difficulty thinking of and visualising solutions) and idea effectiveness (mean of practicality and satisfaction). Language ability was measured at Baseline using the NIH Toolbox Picture Vocabulary Test to control for lexical skill in linguistic tasks. Analyses used linear mixed effects models (lmerTest in R) with Group (LSD vs placebo) and Visit (Baseline, Treatment, Final) as fixed effects and Participant as a random effect, conducted on an intention-to-treat basis. Vocabulary was included as a covariate in AUT and RAT models. Study wave (to index home versus lab administration) was examined in follow-up analyses. Effect sizes were reported as partial eta squared and inter-rater reliability for subjective ratings was assessed via Cronbach's alpha; an alpha > 0.7 was the threshold for acceptable agreement. The extracted text does not clearly report the exact Cronbach alpha values or all details of withdrawals, but notes that 75 participants completed the full course of doses (placebo = 39, LSD = 36) and that a small number of task administrations were missing or corrupted.
Across the multimodal battery there was no evidence of a drug-by-visit interaction indicating acute or durable effects of microdosing on measured creativity. The AUT showed no Group x Visit interaction and no main effects of Group or Visit on fluency, flexibility, elaboration or originality. Vocabulary had a significant effect on all AUT scales (for example, flexibility F = 20.42, p < 0.001, ηp2 = 0.207), with higher baseline vocabulary associated with higher AUT scores. The RAT likewise showed no Group x Visit interaction and no main effects of Group or Visit for number of items correct or attempted; however vocabulary strongly predicted RAT accuracy (F = 45.04, p < 0.001, ηp2 = 0.355), and participants with higher vocabulary got more RAT items correct. The CAT (external assessment of visual-art outputs) produced no interaction or main effects of Group or Visit on either creativity or technical goodness ratings as provided by the teacher raters. The EPSQ showed no Group x Visit interaction on idea generation difficulty or idea effectiveness, but there was a significant main effect of Visit on idea generation difficulty (F = 47.35, p < 0.001, ηp2 = 0.378): participants, regardless of group, rated idea generation as more difficult at Baseline than at the Final visit, with an estimated marginal mean difference of 17 points (SE = 2.47, p < 0.001). Follow-up analyses found no effect of study wave (home versus lab administration) on AUT or RAT outcomes. The extracted text documents some missing or corrupted task data (three AUT missing at one Treatment session in the placebo group; one missing/corrupted RAT instance in the LSD group at Baseline and Final; one CAT missing in the placebo Final) but does not provide participant-level attrition beyond the summary completion counts.
Murphy and colleagues report null findings across multiple objective measures of creativity: neither acute on-dose effects at 240 minutes nor durable effects two days after a six-week microdosing regimen were detected on divergent thinking (AUT), convergent thinking (RAT), externally rated practical creativity (CAT), or the bespoke everyday problem-solving questionnaire (EPSQ). The authors note a clear influence of baseline vocabulary on performance in the linguistic tasks, emphasising the importance of controlling for domain-specific skill when interpreting RAT and AUT results. By contrast, a reduction in self-rated difficulty of idea generation across visits was observed regardless of group, suggesting trial participation or placebo/expectancy effects influenced this self-report measure. Several explanations for the discrepancy between subjective reports of enhanced creativity and the objective null results are considered. First, limited statistical power could conceal small effects given the parallel-group design with 40 participants per arm. Second, the controlled laboratory testing environment may not capture the contexts in which microdosers report benefits — set and setting could moderate creative effects and the standardised clinic setting may suppress them. Third, standard creativity tests have contested construct validity: domain-specific tasks, task impurity (overlapping cognitive processes), closed-problem formats, and lack of measures of effectiveness can all limit sensitivity to the kinds of creative change microdosers describe. For instance, the RAT may be affected in opposing directions by increased semantic associativity and impaired convergent control under psychedelics, producing net null effects. Finally, expectancy effects may play a role: creativity was the domain with highest baseline expectancy among participants and prior analyses of the trial showed dose-day increases in self-rated creativity and wellbeing; however objective task performance did not mirror those subjective shifts. The authors suggest that subjective feelings of creativity may reflect mood-related uplift rather than measurable improvements in creative output, yet they also note potential therapeutic value if such feelings enhance wellbeing or could be harnessed in conjunction with art therapies. Limitations acknowledged include potential underpowering for subtle effects, lack of counterbalancing within the battery (raising fatigue concerns), presentation of the creativity battery as a secondary measure which prevented scheduling at peak subjective effects, restriction to a single low dose (10 µg), absence of a baseline control measure of visual-art skill, unknown test–retest reliability for the specific item sets and timepoints used, and COVID-19-related protocol adaptations that necessitated some remote testing. The authors conclude that further research should explore set and setting manipulations, include domain-specific skill controls or expert samples, and consider alternative or more ecologically valid outcome measures to determine whether the subjective creative benefits of microdosing can be operationalised or are primarily expectancy- or mood-driven.
In this placebo-controlled trial, a six-week microdosing regimen of 10 µg LSD produced no measurable changes in creativity on a multimodal battery administered acutely and two days after the regimen concluded. Participants from the same sample reported feeling more creative on dose days, suggesting a dissociation between subjective creative feeling and objective creative performance. The authors posit that any microdosing-related creative effects may be too small for standard laboratory tests to detect, may be confined to a subjective mood-related experience, or may depend on contextual factors such as set and setting; they recommend future work to manipulate these factors and to assess whether subjective creative feelings can be translated into functional or therapeutic benefit.
The MDLSD trial consisted of eighty healthy male participants randomised into either LSD (n = 40) or placebo (n = 40) groups. Full inclusion and exclusion data is included in the Supplementary Materials. Creativity tasks took approximately 30-45 min and were first undertaken at a drug-free Baseline session (Table). The order of tasks was AUT, CAT, RAT. Task order was not counterbalanced within the battery. Approximately one week later, participants returned for a Treatment session, in which they were administered their first 10 µg LSD microdose under supervision and creativity tasks were repeated at 240 min after taking the dose. Peak subjective effects of 10 µg LSD have previously been observed at 150 min post-dose (Holze et al. 2020), however as the creativity tasks were secondary measures, priority scheduling at peak of effects was given to primary EEG measures. Pharmacodynamic data from the present study (currently in review) show that at 240 min participants' ratings of feeling an effect were at 81.7% of the maximum observed effect, indicating that acute effects could still be regarded as being present at this time point. Participants then self-administered 13 subsequent doses at home on an every-third-day protocol with some flexibility. Two days after their final dose, participants returned for a drug-free Final visit and creativity tasks were repeated. Participants were run in four waves ofto think of this linking word in order to measure convergent thinking and associative spread). It is possible that these tasks used in the controlled studies to-date are not entirely adequate for capturing the experiences of creativity reported by microdoses. Creativity tasks of this kind have been criticised for giving an incomplete picture of creative processes by being too domain specific, presenting closed rather than open problems, and failing to adequately capture the criteria of efficiency and the phenomenon of insight). An alternative approach that may have more construct validity in the microdosing context is testing creative output, such as art or creative writing using the Consensual Assessment Technique (CAT). In this task, participants are given a practical art or writing assignment, and outputs are then scored for their level of creativity by a panel of experts. These experts are asked to rate both the creative and technical elements of the piece in relation to the overall performance of the group. This approach is argued to be more valid than tests like the RAT and AUT because it is a measure of applied creative activities which accounts for effectiveness by being externally rated, and should therefore be considered the gold standard of creativity testing (Amabile 1996;. The limitations of this approach however are that the tests are still domain-specific, and they may be affected by participants' own proficiency in that particular domain. The CAT task has not previously been administered during any studies of microdosing psychedelics. The degree to which creativity is domain-specific or domain-general has been debated, but it is reasonable to conclude that creative processes require elements of both types of abilities and that creative testing in only one domain is inadequate to get a complete picture. In order to overcome domain specificity in creativity testing, the current study implemented a multimodal approach to test across several domains and cognitive processes, an approach that has been recommended in the literature. Our creativity battery included the RAT and AUT in order to replicate existing studies of microdosed psychedelics), a visual art CAT task to add a different modality with more practical creative output and a degree of assessment by others, and daily VAS ratings of participants self-rated feelings of being creative (reported previously in. In this way, the MDLSD study covered all the four assessment domains identified by-divergent thinking, convergent thinking, assessment by others, 4. Originality: the uniqueness of the response (one point for responses given by less than 15% of the sample, two points if given by less than 10%, three points if given by less than 5%). Two independent raters coded each item for the first 10 participants, after which Cronbach's alpha was calculated using the alpha function from the psych package in R. If an alpha value greater than 0.7 was achieved, only one rater then continued with rating the entire dataset.
Each task and questionnaire was analysed with a linear mixed effects model using the lmerTest package in Rwith Group and Visit treated as fixed effects, and Participants as random effect. Linear mixed effects modelling was chosen due to the ability to accommodate missing data without excluding participants, and to account for the random effects of participants variable abilities at the tasks. Language ability was controlled for in the AUT and RAT analyses by including scores on the NIH Toolbox Picture Vocabulary Test) as a fixed effect covariate. The vocabulary test was administered via the NIH Toolbox iPad app at the Baseline visit (National Institutes of Health 2019). Significant results were uncorrected and considered exploratory. Posthoc analyses were conducted by calculating the estimated marginal means using the emmeans package in R. To check whether the change in administration location (home versus in the lab) was significant, a follow up analysis was conducted on the RAT and AUT scores which also included study wave (1-4, see Supplementary Materials for description of varying study wave conditions), as a fixed effect. All analyses were based on intention-to-treat. Effect sizes were calculated as partial Eta squared (proportion of variance explained by fixed effects and interactions separately; η p 2 ) using the effectsize package in R. The inter-rater reliability of AUT and CAT ratings was tested by computing Cronbach's alpha using the alpha function from the psych package in R. Separate alpha values were computed for 'creativity' and 'technical goodness' ratings in the CAT. In the AUT, alpha was calculated for flexibility and originality given that these two ratings are subjective. An alpha value over 0.7 was considered acceptable.
This study tested the acute and durable effect of microdosing on creativity using four different modalities. Consistent with previous controlled and semi-controlled acute experiments, no interaction effect of drug group by visit was seen on divergent thinking as measured by the AUT, nor convergent thinking as measured by the RAT. No effect was seen on externally assessed practical creativity as measured by the CAT, nor on self-assessment as measured by the EPSQ. A significant effect of vocabulary was seen on accuracy in the RAT task and on all scales of the AUT, highlighting the necessary of vocabulary as a control variable in linguistic creativity tasks. There was a significant negative effect of visit on idea generation difficulty, regardless of group, suggesting that placebo effects or an effect of being in the trial affected participant self-ratings of creativity. The null results in this study are in contrast to the significant increase in self-rated creativity on the dose days reported inand to the self-rated experiences of microdosers in the community. The gap between the findings inside and outside the lab may simply be explained by lack of power in the laboratory studies, however there are three other possible explanations: (1) testing conditions in the lab may dampen or mask any potential creativity benefits of microdosing; (2) the tests used are not valid for testing the types of creativity enhanced by microdosing; (3) enhancements to creativity reported in the grey literature are placebo effects.
Create a free account to open full-text PDFs.
Anderson, T., Petranker, R., Christopher, A. et al. · Harm Reduction Journal (2019)
Anderson, T., Petranker, R., Rosenbaum, D. et al. · Psychopharmacology (2019)
Andersson, M., Kjellgren, A. · Harm Reduction Journal (2019)
Baggot, M. J. · Preprints (2015)
Bershad, A. K., Schepers, S. T., Bremmer, M. P. et al. · Biological Psychiatry (2019)
Carhart-Harris, R. L., Erritzoe, D., Williams, T. et al. · PNAS (2012)
Carhart-Harris, R. L., Muthukumaraswamy, S., Roseman, L. et al. · PNAS (2016)
Cavanna, F., Muller, S., de la Fuente, L. A. et al. · Translational Psychiatry (2022)
Daws, R. E., Timmermann, C., Giribaldi, B. et al. · Nature Medicine (2022)
Doss, M. K., Považan, M., Rosenberg, M. D. et al. · Translational Psychiatry (2021)
Mason, N. L., Mischler, E., Uthaug, M. V. et al. · Journal of Psychoactive Drugs (2019)
Mason, N. L., Kuypers, K. P. C., Reckweg, J. T. et al. · Translational Psychiatry (2021)
Molla, H. M., Lee, R., Tare, I. et al. · Neuropsychopharmacology (2023)
Murphy, R., Sumner, R. L., Evans, W. J. et al. · Biological Psychiatry (2023)
Murphy, R. J., Godfrey, K., Shaw, A. D. et al. · BMC Psychiatry (2024)
Murphy, R., Muthukumaraswamy, S., De Wit, H. · Biological Psychiatry (2024)
Polito, V., Stevenson, R. J. · PLOS ONE (2019)
Prochazkova, L., Lippelt, D. P., Colzato, L. S. et al. · Psychopharmacology (2018)
Family, N., Vinson, D., Vigliocco, G. et al. · Language Cognition and Neuroscience (2016)
Girn, M., Mills, C., Roseman, L. et al. · NeuroImage (2020)
Hartogsohn, I. · Drug Science Policy and Law (2017)
Zhang, D. Z., Holze, F., Liechti, M. E. et al. · Clinical Pharmacology and Therapeutics (2020)
Janiger, O., Dobkin De Rios, M. · Journal of Psychoactive Drugs (1989)
Kaertner, L. S., Steinborn, M. B., Kettner, H. et al. · Scientific Reports (2021)
Kiraga, M. K., Mason, N. L., Uthaug, M. V. et al. · Frontiers in Pharmacology (2021)
Kuypers, K. P. C., Riba, &. J., De La Fuente Revenga, &. M. et al. · Psychopharmacology (2016)
Lea, T., Amada, N., Jungaberle, H. et al. · International Journal of Drug Policy (2020)
Wießner, I., Falchi, M., Maia, L. O. et al. · Journal of Psychopharmacology (2022)
Sessa, B. · Journal of Psychopharmacology (2008)
Szigeti, B., Kartner, L., Blemings, A. et al. · eLife (2021)
Papers in Blossom that reference this study
Prochazkova, L., Marschall, J., Lippelt, D. P. et al. · Neuropharmacology (2026)
Murphy, R. J., Wardlaw, M., Smith, T. et al. · Journal of Humanistic Psychology (2025)