No evidence for deleterious effects of psilocybin and MDMA on working memory or related neurophysiological activity in a non-clinical population: An uncontrolled exploratory study

Psychedelic compounds such as psilocybin and MDMA have re-emerged as candidate treatments for selected psychiatric conditions, but their non-acute effects on cognition and brain activity remain poorly characterised. The paper notes that acute administration of both drugs is known to alter EEG rhythms, yet there has been very limited research comparing pre- and post-exposure neurophysiology beyond the acute phase, especially for MDMA. This gap matters because concerns about possible cognitive harm, particularly working memory and executive function impairment, could affect clinical development and safety assessments. Bailey and colleagues therefore set out to examine whether a single exposure to psilocybin or MDMA in a non-clinical sample was followed by changes in working memory performance and related EEG activity. They hypothesised that participants would show practice-like improvements rather than impairment, alongside increases in working-memory-related oscillatory activity, and explored whether the aperiodic 1/f slope of the EEG spectrum also changed after dosing. The study was explicitly exploratory and was designed to provide early evidence about possible enduring neurophysiological effects after controlled psychedelic administration.

The study was an uncontrolled exploratory component of the ASSESS trial, conducted in Melbourne, Australia between 2023 and 2024. Fifty healthy adults consented to participate; 48 were dosed, and 29 participants in each drug condition contributed usable pre- and post-dosing EEG data for the current working memory analyses. Participants chose whether they would receive MDMA or psilocybin in a given session, all individuals in the same dosing group received the same drug, and 16 participants received both drugs in separate sessions after a washout of more than 3 months. The extracted text does not clearly indicate any formal control condition, and the authors state that none was included because of resource and burden constraints. Eligibility criteria excluded pregnancy, unstable medical illness, relevant cardiovascular disease, psychoactive medication use, seizure history, moderate to severe traumatic brain injury, hepatic impairment, insulin-dependent diabetes, psychotic disorders, bipolar disorder, substance dependence, recent use of MDMA or psilocybin, and lifetime heavy use of either compound. Participants were adults aged 31 to 68 years, English-proficient, and had training in psychedelic-assisted psychotherapy. Dosing sessions were held in a low-stimulus clinical room with two experienced therapists and medical support. Psilocybin was given as a single capsule dose of 25 mg for those under 90 kg or 30 mg for those between 90 and 115 kg. MDMA was given as 80 mg initially, with an optional 40 mg supplement after 1 hour if subjective effects were judged inadequate; most included MDMA participants received the extra dose. All drugs were pharmaceutical-grade, GMP-standard products. EEG and behavioural assessments were completed 1 to 16 days before dosing and 5 to 16 days after dosing, with an additional 3-month remote follow-up for the behavioural task only. The main task was a 3-back working memory paradigm in which participants responded when the current letter matched the one shown three items earlier. The study also collected 10-minute eyes-closed resting EEG, using the resting data as a targeted comparison to determine whether any task-related changes reflected more general neurophysiological shifts. EEG was recorded with a 64-channel system and pre-processed using the RELAX pipeline, with artefact rejection, ICA-based cleaning, and baseline correction. Analyses focused on time-resolved frequency power, aperiodic slope and offset derived with FOOOF, oscillatory power after removing the aperiodic component, and behavioural accuracy using a combined measure of correct responses and correct non-responses. Pre- versus post-dosing comparisons were made within each drug condition using paired t-tests, repeated-measures ANOVA, Bayesian statistics, and permutation-based EEG cluster tests to manage multiple comparisons.

In the main EEG analyses of the 3-back task, Bailey and colleagues found a significant increase in global baseline-corrected frequency power from the pre-dosing to post-dosing session across a broad cluster spanning roughly 8 to 45 Hz and about 330 to 820 ms after stimulus presentation. This effect was driven by stronger post-dosing power, but there was no overall difference between drug groups in the topographical distribution of activity, and no significant group-by-session interaction in the topography analysis. When the averaged cluster values were analysed further, there was a significant main effect of session and a significant session-by-drug interaction, indicating that the size of the change differed between psilocybin and MDMA. Follow-up frequency-band analyses suggested that the post-dosing increase in baseline-corrected power was most evident in the MDMA condition across alpha, beta, and gamma ranges, whereas the psilocybin condition did not show significant effects on this specific baseline-corrected measure. After correcting for aperiodic activity, however, both drugs showed clusters of stronger oscillatory power post-dosing. For MDMA, these clusters were in delta, theta, and alpha ranges; for psilocybin, they were in upper theta to alpha, upper alpha to lower beta, and beta ranges. None of the oscillatory measures or their changes correlated with 3-back accuracy. For aperiodic activity, psilocybin produced a steeper 1/f slope in the post-dosing EEG session during the 3-back task. This was significant globally and appeared strongest over occipital electrodes, where the difference was large. The same occipital pattern was also present during eyes-closed resting EEG, where the global effect was only a trend but the occipital-focused comparison was significant. MDMA did not show significant changes in aperiodic slope, and neither drug changed the aperiodic offset. After aperiodic correction, resting EEG showed no significant post-dosing oscillatory differences for either drug. Behaviourally, accuracy on the 3-back task improved after MDMA in the pre- to post-dosing comparison, but not after psilocybin in the primary paired analysis. In the combined repeated-measures analysis across both drugs, there was a significant main effect of session but no main effect of group and no interaction, suggesting similar improvement across conditions overall. When the 3-month follow-up was included, accuracy was again significantly better at later timepoints, including improved performance at post-dosing and at 3 months relative to baseline. The authors also compared the size of this improvement with published practice effects and reported that the observed gains were broadly similar in magnitude. Overall, the results did not show any evidence of working memory impairment after either drug.

The authors interpret the findings as suggestive of enduring neurophysiological effects after a single controlled dose of psilocybin or MDMA, especially because the EEG changes were observed 5 to 16 days after administration rather than during the acute drug phase. They argue that the post-dosing increases in oscillatory power across theta to gamma frequencies, together with the steeper aperiodic slope after psilocybin, could reflect lasting neuroplastic or compensatory changes. They also note that the behavioural data are reassuring: at minimum, the results do not support concern that a single exposure to either drug impairs working memory in a controlled setting. Relative to earlier literature, the authors say their findings differ from the well-known acute effects of serotonergic psychedelics, which typically reduce alpha power and flatten the aperiodic slope. They suggest that the post-acute steepening of the aperiodic slope after psilocybin may indicate a different, possibly compensatory or plastic response occurring days later, and they note that the occipital localisation fits the known distribution of 5-HT2A receptors. They also point out that the MDMA findings are more consistent with some earlier non-acute reports of increased resting alpha and beta power in recreational users, though they stress that their study examined controlled pharmaceutical dosing rather than recreational use. The authors are cautious about interpretation because the study lacked a placebo or active control arm. They acknowledge that the observed EEG and behavioural improvements could reflect practice or familiarity effects from repeating the 3-back task, although they argue that the resting-state psilocybin slope change is harder to explain this way. Other limitations they discuss include the variable timing of post-dosing EEG testing, dose variability within each drug condition, lack of expectancy measures, the fact that 16 participants received both drugs across separate sessions, and the absence of sufficient power for mediation or dose-response analyses. They also note that the sample consisted of healthy volunteers trained in psychedelic-assisted psychotherapy, so generalisability to clinical populations is uncertain. In terms of implications, the authors say the study provides an example of rigorous psychedelic research under controlled conditions and offers preliminary reassurance that single-dose psilocybin and MDMA are unlikely to impair working memory. They suggest future work should include control conditions, larger samples, expectancy measures, and clinical populations, and should test whether the EEG changes relate to cognitive or therapeutic outcomes.

The authors conclude that single exposures to psilocybin and MDMA in controlled clinical settings did not impair working memory and may have been associated with improved performance and persistent EEG changes. They frame the findings as preliminary evidence against cognitive safety concerns in such settings, while emphasising that the absence of a control arm means the effects cannot be attributed definitively to the drugs. They present the study as a methodological template for future controlled research intended to separate drug-specific effects from practice and other confounds.