Psilocybin intoxication did not affect daytime or sleep-related declarative memory consolidation in a small sample exploratory analysis

The study employed a double-blind, placebo-controlled crossover design in 20 healthy volunteers (10 male, 10 female; mean age 36, SD = 8.1), with sessions at least 28 days apart. This cognitive substudy formed part of a larger clinical trial; it was nested in the EEG arm. Participants were screened for psychiatric and medical exclusions using standard instruments and procedures; 14 had prior psychedelic experience and five were naïve. Women were not dosed during menses. One participant was excluded for excessive daytime sleepiness during placebo EEG, and one withdrew from the second part of the study, leaving 19 participants for most analyses and smaller N in some task-specific analyses as noted. Three established neuropsychological tasks were used. The Groton Maze Learning Task (GMLT) assessed spatial/executive memory (five learning trials and a delayed trial with alternate maze forms to reduce practice effects). Verbal memory was assessed with the Rey Auditory Verbal Learning Test (RAVLT), including encoding trials, an interference list, delayed recall and a recognition trial; two alternative Czech forms were counterbalanced. Sleep-related declarative consolidation was probed with a Czech version of the Paired Associative Learning Test (PALT), comprising learning, immediate recall with feedback, short-delay evening recall, and morning recall after sleep. Polysomnography was recorded overnight for the sleep session. On each dosing day participants completed RAVLT and GMLT before dosing, then received oral psilocybin or placebo on an empty stomach. Psilocybin dosage was weight-adjusted at approximately 0.26 mg/kg (1 mg per 5 kg), producing a mean dose of 18.35 mg (SD = 2.21 mg); acute effects lasted roughly 6 hours and were monitored with the Brief Psychiatric Rating Scale (BPRS). After acute effects subsided participants completed delayed RAVLT and GMLT trials, and around 8 hours after dosing they performed the PALT learning phases, with evening and morning recalls surrounding an overnight sleep in the lab. Analyses were exploratory: no a priori power calculation was conducted and the trial was not powered to detect small or medium effects. Outcomes were analysed using mixed-design repeated-measures ANOVA, with treatment (psilocybin versus placebo) and time as within-subject factors, and treatment order and gender as between-subject factors where relevant. Data were inspected for missingness and errors, Box-Cox transformed to improve normality, and extreme outliers removed pairwise. The investigators report varying Ns per analysis due to incomplete data or outlier removal (for example, GMLT analyses included 17 completers; RAVLT retention used N = 20; recognition used N = 18; PALT used N = 19).

Across the tasks, mixed-model analyses did not identify a significant main effect of psilocybin on declarative memory consolidation measures in this sample. For the GMLT (spatial memory), 17 completers were analysed after preprocessing. Time of testing approached significance, with participants making more errors after the delay (F(1,15) = 4.172, p = 0.059), but no main effect of treatment was observed. A strong interaction between treatment and treatment order emerged (F(1,15) = 29.975, p < 0.001), driven by a robust practice effect: participants performed better on their second visit regardless of whether psilocybin or placebo was administered first (pairwise comparisons: placebo-psilocybin p = 0.013; psilocybin-placebo p = 0.036). RAVLT retention scores (N = 20) showed a sex effect: females outperformed males on retention (mean(SD) females = 1.45(1.99) vs males = 2.9(1.97); p = 0.036). There was no main effect of drug on retention. Recognition analyses (one participant did not complete the recognition test and one outlier was removed; N = 18) also found females outperforming males (mean(SD) females = 13.5(1.79) vs males = 11.9(1.54); F(1,14) = 4.712, p = 0.048) but no significant drug effects or interactions in the mixed ANOVA. PALT results probing sleep-related consolidation included N = 19 after exclusions. A main effect of time was present (Greenhouse-Geisser corrected F(1.12,16.78) = 202.463, p < 0.001), reflecting learning and immediate performance changes across the three measurement points; pairwise tests showed differences between immediate and evening and between immediate and morning recalls but not between evening and morning recalls, indicating no detectable benefit attributable to sleep in these data. There was a main effect of gender (F(1,13) = 5.913, p = 0.047) and a sex-by-treatment interaction (F(1,15) = 7.553, p = 0.015): females outperformed males in the placebo condition but this advantage was not present in the psilocybin condition. Brief Psychiatric Rating Scale scores indicated that participants were free of acute psilocybin effects by 6 hours post-dose. The authors emphasise that the study was exploratory and only powered to detect large effect sizes.

The investigators conclude that, contrary to their hypothesis, they found no evidence that a single, weight-adjusted oral dose of psilocybin enhanced declarative memory consolidation either during wake or across the first post-dose night in this sample. At the same time, the data provide no indication that psilocybin impairs consolidation of non-emotional spatial or verbal material; this is presented as an important safety-relevant finding for clinical and laboratory contexts. Nikoli and colleagues note replication of known effects: females performed better than males on verbal memory tests, and a substantial practice effect was observed on the GMLT despite the use of alternate maze versions. They situate their null treatment effects alongside preclinical and clinical literature that is mixed—animal studies have reported both impairments and facilitation depending on timing and design, and prior human trials have shown no persistent memory deficits after dosing. The authors propose that any psilocybin influence on memory consolidation may be domain-specific, potentially affecting emotionally charged memories rather than non-emotional, semantically related material assessed here. Key limitations acknowledged include the small, exploratory sample and lack of power to detect small-to-moderate effects; the absence of control for age and education which can influence RAVLT performance; not testing sleep-dependent consolidation during acute intoxication or including a wake-control condition; and the practical limitations of the GMLT in this study due to an unexpected practice effect and questions about its test-retest reliability. The authors recommend larger samples, inclusion of learning phases before dosing with assessment both after dosing and after first sleep, and consideration of emotionally valenced memory tasks in future work to clarify whether psilocybin selectively affects emotional memory systems.

The study's clinical-trial data in healthy volunteers indicate that acute psilocybin intoxication does not significantly alter spatial or verbal declarative memory consolidation, nor sleep-dependent consolidation across the night following dosing. These findings extend the evidence base regarding the safety of psilocybin with respect to the consolidation of non-emotional declarative memories in a controlled laboratory setting.