The Effects of Daytime Psilocybin Administration on Sleep: Implications for Antidepressant Action

Psilocybin and its active metabolite psilocin are serotonergic psychedelics that act primarily at 5-HT1A and 5-HT2A/C receptors and produce acute alterations of perception, cognition and emotion as well as longer-term improvements in mood and well-being. Because serotonergic psychedelics have been shown to promote neuroplasticity in preclinical work and because several conventional antidepressant treatments alter sleep architecture, sleep offers a plausible in vivo proxy to probe psilocybin-related neuroplastic effects. In particular, slow-wave activity (SWA) during non-rapid eye movement (NREM) slow-wave sleep is considered a marker of synaptic homeostasis and neuroplastic processes, while antidepressants commonly alter rapid eye movement (REM) sleep by reducing REM duration and prolonging REM latency. Dudysová and colleagues set out to fill a gap in the literature: no prior clinical sleep data existed for acute psilocybin administration. The study tested whether a single daytime dose of psilocybin would 1) prolong REM onset latency and 2) reduce REM proportion, and 3) increase SWA in the first sleep cycle. Whole-night polysomnography with 19-channel EEG was used to assess macrostructure and quantitative microstructure (spectral power) on the night following daytime dosing, with additional exploratory tests for subjective sleep measures and potential gender differences.

The investigators used a double-blind, placebo-controlled crossover design. Twenty healthy volunteers (10 women) aged 28–53 years (mean 36 ± 8.1) were enrolled; participants were screened to exclude regular medication use, major medical or psychiatric conditions, family history of psychosis, present sleep disorders, pregnancy and menopausal/postmenopausal status in women. Participants were instructed to avoid psychotropic drugs and excessive alcohol before the experiment; women were tested outside menses. Three participants were excluded from sleep analyses (one for excessive daytime sleepiness identified on placebo, two for technical polysomnography problems), leaving 17 subjects in the final analyses. Each participant completed two morning dosing sessions, psilocybin and inactive placebo, in randomized order with at least 28 days between sessions (mean interval 49 days). Psilocybin was administered around 09:00 in capsules adjusted by body weight at approximately 0.26 mg/kg (dose range ~15–22 mg, mean 18.35 mg, s.d. 2.21). Polysomnography was recorded roughly 12 hours after dosing (the night after administration) when acute psychedelic effects had dissipated. Whole-night recordings followed standard procedures: most participants had an adaptation night beforehand; EEG used a 19-electrode cap (10/20 system) plus EOG, EMG and ECG. Data were sampled and processed (band-pass 0.1–200 Hz, downsampled to 250 Hz) and visually scored by two blinded expert scorers with 87% interrater agreement (Cohen's kappa 0.82); a third scorer resolved epochs with <80% agreement. Objective sleep metrics (latency, stage durations, total sleep time, efficiency) and subjective sleep ratings (4-point Likert scales for latency, duration, quality) were collected. Quantitative EEG spectral analysis used FFT on 5-s artifact-free epochs (0.5–40 Hz filter), with power averaged by stage and log-transformed across frequency bands (delta 0.8–4.6 Hz, theta 4.6–8 Hz, alpha 8–12 Hz, sigma 12–15 Hz, beta1 15–20 Hz, beta2 20–35 Hz). SWA in the first sleep cycle was analysed by visually delimiting the first cycle and applying the same spectral approach. To reduce comparisons, 19 derivations were aggregated into frontal, central, parietal, temporal and occipital regions and also into an average electrode for some analyses. Paired t-tests or Wilcoxon signed-rank tests assessed condition differences; Sidak correction was applied for multiple comparisons in spectral analyses. Effect sizes (r) were reported and exploratory two-way repeated-measures ANOVAs tested gender-by-condition interactions where applicable.

Seventeen participants contributed to the sleep analyses. Psilocybin produced its expected acute psychedelic effects during the daytime session but recordings were made the following night when effects had dissipated. Macrostructure: Objective sleep latency, total sleep time, sleep efficiency and number of sleep cycles did not differ significantly between placebo and psilocybin. A statistically significant prolongation of REM latency was observed after psilocybin (one-tailed, uncorrected: z = -1.66, p = 0.048) with a small effect size (r ≈ -0.28). Sleep stage durations and proportions did not show significant differences after correction, although uncorrected trends indicated decreased REM, N1 and N3 durations and an increased N2 proportion following psilocybin. SWA in the first sleep cycle: Absolute delta power (an index of SWA) during slow-wave sleep in the first sleep cycle was significantly lower after psilocybin at the averaged electrode (t(16) = -3.57, p = 0.003, r = 0.67). Local decreases that survived correction for multiple comparisons were observed at averaged parietal (t ≈ -3.93, p = 0.001), temporal (t ≈ -3.40, p = 0.004) and occipital (t(13) = -3.26, p = 0.006) derivations with large effect sizes (r ≈ 0.70, 0.65, 0.67 respectively). Relative delta power decreases were not robust at the average electrode; a local uncorrected decrease appeared at occipital derivations (t(13) = -2.29, p = 0.039, r = 0.54) but did not remain significant after correction. Whole-night EEG spectra: No significant differences in absolute or relative power were found across N1, N2, N3 or REM stages after correction for multiple comparisons. Some uncorrected increases in relative sigma band power were noted frontally and parietally in overall NREM (N1–N3), but these did not survive correction. For the average electrode, there was a non-significant trend to decreased absolute delta power (t ≈ -1.87, p = 0.081, r = 0.42). Subjective measures: No significant changes in subjective total sleep time or sleep quality were detected. Participants, however, reported longer subjective time to fall asleep after psilocybin (mean increase ~10.5 minutes), a medium-sized effect (t(18) = 2.39, p = 0.028, r = 0.49). Gender analyses: No main effects of gender or gender-by-condition interactions were found for sleep macrostructure. Women showed higher mean relative delta power across electrodes irrespective of condition (F(1,15) = 4.584, p = 0.049), although this did not hold locally after correction. An interaction emerged for subjective sleep quality (F(1,12) = 15.429, p = 0.002): men reported significantly lower subjective sleep quality after psilocybin.

Dudysová and colleagues interpret the primary findings as evidence that daytime psilocybin administration alters sleep in ways that partly resemble conventional serotonergic antidepressants. In particular, the observed prolongation of REM latency and a trend toward reduced REM duration on the night after dosing are consistent with effects reported for SSRIs, SNRIs, TCAs and MAOIs, and such REM changes have been linked previously to antidepressant action. Contrary to the authors' hypotheses and to findings reported with ketamine, psilocybin reduced absolute delta power during slow-wave sleep in the first sleep cycle rather than increasing it. The investigators note that reduction of SWA accords with prior work on 5-HT2 receptor agonists (for example, fenfluramine) and propose several possible explanations: receptor-specific neurochemical effects distinct from ketamine, dose- and time-dependent neurogenic/plasticity effects, differences in pharmacokinetics between intravenous ketamine and oral psilocybin affecting the timing of plasticity processes, or an acute reduction in homeostatic sleep pressure leading to suppressed neural synchrony. They also raise the possibility that SWA suppression could be adaptive in depressed populations if it serves to normalise pathologically elevated SWS, citing prior work where disruption of SWA improved depressive symptoms. The authors therefore call for studies in clinically depressed samples to test associations between SWA changes and mood outcomes. Regarding quantitative EEG, no robust whole-night spectral changes survived correction for multiple comparisons. The uncorrected sigma-band increases in NREM could relate to sleep spindles and processes of memory consolidation and plasticity, but the authors emphasise that these findings are provisional and require larger samples. Subjective sleep continuity and overall sleep quality were largely unchanged, aside from longer perceived sleep latency after psilocybin, a result similar to prior ayahuasca findings. Gender analyses were mostly null for macrostructure, though women had higher relative delta power on average and men reported worse subjective sleep quality after psilocybin; the investigators advise caution interpreting these signals given sample size. Key limitations acknowledged by the study team include the small sample size and resulting limited power, the short-term focus (single-night follow-up only) which prevents conclusions about longer-term or repeated-dose effects, and the exclusion of phasic REM epochs from spectral analysis to avoid EOG artefacts, which may affect REM spectral estimates. The authors conclude that their study provides the first sleep data following psilocybin administration, documenting REM latency prolongation and SWA suppression, and recommend further research in larger and clinical samples to determine whether these sleep changes relate to psilocybin's putative antidepressant effects.