Daytime ayahuasca administration modulates REM and slow-wave sleep in healthy volunteers

Ayahuasca is a traditional Amazonian psychoactive brew that contains N,N-dimethyltryptamine (DMT) together with β-carboline monoamine oxidase inhibitors, a combination that renders DMT orally active. DMT acts on serotonergic receptors (partial agonism at 5-HT1A and 5-HT2A/2C sites), and previous human work with ayahuasca and other psychedelics has documented stimulant-like subjective effects, perceptual changes and measurable alterations in wake EEG and regional cerebral blood flow. Because serotonin plays a central and complex role in sleep–wake regulation and different serotonin receptor subtypes have distinct effects on REM sleep and slow-wave sleep (SWS), the authors identified sleep physiology as a key domain in which ayahuasca might produce measurable effects. This study set out to characterise the acute effects of daytime ayahuasca administration on objective polysomnographic (PSG) measures, sleep EEG power spectra, and subjective sleep quality in healthy volunteers. A placebo-controlled, double-blind, cross-over design was used, with d-amphetamine included as a positive control to benchmark stimulant-induced sleep disruption. The investigators hypothesised that, because of its arousing properties, ayahuasca would impair sleep initiation, maintenance and perceived sleep quality similar to d-amphetamine, but that its serotonergic pharmacology would also produce specific modulations of REM and SWS.

Twenty-two healthy male volunteers were recruited; mean age was 27.1 years and mean weight 67.6 kg. Eligibility required prior use of psychedelics on at least ten occasions without sequelae. Screening included a structured psychiatric interview (DSM-IV), medical and laboratory tests, serology and urine drug screening, and assessment for baseline sleep quality. Exclusion criteria included current or past Axis I disorders and substance dependence. Participants were asked to maintain regular sleep–wake schedules and to avoid alcohol, excessive caffeine, cigarettes, naps and strenuous exercise during specified windows around each session. The local ethics committee approved the protocol and all volunteers gave written informed consent. The interventions were three treatments given in randomised, double-blind, cross-over order with at least 1 week between sessions: placebo, 20 mg d-amphetamine (positive control), and a freeze-dried encapsulated ayahuasca formulation dosed to provide 1 mg DMT per kg body weight. The lyophilisate composition per gram was reported (including harmine, harmaline and tetrahydroharmine). Capsules were prepared so that volunteers received the same number of capsules each session. On each experimental day participants arrived at 07:00, had breakfast prior to 10:00, received the assigned capsules at 12:00 noon and remained under supervision in the laboratory until 12:00 the following day; sleep recordings were made from around 23:00 to 07:00. An adaptation night was performed within two weeks prior to the first session and was not included in analyses. Subjective acute drug effects were assessed by Spanish versions of the Hallucinogen Rating Scale (HRS) and the Addiction Research Center Inventory (ARCI); ARCI was administered pre-dose and at 4 h post-dose and transformed to change scores, while HRS was administered at 4 h post-dose. Sleep was recorded with full polysomnography including six EEG channels (Fp1, Fp2, C3, C4, O1, O2 referenced to mastoids), two EOG leads, chin EMG, respiratory channels and bilateral anterior tibialis for limb movements. PSG was scored visually in 30-s epochs using standard Rechtschaffen & Kales criteria by two independent blinded scorers, with disagreements resolved by a third expert. All-night spectral analysis was performed on digitised EEG (sampling 256 Hz) after filtering; power spectra were computed from 5-s artefact-free epochs using FFT and aggregated into frequency bins (0.4 Hz for 0.2–6.0 Hz and 0.8 Hz for 6.2–26.0 Hz) for derivation C4–A1. Variables derived included slow-wave activity (SWA, 0.5–4.0 Hz), spindle frequency activity (SFA, 11–15 Hz) and delta EOG activity (DEA) during REM. To compare dynamics across unequal NREM/REM period lengths, each NREM period was subdivided into 24 equal parts and each REM period into four parts before averaging across subjects; areas under the curve (AUC) for dynamics were used in comparisons. Statistical analysis grouped PSG variables into clusters (sleep initiation/maintenance; sleep architecture; NREM–REM period variables) plus subjective sleep/awakening and subjective drug-effect clusters. Each cluster was submitted to a repeated-measures multivariate analysis of variance (MANOVA) with treatment as the within-subject factor; Greenhouse–Geisser correction was applied where appropriate. Pairwise comparisons used repeated-measures t tests with Sidak correction. For EEG spectra and AUCs, repeated-measures t tests compared each active treatment with placebo, and significance was set at two-sided p<0.05. Four participants were excluded from PSG analyses due to technical problems, leaving a final analysed sample of 18 with complete PSG data.

The final analysed sample comprised 18 participants for PSG and spectral analyses (four initial subjects excluded for technical reasons); demographic characteristics did not differ between initial and final samples. Acute subjective drug effects showed a significant overall effect (MANOVA Pillai's trace reported; p<0.001). Both d-amphetamine and ayahuasca increased ARCI-A (amphetamine-like effects) and ARCI-MBG (euphoria) subscales, and both produced somatic-dysphoric effects on the ARCI-LSD scale. Ayahuasca, but not d-amphetamine, produced significant increases on HRS subscales of perception and volition, reflecting psychedelic-specific perceptual changes and altered volitional capacity. Sleep initiation and maintenance variables also differed between treatments (MANOVA Pillai's trace p=0.001). d-Amphetamine substantially delayed sleep onset, producing significant increases in latencies to stage 1, stage 2 and REM compared with both placebo and ayahuasca, and showed a trend to increase stage 3 latency. It additionally reduced total sleep period (TSP), total sleep time (TST) and sleep efficiency index (SEI), and increased stage 0 time, total wake time and number of awakenings per TSP, relative to placebo and ayahuasca. Ayahuasca did not differ significantly from placebo on measures of sleep initiation or maintenance, although it tended to increase REM latency. Regarding sleep architecture, significant treatment effects were present (MANOVA p<0.001). d-Amphetamine reduced REM duration (minutes and % of TST), reduced SWS when measured in minutes, and increased stage 1 as a percentage of TST; it also increased stage 2 minutes. Ayahuasca significantly decreased REM duration (both minutes and % of TST) compared with placebo, and showed a tendency to increase stage 2 (minutes and % of TST). NREM–REM period analyses showed both active compounds decreased the number of NREM and REM periods and increased the average duration of NREM periods and sleep cycles versus placebo, but the effects were larger after d-amphetamine (MANOVA p<0.001). All-night EEG power spectra differed by treatment. During NREM sleep, d-amphetamine increased power in the high-frequency range above 15 Hz across the spectrum, whereas ayahuasca produced significant increases limited to the 15–20 Hz band. Stage 2 spectra paralleled the NREM findings. In REM sleep, d-amphetamine reduced power density relative to placebo in the 2.4–2.8 Hz and 5.6–8.4 Hz bands; no comparable REM reductions were reported for ayahuasca. Dynamics of slow-wave activity (SWA) across the night showed a decline over consecutive NREM episodes for all conditions. d-Amphetamine produced a significant reduction in SWA in the first sleep cycle compared with placebo (t=2.41, df=14, p=0.030). Ayahuasca showed a trend toward increased SWA in the first cycle vs placebo (t=1.82, df=17, p=0.086). Because treatment altered the number of cycles, a subgroup analysis was performed including only participants who had the same number of cycles after ayahuasca and placebo (n=8); in this subgroup SWA in the first cycle was significantly increased after ayahuasca (t=2.78, df=7, p=0.027). No analogous reanalysis could be done for d-amphetamine because all subjects had fewer cycles versus placebo. Spindle frequency activity (SFA) and delta EOG activity (DEA) during REM did not show significant treatment-related changes. Subjective sleep and awakening quality (SSA) were significantly impaired after d-amphetamine (MANOVA p<0.001): global SSA score and the subjective sleep quality subscale worsened, subjective sleep latency increased and subjective sleep efficiency decreased relative to placebo and ayahuasca. Ayahuasca did not produce significant subjective deterioration in sleep or awakening quality compared with placebo.

Barbanoj and colleagues interpret the findings as showing that daytime administration of ayahuasca produces measurable changes in sleep physiology without the overt subjective and objective sleep disruption seen after morning d-amphetamine. Both drugs elicited daytime stimulant-like effects on self-report measures, but ayahuasca uniquely increased HRS subscales related to perception and volition, consistent with its psychedelic profile. Objectively, d-amphetamine produced the expected stimulant pattern: delayed sleep onset, impaired sleep maintenance, increased light sleep, reduced SWS (minutes in the first cycle) and marked REM suppression. Ayahuasca did not deteriorate sleep initiation or maintenance, but it did reduce REM duration and tended to delay REM onset; it also reduced the number of NREM/REM periods and lengthened NREM periods and sleep cycles, albeit to a lesser extent than d-amphetamine. Spectral analysis revealed that both active treatments increased high-frequency EEG power during NREM, with d-amphetamine effects extending above 15 Hz and ayahuasca effects concentrated in the 15–20 Hz band. Crucially, whereas d-amphetamine reduced SWA in the first cycle, ayahuasca increased SWA in that cycle (significant in an eight-subject subgroup with equal cycle counts). The authors note that increased SWA in the first cycle is typically associated with greater sleep pressure, for example following sleep deprivation or physical exercise, and suggest several possible mechanisms: DMT agonism at 5-HT1A receptors or functional desensitisation of 5-HT2 receptors could modulate SWS in ways that differ from the expected 5-HT2A/C agonist effect; alternatively, the physiological stress or tiredness following the ayahuasca experience (including elevated cortisol) might augment sleep pressure. In placing their results in context, the investigators compare with older studies of classical psychedelics—some of which reported REM increases after immediate bedtime administration of LSD—highlighting that timing of drug administration (morning versus immediate pre-sleep) and the multi-alkaloid composition of ayahuasca (including harmine and tetrahydroharmine, which have MAOI and serotonin reuptake inhibition properties) may account for differing sleep effects. They also note that reversible MAO inhibitors and selective serotonin reuptake inhibitors are generally associated with REM suppression and increased stage 2, and that SSRI effects on sleep are sometimes more pronounced after morning dosing. The authors acknowledge limitations implicit in their data: the sample was small and restricted to healthy male volunteers with prior psychedelic exposure, four subjects had to be excluded for technical reasons reducing power, and the compound complexity of ayahuasca complicates attribution of effects to DMT alone. Nonetheless, they conclude that daytime serotonergic psychedelic administration produces detectable changes in PSG measures and sleep EEG spectra—specifically REM inhibition and an increase in early-night SWA after ayahuasca—suggesting interaction with brain circuits that regulate REM and SWS and meriting further study.