LSD increases sleep duration the night after microdosing

Microdosing—the repeated self-administration of sub-hallucinogenic doses of psychedelics such as LSD or psilocybin—has become a popular practice with claimed benefits for mood, creativity and productivity. Previous controlled work on macrodoses of serotonergic psychedelics has shown effects on sleep architecture (for example altered REM sleep and slow-wave power), and some older or uncontrolled reports have suggested that low doses of LSD can modify REM and increase body movements during sleep. However, objective, prospectively collected measurements of sleep in contemporary microdosing regimens are lacking, and community surveys give mixed subjective reports about sleep quality following microdosing. Allen and colleagues therefore set out to evaluate whether microdosing LSD alters objective sleep and activity measures. They embedded sleep monitoring into a Phase I, double-blind, randomised, placebo-controlled trial in which healthy volunteers self-administered 10 µg LSD or placebo every third day for six weeks while wearing consumer-grade wearable devices to capture naturalistic sleep and activity data. The trial aimed to provide the first objective assessment of sleep changes associated with a standardised microdosing schedule in an ecologically valid, home-administered setting.

The study (MDLSD) used a double-blind, parallel-group randomised design. Eighty healthy male volunteers aged 25–60 were randomised 1:1 to receive either LSD (10 µg base) or inactive placebo, administered sublingually via 1 mL oral syringe every third day for six weeks (fourteen doses total). Each participant completed screening, a seven-night baseline run-in, a first dosing visit one week after baseline, and a final follow-up visit scheduled two days after the last microdose (day 42). The trial was prospectively registered. Inclusion required male sex and the age range above; key exclusions included high resting blood pressure, extreme body weight, significant renal or hepatic impairment, unstable medical or neurological conditions, lifetime psychotic or bipolar disorder, current anxiety or eating disorders, suicidality, first-degree relatives with psychotic disorders, substance use disorder, use of psychotropic medication, recent serotonergic psychedelic use, or any lifetime history of psychedelic microdosing. Urine drug screens and breathalyser tests were used to verify abstinence at screening and baseline/first dosing respectively. Participants were provided with Fitbit Charge 3/4 devices and instructed to wear them throughout the trial; the Fitbit app synchronised data via participants' phones. The tracker reports sleep in four states: REM, Deep (N3), Light (N1+N2) and Awake (wake after sleep onset). Two composite variables were computed: Asleep (Deep + Light + REM) and Total (Deep + Light + REM + Awake), the latter approximating polysomnographic total sleep time. The team removed 20 duplicated entries from the Fitbit export and compared Fitbit’s Sleep Summary and Sleep Granular outputs, choosing the summary data because discrepancies were small (mean error ~2.11 minutes, max ~11.37 minutes for REM). Sleep events were assigned to dates using a 9am cut-off to align sleep start times with dosing days; Fitbit’s isMainSleep flag filtered out naps and merged interrupted main sleeps. For statistical inference, linear mixed-effects models were fitted with Group (LSD, placebo) and Day (Dose, Dose+1, Dose+2) as fixed effects and participants as a random effect, focusing on the Group × Day interaction. The baseline average over the seven-night run-in was included as a covariate. Analyses used an intention-to-treat approach (all available data analysed, no imputation for missing nights). The authors note that sleep analyses were exploratory because no pre-specified sleep hypotheses were declared; Bonferroni correction was applied within analyses to control for multiple comparisons.

Eighty participants began the trial; five did not complete protocol (four discontinued for mild anxiety, one for unrelated reasons), and three participants inadvertently received an extra dose due to scheduling issues. After data cleaning, 3,231 nights of sleep data were available for analysis: 503 baseline nights, 935 dosing-day nights, 927 nights the day after dosing (Dose+1), and 866 nights two days after dosing (Dose+2). The mean number of nights contributed per participant was 40.39 (sd = 9.96). Linear mixed-effects models adjusted for baseline showed a pattern of increased time across sleep metrics on the Dose+1 night for the LSD group relative to placebo. After Bonferroni correction (alpha threshold 0.00833 for six comparisons), three outcomes reached significance: REM sleep time (p = 0.0037), Asleep time (p = 0.0026) and Total sleep time (p = 0.0027). These corresponded to an additional 8.13 minutes of REM sleep (95% CI 3.34–12.9), 21.1 minutes of Asleep time (95% CI 8.9–33.2) and 24.3 minutes of Total sleep (95% CI 10.3–38.3) on the night after a microdose in the LSD group versus placebo. Similar results were obtained when the baseline covariate was omitted, and analyses of the granular sleep data produced the same pattern. Analyses of the proportion of time spent in each sleep stage found no changes approaching significance (even at an uncorrected p = 0.05), indicating that absolute increases in sleep duration did not reflect shifts in stage ratios. Activity data comprised 3,842 days (556 baseline, 1,102 dosing days, 1,101 Dose+1 days, 1,083 Dose+2 days) and showed no significant Group × Day interaction effects for activity metrics (calories, distance, steps) or activity-state minutes (sedentary, light, moderate, very active). The LSD group visually appeared to have lower moderate and very active minutes overall, but this pattern was present at baseline and no main effect was detected. Subjective tiredness was collected nightly; the authors previously reported a marginally significant increase in self-reported tiredness on Dose+1 days. Sleep-related adverse events were reported by 10 participants in the LSD group and 4 in the placebo group; this difference was not statistically significant in odds-ratio analyses. Semi-structured interviews identified variable reports of energy changes among LSD participants, with some describing post-dose fatigue and others noting bursts of energy; selected participant quotes appear in supplementary materials.

Participants who received 10 µg LSD microdoses showed a statistically significant and clinically meaningful increase in sleep the night after dosing, with an average of 24 additional minutes of total sleep and 8 extra minutes of REM, while no differences were observed on the dosing night itself. The authors interpret the Dose+1 increase as unlikely to be explained by expectancy alone, given the objective nature of the wearable-derived measurements and the null response in the placebo group. Practical implications highlighted include recommending at least one day off between microdoses to allow participants or patients to be well rested before the next dose. The investigators also note that the increase in sleep could plausibly contribute to reported mood benefits of microdosing: sleep disturbances are common in mood disorders, and sleep (including REM) modulates synaptic plasticity, which is implicated in depression. They contrast these effects with conventional antidepressants such as selective serotonin reuptake inhibitors, which often suppress REM and can worsen sleep continuity, suggesting that microdosed LSD might have distinct sleep-related mechanisms if used therapeutically. The authors acknowledge several limitations. Using consumer-grade wearables enabled large-scale, naturalistic data collection (3,231 nights) but constrained measurement: devices do not reliably provide clinical metrics such as sleep or REM onset latency, and their proprietary algorithms are a “black box”. Although meta-analytic comparisons suggest Fitbits are reasonably accurate for sleep duration and staging, they remain inferior to polysomnography. External validity is also limited because the cohort comprised only healthy male participants with relatively homogeneous ethnicity and no mental health or sleep disorders. The study did not include standardised subjective sleep-quality scales, and sleep analyses were exploratory because no pre-specified sleep hypotheses were declared. Finally, the authors note that prior controlled microdosing trials have shown inconsistent physiological effects, though central nervous system penetration of microdoses has been demonstrated with EEG and fMRI in other work. Given these findings and caveats, the investigators argue that objective sleep monitoring should be incorporated into upcoming Phase II trials of LSD microdosing in patients with major depressive disorder to explore clinical relevance and mechanisms further.

The observed increase in total sleep following LSD microdosing provides a compelling rationale to include wearable sleep monitoring in forthcoming Phase II trials of LSD microdosing in major depressive disorder, as the sleep modifications could have clinical and mechanistic implications for potential antidepressant effects.