Effects of psilocybin on time perception and temporal control of behaviour in humans

Temporal processing is essential for perception and motor control, and evidence indicates that different neural mechanisms subserve distinct time ranges and tasks. Previous work has implicated dopaminergic fronto-striatal circuits in timing, and patients with lesions or disorders affecting these systems show impaired accuracy and increased variability in time estimation and motor timing. Hallucinogens such as psilocybin reliably alter the subjective experience of time, yet prior studies had not systematically examined objective, standardised measures of temporal processing under psilocybin in humans. Wittmann and colleagues designed a double-blind, placebo-controlled within-subject experiment to test dose-dependent effects of psilocybin (a mixed 5-HT2A/1A agonist) on multiple aspects of temporal processing. The study employed temporal reproduction, sensorimotor synchronization and tapping tasks—chosen to probe timing below and above an approximate 2–3 second boundary—as well as assessments of spatial working memory and subjective conscious state, with the hypothesis that psilocybin would selectively impair longer-interval timing because of its known effects on attention and working memory.

Twelve healthy, right-handed volunteers (six men, six women; mean age 26.8 years, SD 3.6) recruited in Zürich participated after medical and psychiatric screening and informed consent. Exclusion criteria included hearing or vision problems, major psychiatric disorder in self or first-degree relatives, and regular substance abuse. Six participants reported prior limited psilocybin use; seven had sporadic cannabis use. The Swiss authorities authorised psilocybin administration and the protocol received local ethics approval. A within-subject, counterbalanced design compared placebo, medium dose (MD; 115 µg/kg, mean absolute dose ≈8.2 mg) and high dose (HD; 250 µg/kg, mean absolute dose ≈17.6 mg) psilocybin on three separate days at least 14 days apart. Assessments occurred at baseline (t0), at peak drug effects (t1; about 90 min after administration) and at a post-peak time (t2; 240 min). Psychometric measures (AMRS at 0, 80 and 280 min; 5D-ASC at 110 min) and the spatial span (SSP) from CANTAB (at 0, 100 and 360 min) were included to index subjective state and spatial working memory. Temporal tasks comprised: (1) temporal reproduction of six standard intervals—three short (1500, 2000, 2500 ms) and three long (4000, 4500, 5000 ms)—each presented eight times in random order; participants reproduced intervals by key press after a fixed 2000 ms pause; (2) sensorimotor synchronization to regular tone sequences with inter-onset intervals of 700, 1000, 2000 and 4000 ms, recording mean asynchrony, standard deviation of asynchrony and percent of missed synchronizations (defined as taps occurring >120 ms after tone onset); and (3) tapping tasks measuring self-paced (personal) and maximum tapping speeds (median inter-tap interval and stability measured via coefficient of variance). Spatial working memory was assessed with the SSP (span length recorded). Subjective effects were measured using the 5D-ASC (five main factors including 'oceanic boundlessness' and an item cluster 'altered time sense') and the 60-item Adjective Mood Rating Scale (AMRS) with multiple subscales. Psilocybin and matched placebo capsules were administered orally; subjective effects typically emerged within 20–40 min and peaked around 60–90 min. Statistical analyses used repeated measures ANOVAs with factors of treatment (placebo, MD, HD), measurement time (t0, t1, t2) and, where applicable, interval or subscale. The primary interaction of interest was treatment × measurement time. Planned contrasts compared changes from t0 to t1 and t0 to t2 between drug and placebo (alpha adjusted for the number of contrasts per measure). Data were natural-log transformed to reduce non-normality (a constant added when necessary), and Greenhouse–Geisser corrections were applied when sphericity was violated. Pearson correlations pooled MD and HD data where significant at both doses; only correlations significant at both doses were considered robust.

Temporal reproduction: For short intervals (1.5–2.5 s) only the effect of interval duration was significant, with no treatment or time effects. For long intervals (4–5 s) there were significant main effects of interval and measurement time, and a significant treatment × measurement time interaction (F = 3.2, p = 0.025). Follow-up analyses showed that HD psilocybin produced a significant under-reproduction of long intervals at peak (t1) compared with baseline (contrast F = 13.0, p = 0.004), whereas MD and placebo did not show this time-dependent change. The underestimation had largely returned by t2 (no significant t2 vs t0 contrast). Analysis of reproduction variability found a main effect of measurement time for long intervals but no significant treatment × time interaction for stability measures. Sensorimotor synchronization: Mean asynchrony analyses indicated significant effects of measurement time (p = 0.001) and interval (p < 0.001), and a measurement time × treatment interaction (F = 3.9, p = 0.009). The interaction reflected that MD psilocybin produced a significant change from baseline to peak when compared with placebo (contrast t1 vs t0: p = 0.003); HD showed a marginal difference against placebo (p = 0.057). When examined within dose conditions, MD showed significant effects of both interval and measurement time (measurement-time contrast t1 vs t0: p = 0.002). For HD, interval, measurement time and their interaction reached significance (interval × measurement-time interaction p = 0.012), with a marginal t1 vs t0 contrast (p = 0.039). The authors note that smaller mean asynchronies under drug do not indicate improved synchronization but reflect increased reaction-time asynchronies (missed anticipations). The standard deviation of asynchrony (stability) was influenced strongly by interval length (larger variability for longer intervals) but showed no treatment or time effects. Percent missed synchronizations occurred mainly for the 2000 ms and 4000 ms intervals; there were significant main effects of treatment (p = 0.018) and interval (p = 0.002), but the treatment × measurement time interaction was not significant, although descriptively more missed synchronizations occurred at peak psilocybin. Tapping speed: Personal (self-paced) tapping tempo slowed under HD psilocybin at peak. Median inter-tap intervals increased to 949.0 ms (SD 390.0) at t1 versus 692.2 ms (SD 274.6) at baseline, with a significant contrast t1 vs t0 in the HD condition (p = 0.004); no significant change was observed for MD. A main effect of measurement time on tapping stability (coefficient of variance) was observed, but there was no treatment effect. Maximum tapping speed (~150 ms inter-tap intervals) showed main effects of measurement time (p = 0.001) and treatment (p = 0.026) but no significant measurement time × treatment interaction; tapping stability at maximum speed was unaffected by psilocybin, and the authors suggest the main treatment effect may reflect chance differences across sessions rather than a drug effect. Spatial span (working memory): A significant treatment × measurement time interaction was found for SSP span length (p = 0.003). Planned contrasts indicated that HD psilocybin impaired spatial span at peak compared with placebo (contrast p < 0.011); MD did not produce a significant decrement. Subjective measures: Psilocybin produced dose-dependent increases across 5D-ASC dimensions (treatment main effect F = 14.2, p < 0.001) with a treatment × subscale interaction (p = 0.016). All five subscales showed significant treatment effects at peak after correction; notable increases included 'oceanic boundlessness' (OB), anxious ego dissolution, visionary restructuralization and reduction of vigilance. The ASC item 'altered time sense' increased significantly (ANOVA p = 0.009; HD vs placebo contrast p = 0.017). AMRS subscales revealed multiple treatment × time interactions: significant effects were reported for inactivation, tiredness, dazed state, introversion, emotional excitability and dreaminess, among others, with contrasts often significant at t1 vs t0 for MD and/or HD. Correlations: No significant correlations were found between spatial span length and temporal reproduction underestimation for the long intervals (r between −0.17 and 0.16). A strong negative correlation emerged between spatial span length and the OB subscore (r = −0.73, p < 0.001), and regression analysis implicated the OB items 'depersonalization' (r = −0.83, p < 0.0001) and 'altered time sense' (r = −0.69, p < 0.001) as major contributors. Temporal-processing measures did not correlate significantly with 5D-ASC subscales, although there was a non-significant trend (r ≈ 0.4–0.6) linking greater underestimation of 4.5–5 s intervals with higher 'altered time sense' scores.

Wittmann and colleagues interpret their results as evidence that psilocybin alters objective timing performance in humans, but selectively so: impairments occurred for interval durations longer than approximately 2–3 seconds across temporal reproduction and sensorimotor synchronization tasks, and personal tapping tempo slowed at peak HD psilocybin. No consistent drug effects were found for shorter intervals or for maximum tapping speed. From these patterns the researchers infer that psilocybin's timing effects likely reflect disruption of cognitively mediated stages of temporal processing—attention, short-term/working memory or decision processes—rather than a basic alteration of an internal pacemaker or accumulator mechanism. The authors link their findings to prior literature showing a temporal-integration window near 2–3 seconds and to studies in which working-memory loading selectively impaired timing for intervals above 2 seconds. The concurrent impairment of spatial working memory under HD psilocybin supports the notion that cognitive processes contribute to the observed timing deficits, although the absence of significant correlations between the objective timing measures and working-memory performance is taken to reflect limited power and relatively small effect sizes in this small sample rather than disconfirming the hypothesis. Slowing of self-paced tapping but preserved maximum tapping is discussed in terms of separable neural control modes for voluntary (feedback) versus maximal (feed-forward) movement frequencies and is consistent with cortical rather than cerebellar or basal ganglia–only effects; the authors note prior imaging evidence of left-sided dorsolateral overactivation under psilocybin that related to depersonalization phenomena. Pharmacologically, the study indicates involvement of the serotonin 5-HT2A/1A system in duration processing in the seconds range, either directly or indirectly via interactions with dopaminergic and glutamatergic systems. The authors acknowledge limitations including small sample size, modest effect sizes and the inability to disentangle receptor-specific contributions; they recommend future work using receptor-selective antagonists, larger samples, concurrent probing of different working-memory domains (e.g. verbal vs spatial), and comparisons with clinical populations (for example, unmedicated acute schizophrenic patients) to clarify similarities and differences between psilocybin-induced effects and timing deficits observed in schizophrenia. The researchers also emphasise that the pattern of selective impairment argues against a non-specific loss of task engagement as an explanation.