Effects of varied doses of psilocybin on time interval reproduction in human subjects

Awareness of temporal order and the ability to discriminate durations are features of normal wakeful consciousness, while altered states of consciousness (ASC) commonly produce changes in subjective time experience. Earlier research has reported that psilocybin, a serotonergic hallucinogen acting primarily at 5-HT2A/1A receptors, produces subjective and objective alterations in time perception: shortened reproductions of intervals longer than about 2.5 s, impaired synchronization to slower inter-beat intervals, slowed preferred tapping rates, working-memory deficits, and reported disturbances in subjective time sense. These findings align with evidence that serotonergic modulation of prefrontal cortex—an area implicated in timing behaviour—can alter timing performance. This study applies a parametric modelling approach to duration-reproduction data, using the dual klepsydra model (DKM) of internal time representation. Rather than analysing reproduction errors separately for each stimulus duration, the investigators fit the DKM to whole-response functions and focused on its single key parameter κ, the loss-rate coefficient of the model's leaky accumulators. The paper reports two double-blind, placebo-controlled experiments aimed at testing whether psilocybin alters κ during the peak drug effect and thereby affects the internal representation of supra-second time intervals.

Two double-blind, placebo-controlled experiments with within-subjects designs were conducted. Experiment 1 included N = 12 healthy volunteers (six men, six women, mean age 26.8 years) who completed three sessions separated by at least 14 days; conditions were placebo (P), a medium dose (M: reported in the extracted text as 115 g/kg body weight), and a high dose (H: reported as 250 g/kg body weight), administered in counterbalanced order. Performance was assessed at baseline (T = 0), at T = 90 min to target peak drug effects, and at T = 240 min. Experiment 2 comprised N = 9 participants (five men, four women, mean age 48.2 years) who completed two sessions (placebo and a very low dose, VL: reported as 12 g/kg body weight) separated by at least 14 days; testing occurred at T = 90 min. The extracted text also reports that psilocybin was given in gelatin capsules (1 mg and 5 mg) and that psilocybin is rapidly converted to the active metabolite psilocin; however, the dosing information in the extraction contains inconsistent units and the text does not clearly resolve this discrepancy. The behavioural task in both experiments was a duration reproduction task (DRT). Subjects heard a 500 Hz tone for a presented duration s, then after a constant silent delay w = 2 s the same tone was played and subjects terminated it when they judged the second tone to equal the first. Stimulus durations were s = 1.5, 2, 2.5, 4, 4.5, and 5 s, presented in random order with eight repetitions per duration. Response times r were recorded with 1 ms precision and averaged across repetitions, producing six stimulus–response pairs per subject per session. Each dataset was fit with the DKM; κ (the loss-rate coefficient) was estimated by weighted least squares (WLSQ) using a weighting function g(s) = s^-1 to compensate for increasing variance with longer stimuli. Data handling and statistical analysis: Some DRT datasets were missing in Experiment 1 (five of 108) due to a computer failure and nine fits produced κ = 0 (interpreted as poor model fit); non-zero κ estimates ranged from 0.0017 to 0.056 s^-1 (median ≈ 0.015 s^-1). In Experiment 2 all κ estimates were positive and ranged from 0.0013 to 0.04 s^-1 (median ≈ 0.02 s^-1); one subject was post hoc excluded (the extracted text does not clearly state the reason). To remove inter-individual baseline differences, κ values were converted to intra-individual ratios q = κ_T/κ_0 for Experiment 1 (so q_0 = 1) and to ratios versus placebo for Experiment 2; missing κ values were replaced conservatively (geometric mean of available values or replication of a single known value when two were missing). Ratio measures were log-transformed to symmetrise distributions and tested against zero with one-sample, two-tailed t-tests. The sessions also included psychometric scales (Altered States of Consciousness Rating Scale and the Adjective Mood Rating Scale), which are reported for context in the extracted text.

Model fits and baseline κ estimates: Across datasets the DKM produced κ estimates consistent with prior work. In Experiment 1, after excluding zero-fit cases, κ ranged from 0.0017 to 0.056 s^-1 (median ≈ 0.015 s^-1). Experiment 2 κ estimates ranged from 0.0013 to 0.04 s^-1 (median ≈ 0.02 s^-1). Experiment 1 (N = 12): At the anticipated peak drug time (T = 90 min), the medium dose produced a modest increase in κ of approximately 16%, while the high dose produced a significant increase in κ of approximately 44% relative to baseline (high dose, P < 0.02). No significant drug effects on κ were observed at T = 240 min, except for a marginal decrease of κ under placebo (P ≈ 0.05), which the authors note might reflect practice effects leading to more precise reproductions. Experiment 2 (N = 9): At T = 90 min the very low dose yielded a significant increase in κ of approximately 83% compared to placebo, despite the fact that most volunteers reported little subjective difference between placebo and this low dose on the psychometric scales; the only reported psychometric change was an increase on the AMRS 'introversion' scale. Analytical notes and caveats: The investigators emphasised that the effect-size measures used in the two experiments were defined differently and therefore are not directly comparable. They also noted substantial age differences between experimental samples (mean age 26.8 years in Experiment 1 versus 48.2 years in Experiment 2) and discussed that age-related declines in 5-HT2A receptor expression could influence sensitivity to psilocybin, complicating cross-experiment comparisons. Missing and zero-fit datasets were treated conservatively in a manner likely to bias group means toward the null, so the reported group effects may be underestimated.

Using the dual klepsydra model to summarise reproduction behaviour, the investigators conclude that psilocybin increases the DKM loss-rate parameter κ, indicating a faster decay of the internal representation of supra-second durations. The effect was dose-dependent in Experiment 1 (significant at the higher dose) and was also observed in Experiment 2 at a very low dose relative to placebo, despite minimal self-reported subjective change. From a modelling perspective, the authors view κ as a sensitive parameter for detecting psychoactive effects on internal time representation in the supra-second range. The authors relate their findings to earlier behavioural reports of subjective shortening of remembered durations and to neuropharmacological evidence linking psilocybin's serotonergic action—and consequent modulation of prefrontal cortex activity—to timing processes. Nevertheless, they emphasise that the neural substrate of the hypothetical DKM accumulators and their sensitivity to environmental or organismic factors remain unclear; thus mechanistic interpretation is tentative. Age-related differences in 5-HT2A receptor density and methodological differences between the two experiments prevent straightforward comparison of effect sizes across studies. The authors acknowledge limitations including missing datasets and poor model fits in some cases, conservative missing-value replacement that biases toward the null, inconsistent or unclear reporting of dose units in the extracted text, and the lack of direct evidence tying κ changes to specific neural mechanisms. Finally, the investigators suggest that an increased loss rate together with increased input noise might account for extreme subjective phenomena in altered states—such as the experience of ‘‘time standing still’’ or a loss of temporal order—by effectively collapsing the reliable horizons of duration experience. They propose that the DKM framework and κ may help to quantify and better understand qualitative alterations of time perception reported in ASCs and certain psychopathologies.