The sample demographics for the different conditions are presented in Table 1. Participants in the different groups did not differ in gender distributions, χ23 = 2.29, p = .52, Cramer’s V = .22, or age, F3,47 = .29, p = .83, η2 = .02. Participants completed the temporal reproduction task at variable time points post-dosing (M = 170 min; CIs 156, 183) (Fig. 1). Nevertheless, there were no significant differences in task completion times (time post-dose at which the task was completed) across drugs (placebo: M = 171, SD = 40; LSD: M = 172, SD = 50), F1,24.26 < .01, p = .95, η2s < .01, or doses (placebo: M = 171, SD = 40; LSD 5 μg: M = 175, SD = 58; LSD 10 μg: M = 158, SD = 46; LSD 20 μg: M = 184, SD = 47), F1,22.91 = .58, p = .64, η2s = .03. These results suggest that the task was completed at a relatively uniform post-dose time across conditions.
Subjective drug effects (SDEVAS)
We analysed five self-report measures concerning subjective changes in cognition and perception (see Fig. 1) with two sets of mixed model ANOVAs with (post-dose) time as a repeated measures independent variable and drug or dose as between-group independent variables, respectively. In addition to two participants who were excluded because of outlying performance in the temporal reproduction task, six participants were excluded because of missing data at one or more time points. Although there were numerical tendencies for larger reported drug effects in the LSD condition, there were no significant main effects of drug, Fs < 1.38, ps > .25, ηp2s < .04, or time, Fs < 2.26, ps > .06, ηp2s < .06, or time × drug interactions, Fs < 2.4, ps > .05, ηp2s < .07, on any of the measures. Similarly, there were no significant main effects of dose, Fs < 1.34, ps > .28, ηp2 s < .10, or time, Fs < 3.3, ps > .05, ηp2s < .22, except on drug effects, F10,360 = 4.91, p < .001, ηp2 = .12, which reflected a tendency for reports to increase from baseline to 120–180 min and subsequently decline again, and no significant time × dose interactions, Fs < 1.5, ps > .18, ηp2s < .12. Insofar as there appeared to be numerical differences across conditions in SDEVAS scores that overlapped with the task completion time, particularly for drug effects and feeling high, we performed exploratory condition contrasts (α < .01) on SDEVAS ratings at 180 min post-dosage. There was a suggestive main effect of drug on drug effects, F1,38 = 5.18, p = .030, η2 = .04, but no significant effects for any of the other SDEVAS scores, Fs < 3.6, ps > .06, η2s < .04. Similarly, there were no significant effects of dose on any of the SDEVAS scores, Fs < 2.3, ps > .12, η2s < .07. Cumulatively, these results suggest that microdose LSD did not produce robust changes in a range of conscious states, as indexed by self-report measures.
Reproduction times in the temporal reproduction task are presented in Fig. 2 as a function of drug. There was a main effect of interval, F8,352 = 411.63, p < .001, ηp2 = .90, reflecting longer reproduction times as the stimulus interval increased. A main effect of drug was also observed, F1,44 = 6.18, p = .017, ηp2 = .12, with participants displaying longer reproduction times in the LSD, relative to the placebo, condition. These effects were further mediated by a drug × interval interaction, F8,352 = 5.13, p = .003, ηp2 = .10. Subsidiary analyses (α < .01) revealed drug effects on reproduction times at three intervals, with relatively uniform effect sizes (~ .16): 2800 ms: F1,45 = 7.53, p = .009, η2 = .15; 3200 ms: F1,45 = 8.34, p = .006, η2 = .16; and 3600 ms: F1,45 = 8.47, p = .006, η2 = .16. Main effects of drug were also suggestive (α < .05) for the 2000, 2400, and 4000 ms intervals with comparable effect sizes, Fs < 5.38, ps < .04, η2s > .11, whereas they were reliably non-significant, with negligible effect sizes, for the 800–1600 ms intervals, Fs < .59, ps > .70, η2s < .01.
In order to determine whether temporal over-reproduction in the LSD condition was independent of potential differential self-reported drug effects across conditions, we next performed a hierarchical regression on reproduction times in each interval separately with drug effect ratings at 180 min post-dosage included in the first step as a nuisance variable and drug condition in the second step. Drug effects did not significantly predict reproduction times for any of the stimulus intervals in the first step, R2s < .06, ps > .17, although there was a weak trend toward significance in the 2800 ms interval condition, R2 = .09, p = .064. In the second step, drug condition did not improve the models for 800–1600 ms intervals, ∆R2s < .06, ps > .18, but reliably significantly improved the models for all other intervals (2000–4000 ms), ∆R2 range .11–.20, p range .025–.002. In all cases, the LSD condition was positively associated with reproduction times. These results corroborate the foregoing ANOVAs and indicate that differences in temporal reproduction across drug conditions are independent of potential differential subjective drug effects across conditions.
To further clarify whether LSD was associated with steeper reproduction slopes, we performed within-participant regression analyses on participants’ individual reproduction data with stimulus interval as a predictor and subsequently contrasted beta coefficients across conditions. The analyses revealed a main effect of drug, F1,44 = 8.53, p = .005, η2 = .16, reflecting steeper slopes in the LSD condition, which further corroborates the tendency for temporal over-production to be more pronounced selectively at longer stimulus intervals (Fig. 2).
Reproduction variability data, as indexed by coefficients of variation (CVs), are presented in Fig. 2. There was a main effect of interval, F8,352 = 11.41, p < .001, ηp2 = .21, reflecting a reduction in variability as the stimulus interval increased; however, there were neither significant main effects of drug, F1,44 = 0.70, p = .41, ηp2 = .02, nor a drug × interval interaction, F8,352 = .47, p = .81, ηp2 = .01. This suggests that LSD is not significantly associated with atypical variability in temporal reproduction.
Analyses of reproduction times as a function of dose (Fig. 2) revealed a main effect of interval, F8,336 = 584.017, p < .001, ηp2 = .93, as previously observed, and a main effect of dose, F3,42 = 3.22, p = .032, ηp2 = .19. These effects were further mediated by a dose × interval interaction, F24,336 = 2.15, p = .037, ηp2 = .13. Subsidiary analyses did not reveal effects of dose on reproduction times at specific intervals (α < .01), although there were suggestive effects (α < .05) with uniformly strong effect sizes (~ .21) at 2000 ms, F3,45 = 3.74, p = .018, η2 = .21, 3200 ms, F3,45 = 3.67, p = .019, η2 = .21, and 3600 ms, F3,45 = 3.5, p = .023, η2 = .20. In each case, reproduction times were longer in the 10 μg dose condition than in the placebo condition, post hoc Tukey’s p range .01–.029. All other dose effects were non-significant, Fs < 3.17, ps > .058, η2s < .18.
Although self-reported drug effects did not significantly differ across dose conditions at 180 min post-dosage, as in the drug-specific analyses, we performed hierarchical regressions to assess whether the observed dose-specific differences in temporal reproduction were independent of self-reported drug effects. As described above, in the first step, drug effects did not significantly predict reproduction times in any of the stimulus intervals. In the second step, dose did not improve the models for 800–1600 ms or 4000 ms intervals, ∆R2s < .07, ps > .10, but significantly improved the models for all other intervals (2000–3600 ms), ∆R2 range .11–.20, p range .040–.004. In all cases, an increased dose was positively associated with reproduction times. These results suggest a positive linear dosage effect on temporal reproduction that is independent of potential differential subjective drug effects across conditions.
A re-analysis of within-participant regression beta coefficients as a function of dose (Fig. 2) revealed a significant main effect of dose, F3,42 = 2.88, p = .047, η2 = .17. Post hoc Tukey’s tests revealed suggestively larger coefficients in the 20 μg relative to the placebo condition, p = .055, but no other significant differences, ps > .10. This suggests a weak tendency for over-production in the 20 μg condition to be more pronounced for longer stimulus intervals.
The re-analysis of CVs as a function of dose (Fig. 2) replicated the main effect of interval, F8,336 = 16.66, p < .001, ηp2 = .28, but there was no significant effect of dose, F3,42 = .242, p = .87, ηp2 = .017, nor a dose × interval interaction, F24,336 = .87, p = .6, ηp2 = .06. These data corroborate the previous results and suggest that LSD does not significantly impact reproduction variability.