Dreamlike effects of LSD on waking imagery in humans depend on serotonin 2A receptor activation

Psychedelic drugs such as lysergic acid diethylamide (LSD) produce altered states of consciousness that have long been described as dreamlike, sharing features with REM sleep dreaming including vivid imagery, altered thought processes and changes in self-experience. Previous work—behavioural, neurophysiological and pharmacological—has suggested overlaps between psychedelic states and dreaming, but the phenomenological heterogeneity of psychedelic imagery has made it difficult to demonstrate a stringent link using objective measures. Dreaming can be operationalised using formal indices such as cognitive bizarreness, which captures the improbable, illogical or fantastical aspects of narrative dream mentation and can be measured reliably in verbal reports. Kraehenmann and colleagues set out to test three linked hypotheses: that LSD produces dreamlike waking imagery measurable as increased cognitive bizarreness; that any such effect depends on activation of the serotonin 2A (5-HT2A) receptor; and that the degree of dreamlike imagery relates to concurrent subjective changes in perception, mood, cognition and sense of self. To address these aims they combined a standardised guided imagery task with quantitative scoring of bizarreness and with pharmacological blockade of 5-HT2A receptors using ketanserin in a double-blind, within-subject crossover design.

Twenty-five healthy adult volunteers were enrolled after psychiatric and medical screening; 9 participants (36%) reported prior classic psychedelic use and participants abstained from other drugs throughout the study. The design was a double-blind, placebo-controlled, within-subjects crossover with three sessions in balanced order and at least 2 weeks between sessions. Treatments were: placebo pre-treatment + placebo (Pla), placebo pre-treatment + LSD 100 mcg orally (LSD), and ketanserin 40 mg orally given 1 h before LSD 100 mcg (Ket+LSD). Ketanserin dosing and timing were chosen to maximise 5-HT2A receptor occupancy prior to LSD. LSD and ketanserin capsules were matched and administered in single oral doses. Sessions began in the morning with pre-treatment at 08:00 and treatment at 09:00. A standardised lunch and dinner were provided; the guided mental imagery task was performed ~7 h after LSD administration (at 16:00) during the descending phase of acute LSD effects to allow verbal reports. The mental imagery procedure was a 30 min standardised guided affective imagery protocol: participants lay with eyes closed and were prompted with an initial motif (meadow, creek or house across sessions) and given open-ended questions by a trained guide to facilitate ongoing imagery; imagery reports were audio-recorded and transcribed, yielding 75 reports (three per participant). Subjective state was measured with the five-dimensional Altered State of Consciousness questionnaire (5D-ASC). A 45-item short version was administered at multiple time points during the session and the short-version score closest to the imagery task was used to quantify contemporaneous subjective experience. Global intensity of drug effect (G-ASC) and five lower-order scales (elementary imagery, complex imagery, euphoric mood, loss of self-boundaries and cognitive control, changed meaning of percepts) were analysed. After imagery, participants rated aspects of their imagery on visual analogue scales (vividness, arousal, positive/negative emotions, insight, relaxation). Cognitive bizarreness was quantified using an adapted version of Hobson's two-stage scoring system for dream reports. Two blinded external judges scored a sample of reports to establish inter-rater reliability (intra-class correlation coefficient = 0.95) and then independently scored the remaining reports. Bizarre items were classified into main categories (plot, cognition, feeling state) and subcategories; bizarreness density (BD) was calculated by dividing number of bizarre items by transcript word count to correct for report length. Statistical analysis used repeated-measures ANOVAs for drug × scale/time/category comparisons, Pearson correlations for change–change relations (ΔBD, Δ5D-ASC, ΔVAS) and multiple linear regression to test which 5D-ASC changes predicted ΔBD. The extracted text contains minor inconsistencies in the reported minute values for some assessment time points (e.g., T3 appears as 390 min in one passage and 420 min in another); the authors state the imagery task occurred about 7 h after treatment and used the 5D-ASC short form completed at the time nearest the imagery task.

At the time of the imagery task LSD continued to produce marked subjective alterations. Repeated-measures ANOVA of the 5D-ASC short version showed a significant main effect of drug (p < 0.001) and a drug × scale interaction; post-hoc tests indicated higher scores on all five 5D-ASC scales in the LSD condition compared with Pla and Ket+LSD (all p < 0.05). No differences were observed between Pla and Ket+LSD, indicating that ketanserin pre-treatment blocked LSD effects on these scales. Global intensity (G-ASC) also showed significant drug and drug × time effects (all p < 0.001). G-ASC was greater in the LSD condition at all measured time points than in Pla and Ket+LSD (all p < 0.001). In the LSD condition, mean (SD) G-ASC at the imagery time and at peak were reported as 23.78% (24.25) and 37.95% (21.89), respectively, giving a relative intensity at imagery time of 62.6% of peak. Subjective ratings of the imagery experience (VAS) showed a significant main effect of drug (p < 0.001) and a drug × scale interaction. Post-hoc comparisons indicated greater vividness and arousal in the LSD condition than Pla, and greater vividness in LSD than Ket+LSD (all p < 0.05). Again, Pla and Ket+LSD did not differ. Cognitive bizarreness density (BD) was significantly increased under LSD versus Pla and Ket+LSD (main effect of drug p < 0.001; post-hoc LSD > Pla and LSD > Ket+LSD, all p < 0.001). BD did not differ between Pla and Ket+LSD for total BD. A drug × category ANOVA found that the increase in BD was concentrated in two of the three main categories (plot and cognition) while the feeling-state category did not differ across conditions. At the subcategory level, BD increases under LSD occurred in four of eight subcategories (all p < 0.05); ketanserin pre-treatment again blocked these effects. Mean word counts for reports were provided (approximately 4,000–4,350 words per condition), and inter-rater reliability for bizarreness scoring was high (ICC = 0.95). Correlational analyses showed a significant positive association between LSD-induced change in total BD and the LSD-induced change in the 5D-ASC loss of self-boundaries and cognitive control scale (r = 0.44, N = 25, p < 0.05 uncorrected). Trend-level relations were reported for euphoric mood (p = 0.05 uncorrected) and emotional arousal VAS (p = 0.08 uncorrected). A multiple linear regression indicated that the loss-of-self-boundaries and cognitive control scale was the only significant predictor of ΔBD in the full model, accounting for about 19% of variance (adjusted R2 = 0.19, F(1,23) = 5.50, p = 0.03).

Kraehenmann and colleagues interpret their findings as quantitative evidence that LSD produces dreamlike waking imagery in healthy volunteers, operationalised as increased cognitive bizarreness in guided imagery reports. They emphasise that the effect was pharmacologically specific: ketanserin pre-treatment completely blocked LSD-induced increases in cognitive bizarreness, subjective state changes on the 5D-ASC, and changes in imagery vividness and arousal, implicating stimulation of the 5-HT2A receptor as a key mechanism. The authors discuss several plausible proximal mechanisms. First, they propose that LSD-induced disturbance of the sense of self (depersonalisation) and impaired cognitive control can alter fundamental organising structures of consciousness (space, time, causality, selfhood) in ways that generate bizarre narrative elements; the observed correlation between BD and the loss-of-self-boundaries/cognitive-control scale supports this link. Second, changes in basic visual processing under LSD—activation of primary visual cortex—could facilitate formation of simple geometric hallucinations that propagate into complex scenes, but the authors argue their data do not support a purely bottom-up visual-noise account because BD was not associated with elementary or complex imagery scales nor with VAS vividness; instead, increases were concentrated in plot-related categories, implying preserved narrative structure akin to dreaming. Third, enhanced creative or divergent thinking is proposed as a contributor, with participants reporting less logical constraint and more dynamic recombination of memory fragments; discontinuity and incongruity subcategories accounted for a large portion of BD. Despite strong evidence implicating 5-HT2A receptors, the investigators note unresolved questions about the roles of other receptors affected by LSD, including dopamine D2/D1 and 5-HT1A, especially given arguments from animal and clinical data that dopaminergic and 5-HT1A mechanisms can influence dreaming and hallucinations. They therefore recommend future pharmacological studies using selective antagonists or non-hallucinogenic 5-HT2A agonists to disentangle receptor contributions. Key limitations acknowledged by the authors include the predominance of psychedelic-naïve subjects (which may affect ability to navigate imagery), the possibility that LSD effects on language production or attention could have influenced narrative reports rather than imagery per se, and the timing of the imagery task in the late/subacute phase of LSD effects when dopaminergic contributions might differ from earlier phases. The authors also speculate on clinical implications, suggesting that the psychedelic state combines lucid awareness with loss of self-boundaries and cognitive control and that this combination might facilitate psychological insight; they caution this is speculative and call for further research. Overall, the study’s interpretation remains cautious and anchored to the data: LSD increases dreamlike bizarreness in waking imagery via mechanisms dependent on 5-HT2A receptor stimulation, and this effect relates most strongly to disruptions in self-experience and cognitive control.

Using a standardised, formal measure of dreaming cognition in healthy subjects, the investigators found that LSD produced mental imagery comparable to dreams, primarily via activation of the 5-HT2A receptor. Dreamlike effects of LSD were related to loss of self-boundaries and cognitive control, but not to measures of simple visual hallucinations. The authors conclude that future studies should further assess the contributions of specific 5-HT and dopamine receptor subtypes to cognitive bizarreness.