Tripping on context: Characteristics and predictors of placebo and nocebo psychedelic effects

Psychedelic compounds are being studied both as potential treatments for mental health conditions and as tools for understanding consciousness, but the role of extra-pharmacological factors such as mindset and environment remains difficult to isolate. Earlier research has recognised that context shapes psychedelic experiences, yet it has rarely been tested as a causal factor in its own right. In particular, it has been unclear whether context alone can produce placebo-like psychedelic experiences and nocebo-like side effects under controlled conditions, and whether individual traits linked to psychedelic responding also predict such effects. Stein and colleagues set out to test whether a carefully constructed psychedelic context, without active drug action, could induce psychedelic effects and side effects. They also aimed to examine whether trait variables previously associated with psychedelic responding, including responsiveness to verbal suggestion, dissociative traits, somatosensory amplification, childhood trauma, and compliance, would predict the magnitude of these placebo and nocebo effects. The study was pre-registered and used a placebo model based on inhaled medical air presented as nitrous oxide.

The researchers conducted a pre-registered, counterbalanced repeated-measures placebo-controlled experiment approved by King’s College London. Seventy-eight non-clinical participants from the greater London area were recruited through institutional sites, flyers, social media, and word of mouth. Participants were screened by telephone for medical and psychiatric eligibility, prior nitrous oxide use, nausea proneness, and adverse reactions, and any uncertainty was resolved with a psychiatrist. The sample was predominantly female, aged 18 to 44 years, and ethnically and nationally diverse. Before the experiment, participants completed baseline trait measures. These included responsiveness to verbal suggestion, assessed with the Brief Suggestibility Scale; childhood trauma, using the Adverse Childhood Experiences checklist; dissociative absorption, depersonalisation/derealisation, and amnesia, using the Dissociative Experiences Scale; somatosensory amplification; and compliance tendencies. The authors note that one pre-registered measure, the Acquiescence Response Set, was dropped because of poor internal consistency. Each participant completed two 20-minute inhalation conditions in counterbalanced order. In the placebo condition, medical air was presented as Entonox, described as a mixture of N₂O and oxygen expected to produce psychedelic effects and minor side effects. In the control condition, medical air was correctly identified as such. The placebo manipulation was embedded in a broader psychedelic-style experimental context, including screening, consent, instructions, visible gas-cylinder labelling, physiological monitoring, and scripted verbal suggestions about effects and side effects. Each inhalation period included an adjustment phase, a temporal reproduction task, and a four-minute resting-state epoch, followed by a 20-minute washout period. During the resting state, participants completed outcome measures referring to that epoch. State outcomes were measured with the 5D-ASC for altered states of consciousness, the CADSS for dissociation, the Ego-Dissolution Inventory, and the General Assessment of Side Effects. Blinding was assessed with a treatment-guess questionnaire. Temporal perception was tested using a temporal reproduction task across three blocks of 30 trials per condition. Analytically, paired-samples permutation t-tests compared placebo and control conditions, with Hedges’ g and bootstrap confidence intervals as effect sizes. Order effects were examined with permutation ANOVAs. Temporal reproduction was analysed using individual-level regression slopes and repeated-measures ANOVA. Trait predictors were explored using Spearman correlations and regularised regression models, with model selection based on cross-validation. False discovery rate correction was applied to the relevant statistical tests. Missing data were imputed where appropriate, and some self-report and timing data were excluded because of missingness or technical issues.

The placebo condition produced higher scores than the control condition on all phenomenological state measures. The largest effect was for depersonalisation on the CADSS, with a large effect size. Moderate effects were reported for altered states of consciousness on the 5D-ASC, dissociation subscales, ego dissolution, and side effect severity on the GASE. A weaker effect was observed for dread of ego dissolution. Overall, the authors describe the placebo and nocebo effects as generally moderate in magnitude. Order moderated some outcomes. Significant condition-by-order interactions were found for four of the 12 measures: the Ego-Dissolution Inventory, overall CADSS, and the depersonalisation and derealisation subscales. In these cases, placebo effects were significant and larger when the placebo condition came first, whereas effects were small and non-significant when the control condition came first. The authors interpret this as suggesting attenuation with greater familiarity with medical air inhalation. They also report that just under half of participants remained blind to the placebo manipulation, and that blinding status was mainly associated with whether participants experienced perceptual effects and side effects. The placebo condition did not significantly affect temporal reproduction performance. The extracted text does not report a clear placebo effect on this task, indicating that the suggested context altered subjective experience but not this measure of time perception. Trait-state analyses showed that difference scores for the phenomenological outcomes were moderately to strongly intercorrelated, although side effects were only related to altered states of consciousness and its subscales. Trait dissociation and compliance were the only baseline measures consistently correlated with placebo effects, and these correlations were generally weak. In the regularised regression models, responsiveness to verbal suggestion emerged as the strongest predictor of altered states of consciousness, while absorption and compliance were also retained as positive but weak predictors. Similar patterns were observed for some 5D-ASC subscales. Models for placebo-induced dissociation and ego dissolution explained only modest in-sample variance and performed poorly in cross-validation. The model for side effects also explained only a modest amount of variance. Across models, compliance and trait dissociation were repeatedly retained, but effects were small, and there was limited evidence for reliable trait predictors of nocebo effects.

The authors conclude that a psychedelic context alone, without pharmacological action, can reliably elicit both psychedelic-like experiences and side effects on standardised measures. They interpret this as evidence that context is not merely a background moderator but can act as a causal contributor to psychedelic phenomena. They also argue that the moderate nocebo effects they observed show that side effects in psychedelic research are not solely drug-driven but are partly shaped by contextual features of the experimental or therapeutic setting. In relation to earlier research, the authors present their findings as consistent with broader placebo and nocebo literature and with previous work suggesting that suggestions and context can generate anomalous experiences resembling psychedelic states. They note that the largest placebo effects were seen for depersonalisation and derealisation, followed by ego dissolution and altered states of consciousness, and suggest that these effects may reflect the content of the verbal suggestions and broader cultural associations with nitrous oxide. By contrast, the lack of an effect on temporal reproduction suggests that the manipulation influenced suggested phenomenology rather than unrelated cognitive domains. The authors argue that several aspects of the design may have reduced observed effect sizes. They suggest that the cross-over, repeated-measures design may have made the effect smaller at the aggregate level, especially when participants experienced the control condition first. They also propose that reduced blinding may have attenuated effects, and that the side-effect measure may have underestimated nocebo responses because it included symptoms that were not always plausible or relevant in the experimental setting. On individual differences, they report that trait responsiveness to verbal suggestion predicted placebo psychedelic effects on altered states of consciousness, while absorption and dissociative traits were more consistent predictors across outcomes. Compliance also predicted dissociative responses, raising the possibility that some self-reported psychedelic effects may be inflated by compliant responding. They view these findings as supporting the idea that responsiveness to context and suggestion partly explains variability in psychedelic outcomes. The authors also emphasise methodological implications. They identify order effects, washout adequacy, and the need to model order statistically as important considerations for future studies. They further note that blinding guesses were influenced by whether participants experienced perceptual effects and side effects. The main limitations they acknowledge are the complexity of the placebo induction procedure, which may limit generalisability to more common routes of administration such as oral dosing, possible limited power for subgroup analyses and blinding predictors, and the heavy reliance on self-report outcomes. They suggest that future work should more directly examine context, suggestion, and nocebo processes in psychedelic research, both for experimental validity and for clinical interpretation.

The authors conclude that psychedelic context alone can produce both psychedelic effects and side effects in the absence of an active drug. They state that individual difference factors linked to psychedelic responding explained only a modest amount of variance but did predict some placebo-psychedelic outcomes. On this basis, they argue that context should be treated not only as a moderator of psychedelic experiences but as a causal influence that can be studied independently of pharmacology.