Naturalistic psychedelic therapy: The role of relaxation and subjective drug effects in antidepressant response

Psychedelic-assisted therapy (PAT) with psilocybin and LSD has shown promise for depression, anxiety and other neuropsychiatric disorders, and Switzerland permits PAT under a limited medical use programme. Calder and colleagues note that data from this programme provide a naturalistic, real-world complement to controlled trials because patients vary in diagnoses, concomitant medications, dosing schedules and number of sessions. The authors highlight ongoing debate about which acute subjective drug effects underpin therapeutic benefit: mystical-type experiences measured retrospectively (e.g. with the Mystical Experience Questionnaire, MEQ), overall intensity of the experience, ego dissolution, or the valence of acute effects (pleasant versus challenging). They also note that subjective effects are usually measured retrospectively but could be assessed in real time, which may be valuable clinically. This study set out to characterise acute subjective effects during PAT in Swiss clinical practice, to compare those effects with data from a randomized double-blind crossover trial in healthy volunteers receiving the same drug formulations, and to examine relationships between acute subjective effects and rapid antidepressant response in patients. Specific aims included analysing dose–response relations for LSD and psilocybin in patients, comparing patient and healthy-subject subjective profiles, assessing how hourly (real-time) ratings relate to retrospective MEQ scores, and testing whether particular acute effects predict changes in depressive symptoms as measured by the Montgomery–Åsberg Depression Rating Scale (MADRS).

This was a mixed naturalistic/experimental comparison using two data sources. Patient data came from PAT administered within the Swiss limited medical use programme at the Freiburg Network for Mental Health. Each PAT round included two preparatory consultations, one dosing session and 1–3 post-dose integration sessions; patients could repeat sessions as needed (minimum 3 months between sessions). Most patients continued concurrent long-term psychotherapy. The patient sample consisted of 28 individuals (50% female, mean age 49 years) with treatment-resistant conditions (86% had treatment-resistant major depressive disorder); 78.6% were taking psychiatric medications. Exclusion criteria included personal or family history of psychosis, bipolar I disorder, borderline personality disorder, epilepsy, dementia and pregnancy. All patients provided written informed consent. Healthy participants were drawn from a double-blind, placebo-controlled crossover trial that administered two doses each of LSD and psilocybin across five dosing days, with randomised order and ≥10 days between doses. The healthy sample included 28 participants (50% women, mean age 35 years), with no major psychiatric disorders and exclusion for recent or heavy illicit drug use; 50% had prior psychedelic experience. Procedures for healthy volunteers included at least hourly assessments of subjective effects during the acute period and discharge 24 h after dosing. Drug formulations were GMP-grade LSD base (>99% purity) in ethanol solution and psilocybin (99.7% purity) in capsules (5 mg per capsule). In the clinical sample the standard initial doses were 100 µg LSD or 15 mg psilocybin, adjustable based on response; total single-session doses ranged 100–250 µg LSD (median 100 µg) and 10–25 mg psilocybin (median 20 mg). Patients typically received treatment between 08:00 and 10:00, were encouraged to relax with eyeshades and music, and the treating psychiatrist was present and recorded hourly verbal ratings. Healthy participants received dosing at 09:00 in a hospital room with an investigator present. Primary assessments included hourly real-time ratings of acute effects. Patients rated five items verbally on a 0–10 scale (overall intensity, good effects, bad effects, ego dissolution, relaxation), converted to 0–100 for comparability; healthy participants used 0–100 visual analogue scales for the same items except relaxation. The MEQ30 (German version) was completed retrospectively after sessions (MEQ available for all healthy sessions and 41/55 PAT sessions). MADRS was used to assess depressive symptoms in patients at baseline and immediately after dosing; a subset provided follow-up MADRS about 7.7 days after PAT. Adverse effects were assessed with a preliminary Swiss Psychedelic Side Effects Inventory (SPSI) for 31/55 PAT sessions, recording symptom presence, duration, relatedness, valence and impact. Statistical analyses used R with linear mixed-effects models (lme4), including subject ID as a random effect to account for repeated measures. Area under the curve (AUC) for the first 8 h post-dose was calculated for real-time ratings. Psilocybin doses were converted to LSD-equivalent doses (1 mg psilocybin = 5 µg LSD) for comparisons; the Benjamini–Hochberg procedure controlled the false discovery rate. Where second doses or sedatives were given 1–3 h after the initial dose, timepoints after the second dose were treated as missing and mean imputation by group/drug/dose was used to enable AUC calculation.

Dose–response and patient–control comparisons: In patients, real-time ratings of "any drug effect" (AUC first 8 h) increased significantly with dose (F(5,30)=6.884, p<0.001). For direct comparisons, analyses used sessions with 100 µg LSD, 200 µg LSD or 15 mg psilocybin and adjusted for age (patients were older than healthy controls). Patients reported significantly lower ego dissolution than healthy volunteers (β = -183.50, SE = 36.01, p < 0.001). Ratings of bad drug effects were numerically higher in patients but did not reach statistical significance (β = 53.70, SE = 21.80, p = 0.09). No significant group differences were found for overall drug effect (β = -57.23, SE = 34.08, p = 0.30) or good drug effect (β = -98.61, SE = 46.59, p = 0.16). Age was positively associated with ego dissolution (β = 4.31, SE = 1.37, p = 0.02). Mystical experiences and relation to real-time ratings: "Complete" mystical experiences (MEQ30 criteria) were rare: 8 of 112 dosing sessions in healthy participants met the criteria, and none of the PAT sessions did. There were no statistical differences between groups in rates of complete mystical experiences (p = 0.99) or overall MEQ scores (β = -5.87, SE = 6.40, p = 0.58). Across samples, real-time ratings explained a substantial portion of variance in subsequent MEQ scores (marginal R2 = 0.55, conditional R2 = 0.67), and of the real-time items only ego dissolution significantly predicted MEQ scores (β = 0.11, SE = 0.03, p < 0.001). Antidepressant effects: In the subset with treatment-resistant depression, baseline MADRS averaged 20.9 (moderately severe). For 41 PAT sessions in 21 patients with baseline and immediate post-PAT MADRS, the mean decrease immediately after PAT was 6.2 points (29% reduction), a statistically significant improvement (β = 7.44, SE = 1.42, p < 0.001). Improvement remained at the follow-up assessment (~7.7 days), with a significant effect (β = 7.74, SE = 1.64, p < 0.001). There was no dose–response relationship for antidepressant effect in the analysed dose range (immediate: β = -0.01, SE = 0.02, p = 0.83; follow-up: β = 0.04, SE = 0.05, p = 0.58), and after controlling for dose there was no difference in efficacy between LSD and psilocybin immediately or at follow-up. Predictors of MADRS change: In complete-case analyses, the five hourly real-time ratings together explained 28% of the variance in immediate MADRS improvement (marginal R2 = 0.28, conditional R2 = 0.35). Of those ratings, only relaxation predicted greater MADRS reduction (β = -0.04, SE = 0.01, p = 0.03). By contrast, MEQ scores explained only 4% of variance in MADRS change (R2m = 0.04) and were not significant. None of the subjective effects predicted MADRS improvement at the later follow-up. Medication effects and adverse events: Exploratory analyses controlling for dose found no significant differences in acute subjective effects between patients taking psychiatric medications and those not taking them; the authors note limited power and high polypharmacy. Adverse-effect data (SPSI) were available for 31 sessions: 93.5% of patients experienced at least one side effect. The most common were mild headache (32.3%), fatigue (29%) and nausea (29%). Most side effects were reported as unpleasant but transient and resolved within 48 h in all but one case. One patient experienced several days of anxiety, derealisation, non-psychotic hallucinations and passive suicidal ideation after a challenging 15 mg psilocybin session for OCD; symptoms resolved after 6 days with intensive outpatient psychotherapy and no medication changes, and no long-term adverse effects were evident at 18 months.

Calder and colleagues interpret these naturalistic data as indicating that PAT with LSD and psilocybin can produce rapid antidepressant effects in a heterogeneous, treatment-resistant clinical population and appears generally safe when delivered with supportive psychotherapy. Patients showed broadly similar overall drug-effect and MEQ profiles to healthy volunteers, but displayed significantly lower ego dissolution and a non-significant tendency toward more challenging (bad) drug effects. The authors suggest several potential explanations for reduced ego dissolution in patients, including residual effects of psychiatric medications, greater anxiety or confrontation with difficult psychological material in a clinical setting, and personality factors such as higher neuroticism. They also note an unexpected positive association between age and ego dissolution and recommend further investigation across broader age ranges. Regarding mechanisms, the study did not find that retrospective mystical-type experiences (MEQ) predicted antidepressant response in this sample; instead, simple real-time measures collected hourly during sessions explained more variance in MEQ scores and in immediate clinical improvement. Relaxation during the session emerged as the strongest real-time predictor of immediate MADRS reduction, leading the authors to highlight relaxation as a potential therapeutic process variable. They propose that relaxation may reflect feelings of safety, capacity for "letting go", or effective coping that enables engagement with therapeutic material, and suggest that teaching relaxation techniques in preparatory sessions could be helpful. The authors acknowledge important limitations: the naturalistic dataset is heterogeneous with variable doses, diagnoses, concomitant medications and session frequency; the patient arm was open-label, non-randomised and uncontrolled and all patients received concurrent psychotherapy, limiting causal inference. Sample sizes were modest for several analyses, some data (MEQ, follow-up MADRS) were incomplete, follow-up intervals were short, and flexible dosing complicates dose–response interpretation. Differences in clinical setting between the patient and healthy samples may also have influenced subjective effects. Finally, the high rate of polypharmacy limited the ability to test effects of individual medication classes. Given these caveats, Calder et al. recommend cautious interpretation and call for larger, controlled studies to clarify the roles of relaxation, mystical-type experiences, ego dissolution and concomitant medications in PAT outcomes.

In conclusion, the study suggests that PAT with LSD and psilocybin, delivered within a real-world Swiss medical programme alongside psychotherapy, was associated with rapid reductions in depressive symptoms and an acceptable short-term safety profile in a heterogeneous, treatment-resistant cohort. Patients displayed a somewhat different subjective-effect profile from healthy volunteers, notably less ego dissolution. Hourly, real-time ratings—particularly relaxation—were better predictors of immediate antidepressant response than retrospective MEQ measures, and the authors recommend further research into the therapeutic importance of relaxation during PAT.