Enhancing mindfulness and compassion through an ayahuasca-inspired formulation containing N,N-DMT and harmine: A randomized controlled trial in healthy subjects

Aicher and colleagues frame the study against evidence that mindfulness and compassion training improve psychological well-being but require time and practice that some patients with mood disorders cannot sustain. Previous studies report that serotonergic psychedelics and traditional ayahuasca can rapidly improve mood and increase mindfulness-related capacities and prosocial behaviour; however, the mechanisms by which ayahuasca's principal components—N,N-dimethyltryptamine (DMT) together with MAO-A inhibiting beta-carbolines such as harmine—affect mindfulness and compassion remain unclear. This randomised, double-blind, placebo-controlled within-subject trial tested whether an ayahuasca-inspired formulation combining intranasal DMT with sublingual/orodispersible harmine (DMT/HAR) acutely increases mindfulness (MINDSENS Composite Index, MCI) and compassion (Self and Others Compassion Scales, SOCS-S and SOCS-O) in healthy volunteers, compared with harmine alone (HAR) and placebo (PLA). The authors hypothesised greater increases in mindfulness and compassion after DMT/HAR than after HAR or PLA, and stronger effects in participants classified as high versus low sensitivity to the drug.

The trial used a within-subjects, double-blind, randomised, placebo-controlled design. Thirty-one healthy male participants completed three experimental sessions, each separated in time, in which they received in counterbalanced order: (1) DMT + harmine (DMT/HAR), (2) harmine + placebo intranasal spray (HAR), and (3) placebo + placebo (PLA). Six additional recruits dropped out and were replaced so that 31 participants (mean age 25.39 ± 4.21 years) completed all three arms. Inclusion criteria included age 20–40, BMI 18.5–30, male sex, absence of current or past somatic, neurological or psychiatric disorder, no family history of psychosis or bipolar disorder, no interacting medication, and no current or regular drug use; participants provided written informed consent and received monetary compensation. Interventions were designed to mimic key pharmacology of ayahuasca while avoiding oral administration. Harmine HCl was administered as a 100 mg orodispersible tablet for buccal delivery, 30 minutes before the intranasal dosing sequence. DMT was delivered as an intranasal spray in incremental doses totalling 100 mg across 10 administrations (10 mg every 15 minutes over 150 minutes). Matching placebos with comparable texture/taste were used. The setting was a single-participant, soundproof, temperature-controlled laboratory room with a relaxed living-room ambience and standardised background music; a medically trained experimenter supervised each session. Participants could skip individual DMT doses for tolerability; this occurred twice across the study. Primary outcome measures were the MINDSENS Composite Index (MCI; derived from selected items of the FFMQ and the Experiences Questionnaire) assessing mindfulness subdomains (observing, non-reacting, decentering), and the Sussex-Oxford Compassion Scales for others (SOCS-O) and self (SOCS-S). Assessments occurred at baseline, 24 hours post-experience, and at 1- and 4-month follow-ups. For analysis, the investigators used (generalized) linear mixed-effects models estimated by REML via the lme4 package in R. Participants were dichotomised into high- versus low-sensitivity groups by a median split on maximum subjective intensity ratings in the DMT/HAR condition. Fixed effects included drug condition (DMT/HAR, HAR, PLA), sensitivity group, and their interaction; models were fit on standardized data to yield standardized parameters. Type III repeated-measures ANOVA tables with Satterthwaite or Kenward-Roger degrees of freedom were used to test effects, post hoc contrasts were computed with the emmeans package, significance was set at 0.05, and p-values were corrected for multiple comparisons using the Benjamini-Hochberg method.

Sample flow and tolerability: Thirty-one male participants completed all three conditions; six recruits initially enrolled dropped out and were replaced. Only a few and small side effects were reported, and two dose skips occurred. The extracted text does not provide a detailed adverse-event table or physiological safety data within this section. MINDSENS Composite Index (MCI): The mixed model showed substantial total explanatory power (conditional R2 = 0.86) and a fixed-effects contribution (marginal R2 = 0.19). There were significant main effects of drug (F(2,56.19) = 3.41, p = 0.04) and sensitivity (F(1,29.08) = 6.54, p = 0.016), plus a significant drug × sensitivity interaction (F(2,56.19) = 7.87, p < 0.001). Post hoc contrasts indicated higher MCI after DMT/HAR versus placebo (mean difference = 0.14, SE = 0.05, df = 56.1, t = 2.58, p = 0.04). No significant difference emerged between DMT/HAR and HAR (p = 0.17) or between HAR and PLA (p = 0.35) in the pooled sample. When stratified by sensitivity, the high-sensitivity subgroup showed pronounced increases: DMT/HAR exceeded HAR (Mdiff = 0.28, SE = 0.07, p < 0.001) and PLA (Mdiff = 0.31, SE = 0.07, p < 0.001), whereas the low-sensitivity subgroup showed no significant condition differences. Self-compassion (SOCS-S): The model had conditional R2 = 0.90 and marginal R2 = 0.14. There were significant effects of drug (F(2,56.04) = 7.53, p = 0.001), sensitivity (F(1,28.96) = 4.21, p = 0.05), and their interaction (F(2,56.04) = 4.58, p = 0.014). Post hoc tests demonstrated that both DMT/HAR (Mdiff = 2.83, SE = 0.84, p < 0.01) and HAR (Mdiff = 2.82, SE = 0.85, p < 0.01) produced higher SOCS-S scores than PLA, with no significant difference between DMT/HAR and HAR (p = 0.99) in the overall sample. In the high-sensitivity group, DMT/HAR and HAR each exceeded PLA (DMT/HAR vs PLA: Mdiff = 5.19, SE = 1.18, p < 0.001; HAR vs PLA: Mdiff = 3.26, SE = 1.18, p = 0.01), while the low-sensitivity group showed no significant contrasts. Compassion for others (SOCS-O): The model yielded conditional R2 = 0.88 and marginal R2 = 0.13, with a main effect of drug (F(2,56.14) = 3.37, p = 0.042) and sensitivity (F(1,29.05) = 4.42, p = 0.04); the drug × sensitivity interaction was not significant (p = 0.20). Overall, DMT/HAR was higher than PLA (Mdiff = 2.28, SE = 0.88, p = 0.04). In subgroup analyses, the high-sensitivity group reported greater SOCS-O after DMT/HAR than after PLA (Mdiff = 3.69, SE = 1.24, p = 0.01); low-sensitivity participants again showed no significant condition differences. Subscale findings and temporal pattern: Across mindfulness subscales, non-reacting showed a strong drug effect, while observing and decentering increased in the high-sensitivity group. For SOCS subscales, recognizing suffering and feeling for the person suffering increased for SOCS-O in the high-sensitivity group; for SOCS-S the recognizing suffering subscale increased. Notably, self-compassion rose after both HAR and DMT/HAR, consistent with a nonspecific serotonergic contribution of harmine. No sustained effects were detected at the 1-month and 4-month follow-ups according to the extracted text.

Aicher and colleagues interpret the findings as evidence that an ayahuasca-inspired DMT/harmine formulation can elicit acute increases in measures of mindfulness and compassion in healthy volunteers tested in a controlled laboratory setting. The investigators note that the observed magnitude of change on the MINDSENS index is comparable to effects seen in meditation practitioners for whom the index was developed, suggesting that pharmacological induction of altered states may transiently produce similar capacities such as non-reactivity and decentering without concurrent meditation training. The authors highlight that effects were concentrated in participants classified as high-sensitivity to the DMT/HAR condition; these individuals showed robust increases across mindfulness and compassion measures, whereas low-sensitivity participants did not exhibit condition-dependent changes. Harmine alone produced increases in self-compassion, which the team attributes to nonspecific serotonergic effects of the MAO-A inhibitor. The absence of persistent effects at 1 and 4 months is emphasised, and the authors argue that longer-term stabilisation of mindfulness/compassion likely requires integration or concomitant therapeutic/mindfulness training. The investigators place their results alongside prior ayahuasca and psychedelic research, noting overlap with reports of enhanced empathy, prosociality, and antidepressant effects, while acknowledging that prior naturalistic studies may be confounded by ceremonial or group contexts. They propose that pharmacological formulations such as DMT/HAR may transiently increase openness, acceptance and affiliative attitudes even outside ceremonial settings, which could be therapeutically useful if combined with psychotherapy or compassion-focused interventions. Limitations acknowledged by the authors include the homogeneous sample (male-only, young, highly educated), the laboratory rather than therapeutic setting, moderate dosing chosen to permit concurrent neurobehavioural testing, potential unblinding especially among high-sensitivity participants, and lack of formal assessment of blinding efficacy. The investigators recommend dose–response work, examination of personality moderators such as trait absorption, and trials in clinical populations to determine effective doses and to test whether DMT/HAR–assisted interventions can produce clinically meaningful and durable changes in mindfulness, compassion, and psychiatric outcomes.