Mystical experiences during magnesium-Ibogaine are associated with improvements in PTSD symptoms in veterans

Posttraumatic stress disorder (PTSD) is common among combat-exposed veterans and co-occurs frequently with traumatic brain injury (TBI). Conventional pharmacological advances for PTSD have been limited over recent decades, and combat-related PTSD—particularly when comorbid with TBI—often responds poorly to standard treatments. In response, there has been renewed interest in psychedelic and related compounds, with trials of MDMA and ketamine showing promising results; a recent open-label investigation of ibogaine combined with magnesium (magnesium-ibogaine) also reported substantial PTSD symptom improvements after a single session in combat-exposed veterans with TBI. Despite these clinical signals, the cognitive and neurophysiological mechanisms by which magnesium-ibogaine may exert therapeutic effects remain unclear. The present study, a secondary analysis of data from the Stanford Traumatic Injury to the CNS (MISTIC) project, examined whether the intensity of mystical experiences during magnesium-ibogaine treatment is associated with clinical and neurophysiological changes. The primary registered hypothesis was that higher scores on the Mystical Experiences Questionnaire (MEQ30) would predict greater reductions in clinician-rated PTSD severity (CAPS-5) immediately after treatment; secondary and exploratory hypotheses extended this to MEQ30 subscales, depression and anxiety severity, functional disability, and selected EEG markers (including peak alpha frequency, PAF). E. and colleagues framed this work to probe whether subjective mystical phenomena relate to therapeutic outcomes and to investigate potential neural correlates of those experiences following magnesium-ibogaine in a treatment-resistant, high-risk veteran sample.

This is a secondary analysis of an open-label, prospective investigation (MISTIC) of magnesium-ibogaine conducted with 30 male Special Operations veterans who independently enrolled in treatment at a clinic in Mexico. In-person assessments were conducted at Stanford 2–3 days before dosing (baseline), 3.5–5 days after dosing (immediate-post), and at 1 month post-treatment. Inclusion criteria included age 18–70, capacity to consent, history of head trauma and combat/blast exposure; exclusions included psychotic history, neurological disorders (other than TBI sequelae), MRI contraindication, and high suicidal risk. Psychiatric diagnoses were determined with the MINI; combat exposure, TBI history, PTSD severity, depression, anxiety, and functional disability were assessed with the CES, BAT‑L, CAPS‑5, MADRS, SIGH‑A, and WHODAS‑2.0 respectively. The magnesium-ibogaine protocol spanned 5 days with dosing on Day 2. Preparatory measures included fasting and intravenous magnesium sulfate (1 g) followed by test and incremental oral/intravenous ibogaine dosing up to a maximum of 14 mg/kg (one participant later received an additional 4 mg/kg). Participants were monitored for 72 hours post-dose and received postdosing integration sessions and optional wellness activities; separate preparatory and integration psychotherapy sessions were provided virtually by experienced therapists. Mystical experiences were measured with the MEQ30, administered as an online questionnaire at the immediate-post time point (three participants completed the MEQ30 substantially later; sensitivity analyses addressed this). Resting-state eyes-open EEG (six minutes) was recorded at baseline (N = 30), immediate-post (N = 30), and 1-month (n = 27) using a 64-channel system. Preprocessing steps included downsampling, filtering, artifact rejection (including ICA), epoch rejection, and re-referencing. EEG outcomes of interest—selected based on prior ibogaine findings—included spectral power in canonical bands (delta–gamma) normalised to broadband power, theta/beta ratio, global peak alpha frequency (PAF; identified as the local maximum 7–14 Hz per channel then averaged), and Lempel–Ziv complexity (LZc). Spectral analyses used Welch's method; analyses of power and theta/beta ratio were focused on medial frontal and medial posterior regions of interest. For some PAF analyses, participants without identifiable alpha peaks in >20% of channels were excluded, reducing the effective EEG sample in those models. The primary statistical approach used linear mixed models (LMMs) with CAPS‑5 scores at the three timepoints as the dependent variable, MEQ30 total score, time point (factor), and their interaction as fixed effects, and a random intercept for each participant. Age, CES total, and number of TBIs were included a priori as covariates. MEQ30 scores were entered as per‑item averages (0–5) to aid interpretability. Parameters of interest were the MEQ30 × time interactions for baseline→immediate-post (primary) and baseline→1‑month (secondary). Analyses also substituted MEQ30 subscales or alternative dependent variables (e.g., MADRS, SIGH‑A, EEG measures) as appropriate. Significance was interpreted at p < .05 without formal alpha adjustment; assumptions of the LMMs were inspected and no major violations were reported. Spearman correlations complemented model results. Sensitivity analyses excluded the three participants with delayed MEQ30 completion and tested dose (mg/kg) as a potential confound.

Sample and descriptive findings: All 30 participants completed baseline and immediate-post clinical assessments; 23 met diagnostic criteria for PTSD at baseline on the MINI. The BAT‑L TBI counts included extreme outliers (three participants reported TBI counts ‘‘in the thousands’’), and the researchers imputed those three values as five times the standard deviation for statistical purposes. MEQ30 completion was mostly within 3–7 days post-dosing except for three delayed completions (34–35 days), which were examined in sensitivity analyses. Primary clinical outcome: The LMM showed a significant MEQ30 × time interaction for baseline→immediate-post on CAPS‑5 scores: unstandardised B_adj = −5.89, F(1,56) = 13.65, p < .001, 95% CI [−9.08, −2.70]. In practical terms, every 1‑point increase on the MEQ30 average (0–5 scale; equivalent to a 30‑point increase on the raw MEQ30) was associated with an expected 5.89‑point greater reduction in CAPS‑5 from baseline to immediate-post. A Spearman correlation corroborated this (rho = −0.53, p = .003). The association persisted at 1 month: MEQ30 × time interaction B = −4.45, F(1,56) = 7.79, p = .007, 95% CI [−7.64, −1.26], with Spearman rho = −0.49, p = .007. MEQ30 subscales: LMMs for the MEQ30 subscales (mystical, positive mood, transcendence of space/time, ineffability) largely showed similar associations with CAPS‑5 reductions; all subscale × time interactions were significant except for the ineffability subscale at the baseline→1‑month contrast. The authors interpreted this as no single subscale driving the effect and consistent with MEQ30 functioning as a largely unidimensional measure. Secondary clinical outcomes: MEQ30 total score was significantly associated with reductions in depression (MADRS) and anxiety (SIGH‑A) at both immediate-post and 1‑month timepoints. No significant association was found between MEQ30 total score and change in functional disability (WHODAS‑2.0). EEG/neurophysiological outcomes: For global PAF, the MEQ30 × time interaction was non‑significant for baseline→immediate-post (B = −0.098, F(1,45) = 0.561, p = .46, 95% CI [−0.36, 0.17]) but significant for baseline→1‑month (B = −0.38, F(1,45) = 8.15, p = .006, 95% CI [−0.65, −0.11]). Spearman correlation between MEQ30 total and change in PAF baseline→1‑month was rho = −0.49, p = .019. No significant MEQ30 × time interactions were observed for the other EEG measures of interest (band power, theta/beta ratio, LZc). Sensitivity and supplementary analyses: Excluding the three participants who completed the MEQ30 substantially later did not materially alter the findings, and correlations between MEQ30 scores and delay to completion did not indicate an effect of completion timing. Analyses including ibogaine dose (mg/kg) as a covariate did not show dose effects on primary outcomes. Age, CES total, and number of TBIs were not significant predictors in the LMMs.

E. and colleagues interpret these results as evidence that, within this open‑label cohort of combat veterans with TBI, greater intensity of mystical experiences during magnesium‑ibogaine treatment was associated with larger reductions in PTSD symptoms both immediately and one month later. The observed effect size for the primary correlation was moderate (rho ≈ 0.53) with a confidence interval spanning from weak to strong effects; the association attenuated slightly but remained significant at one month. The authors note that causal inference is not possible from these correlational data and acknowledge that mystical experiences might be an epiphenomenon rather than a mechanistic mediator. The authors offer possible psychological mechanisms by which mystical experiences may facilitate PTSD improvement: profound feelings of connection or sacredness may create a safer context for trauma reprocessing, experiences of ‘‘ego dissolution’’ could loosen rigid self‑schemas (e.g. beliefs of being irreparably damaged), and enhanced connectedness may reduce dissociation. On the neurophysiological side, the association between higher MEQ30 scores and sustained reductions in global PAF at one month suggests a potential neural correlate. The authors link slowed PAF to reductions in hyperarousal and to decreased top‑down inhibitory control, invoking the REBUS model (which proposes that psychedelics relax top‑down predictive priors) as a conceptual framework; sustained PAF slowing may reflect longer‑term relaxation of top‑down constraints with relevance to arousal and attention systems. The authors discuss strengths of the work, notably that this is the first study to examine mystical experience in relation to clinical and EEG outcomes after magnesium‑ibogaine in a TBI/PTSD veteran population, using clinician‑administered PTSD assessment and objective neurophysiology. They also acknowledge multiple limitations: the open‑label design and absence of a control condition preclude disentangling pharmacological effects from expectancy; the MEQ30 may not capture relevant constructs such as connectedness, may emphasise positive rather than adverse aspects of the experience, and its overlap with other acute subjective effects complicates specificity claims. Timing of MEQ30 completion (most participants completed it after integration sessions) could bias recall. The ineffability subscale's lack of significance at one month may reflect limited reliability due to only three items. The authors also note that animal data showing antidepressant-like effects from some non‑hallucinogenic analogues raise the possibility that subjective mystical content may not be strictly necessary for therapeutic outcomes, emphasising uncertainty about mechanism. For future research, E. and colleagues recommend longer follow‑up to assess durability, randomized controlled trials (including dose or active control conditions) to address expectancy and causal inference, larger and more diverse samples for generalisability, EEG recording during the dosing session to better characterise neural correlates of the acute subjective experience, psychometric work on the MEQ30 in the context of ibogaine (and use of alternative or additional measures that capture connectedness, emotional breakthrough, and negative acute effects), and broader investigation of subjective phenomenology that may mediate clinical benefit.