Intact neurophysiological markers of death denial in long-term ayahuasca users

The authors frame death as a universal existential concern that humans manage through cultural, religious, contemplative and pharmacological practices. Recent interest in classic psychedelics—notably 5-HT2A agonists—has highlighted claims that these substances can reduce fear of death and increase acceptance, with potential clinical relevance for end-of-life distress. However, much of the existing evidence relies on self-report, retrospective surveys or small clinical studies; the authors note a gap concerning implicit, behavioural and neurobiologically grounded measures of mortality-related processing, especially at unconscious, automatic levels. To address this gap, the study applied a previously validated no-report magnetoencephalography (MEG) visual mismatch response (vMMR) paradigm that indexes a self-specific, prediction-based neurophysiological marker of "death denial" (the Death Self-Disadvantage, DSD). The primary aim was to test whether long-term naturalistic ayahuasca users show neurophysiological evidence of reduced death denial (i.e. greater acceptance) similar to experienced meditators, or whether implicit perceptual markers remain intact despite changes on explicit measures. Secondary objectives included exploratory correlations to examine the construct validity of the DSD and links with affect, thought accessibility and well-being.

Design and participants: The study recruited 50 healthy adult long-term ayahuasca users (17 female; mean age 38.4 ± 8.93, range 25–68) who reported multiple uses (≥ 8 ceremonies) and considered ayahuasca their primary psychedelic. Mean lifetime ceremony count was 58.4 ± 84.8. An a priori power calculation (G*Power) targeting a medium effect (f = 0.25) and 95% power at α = 0.05 indicated n = 36 was sufficient, making the final n = 50 adequate. Exclusion criteria included current psychoactive medication, recent psychedelic use (within 28 days), frequent cannabis use (≥ twice monthly), neurological/active psychiatric illness, recent bereavement (loss of first-degree relative within 12 months), and conditions affecting MEG data quality. The study had institutional ethics approval and participants were compensated. The authors also re-analysed previously published MEG datasets from a general population sample and an experienced meditator cohort collected in the same laboratory and apparatus for group comparisons. vMMR task and behavioural measures: The core task was a fully randomized within-subject 2 × 2 vMMR paradigm crossing PRIME (DEATH vs NEGATIVE words) and IDENTITY (SELF vs OTHER faces). Trials began with a prime word (600 ms), followed by 3–6 standard face presentations (either self or other) and then a deviant face (50% self–other morph). Faces were displayed for 250 ms with a 350 ms blank inter-stimulus interval. Participants performed a vigilance task (press a button when a face with sunglasses appeared) to maintain attention. The session comprised 360 trials across three blocks, including 90 randomly appearing target trials and an average of ~67.5 trials per condition with a 1:4.5 standard:deviant ratio. Self-report and behavioural instruments: Explicit and implicit measures of fear of death used a modified Fear of Personal Death Scale (FPDS) presented as forced-choice yes/no items; explicit fear was the percentage of "yes" responses (FPDS_P) and implicit fear was indexed by reaction-time based measure (FPDS_RT) normalised against responses to a dental pain control questionnaire. Death acceptance was assessed with the Life Attitude Profile Revised Death Acceptance subscale (LAPR-DA). Death Thought Accessibility (DTA) used a 20-item fragment completion task; higher counts of death-related completions indicate greater accessibility of death thoughts. Death anxiety used the Death Anxiety Scale (DAS). Well-being and life satisfaction were measured with the General Well-Being Scale (GWBS) and the Satisfaction With Life Scale (SWLS). Additional measures included afterlife belief certainty (ABC), ego-dissolution experiences (Ego Dissolution Inventory, EDI), and ayahuasca usage history. MEG acquisition and preprocessing: Continuous MEG was recorded with a whole-head 248-channel magnetometer array (Magnes 3600 WH) sampled at 1017.23 Hz (0.1–400 Hz online band-pass). Head position was monitored and visual stimulus timing marked via a photosensitive diode. Preprocessing used FieldTrip and custom MATLAB scripts: environmental noise and artifacts (power-line, mechanical, SQUID jumps, heartbeat) were removed with a predefined algorithm; two malfunctioning channels were excluded. Data were epoched −100 to 500 ms around face onset, baseline-corrected using the −100 to 0 ms interval, low-pass filtered at 40 Hz (two-pass Butterworth fourth-order), and planar gradient transformed to aid sensor-level interpretation. Event-related fields (ERFs) were computed and deviant minus standard difference waveforms generated. Trials were balanced per condition for each participant. Statistical analyses: Time–sensor selection and hypothesis testing employed non-parametric cluster-based permutation tests (1,000 permutations) to control family-wise error across space/time. Primary inference on condition effects used repeated-measures and mixed-design ANOVAs with factors LOCATION (left/right posterior cluster), IDENTITY (self/other), and PRIME (death/negative), and group (ayahuasca/meditators/controls) for cross-group comparisons. Post-hoc t-tests were corrected using the Holm–Bonferroni procedure. Behavioural vigilance performance (hit rates and reaction times) was also analysed to rule out attentional confounds. Correlations (Pearson or Spearman depending on distribution) examined associations between the DSD index and self-report/behavioural measures; the authors note multiple-comparison considerations for exploratory correlations.

Sensor-level vMMR effects: Cluster-based permutation testing identified a primary deviant > standard vMMR time window between 250 and 301 ms post-stimulus (p = 0.027), with significant posterior-left (5 sensors, p = 0.016) and posterior-right (21 sensors, p = 0.002) clusters. A later time window (424–464 ms) was also detected (p = 0.042) but the corresponding left-posterior spatial cluster was only marginally significant (3 sensors, p = 0.054) and not emphasised in the main text. Within-group condition effects in ayahuasca users: Collapsing across posterior clusters, a repeated-measures ANOVA showed a significant interaction of IDENTITY × PRIME (F(1,49) = 8.05, p = 0.0007, ηp2 = 0.141), indicating the combined influence of face identity and prime type on the vMMR. Post-hoc analyses revealed that, contrary to the hypothesis, long-term ayahuasca users did not show a significant vMMR for the death+self condition (DS), mirroring the pattern previously observed in the general population sample (interpreted as a neurophysiological marker of death denial). The extracted text does not clearly report the specific t statistic for the DS within-group test in the ayahuasca sample. Group comparisons (ayahuasca, meditators, controls): Mixed ANOVA analyses using the shared left-posterior cluster data (normalized as z-scores) yielded a significant GROUP × IDENTITY × PRIME interaction (F(2,109) = 5.306, p = 0.006, ηp2 = 0.089). One-way ANOVAs performed separately per condition showed a significant group effect only for the DS condition (F(2,109) = 4.64, p = 0.012, Holm–Bonferroni pHB = 0.048, ηp2 = 0.078). Post-hoc comparisons indicated that DS values were lower (smaller DS vMMR) for ayahuasca users than for meditators (t = 2.51, p = 0.013, pHB reported, d = 0.54, 95% CI [0.11, 0.99]), but ayahuasca users did not differ from controls (t = 0.66, p = 0.512, n.s.). Mixed ANOVAs comparing ayahuasca users directly with meditators confirmed a significant GROUP × IDENTITY × PRIME effect (F(1,86) = 8.554, p = 0.004, pHB = 0.008, ηp2 = 0.09), consistent with meditators showing a pronounced DS vMMR (interpreted as acceptance) while ayahuasca users resembled controls in showing attenuation for DS (interpreted as denial). Behavioural vigilance and stimulus control checks: Target-detection performance showed high attentiveness (> 96% hit rate across conditions) and no condition-related differences in hit rates or reaction times (ps > 0.05). Word lists for death and negative primes were balanced for valence, arousal, length and frequency, and the number of trials per condition did not differ across the current and comparison datasets. Self-report and behavioural measures of death-related affect: On the modified FPDS explicit measure (percentage of "yes" responses), a one-way ANOVA across the three groups was significant (F(2,108) = 10.9, p < 0.001, ηp2 = 0.17). Post-hoc tests indicated that ayahuasca users reported less explicit fear of death than controls (t = -4.65, p < 0.001; 95% CI for the difference reported as [-1.68, -0.64]); ayahuasca users also reported less fear than meditators (t = -2.05, p = 0.043, d = -0.44). For the FPDS_RT implicit latency measure there was a marginal omnibus effect (F(2,108) = 3.04, p = 0.052, ηp2 = 0.053); ayahuasca users responded faster than controls (t = -2.03, p = 0.045, d = -0.51, 95% CI [-1, -0.01]), interpreted as less implicit fear. No reliable correlations were found between the neurophysiological DSD index and measures of fear and anxiety of death (all ps > 0.05). Construct validity and exploratory correlations: The DSD showed a modest inverse correlation with a self-report death acceptance measure (LAPR-DA; p = 0.03) and a marginal inverse correlation with the behavioural Death Thought Accessibility measure (DTA; p = 0.053), such that larger neurophysiological denial signals were associated with less self-reported death acceptance and fewer death-related word completions respectively. The DSD also correlated with satisfaction with life (p = 0.01). The authors note that none of these correlations survived correction for multiple comparisons and should therefore be regarded as exploratory. Summary of pattern: Overall, ayahuasca long-term users exhibited reduced explicit and implicit fear-of-death on reflective/behavioural measures but retained the fast, automatic neurophysiological signature of death denial (attenuated DS vMMR) similar to the general population, whereas experienced meditators displayed an opposite pattern (a robust DS vMMR interpreted as acceptance).

The authors interpret their findings as showing a dissociation between levels of cognitive processing in mortality-related responses: long-term ayahuasca users reported and behaviourally displayed reduced fear of death, yet their automatic, self-specific perceptual neurophysiology retained the death-denial signature observed in the general population. This outcome was contrary to the authors' primary hypothesis that ayahuasca users would show a DS vMMR more like meditators (i.e. evidence of implicit acceptance). They argue these results suggest that psychedelics may modulate higher-order, reflective aspects of death-related cognition and affect (consistent with reductions in self-reported fear and faster responses on the FPDS_RT), but that earlier, faster, non-conscious self-specific perceptual processes can remain unchanged despite repeated naturalistic ayahuasca exposure. The authors present the meditators' data as evidence that the vMMR death-denial mechanisms are not "hardwired" and therefore amenable to change, and they propose a hypothesis to account for the group differences: long-term alterations in embodied, self-specific perceptual processing may require sustained attentional training (as in meditation) rather than pharmacological perturbation alone, even when the latter produces powerful acute self-dissolution experiences. The authors report preliminary support for the construct validity of the DSD: modest inverse associations with explicit death acceptance (LAPR-DA) and death-thought accessibility (DTA) suggest the neurophysiological index relates to other measures of death-related processing, and a positive association with life satisfaction hints at a possible adaptive function of mortality-denial mechanisms. However, these correlations were small and did not survive correction for multiple comparisons, so the authors present them as exploratory and tentative. Acknowledged limitations include the cross-sectional design and reliance on secondary comparisons to previously collected control and meditator data rather than head-to-head contemporaneous randomised groups; the potential for self-selection bias in long-term ayahuasca users; and sample specificity (highly experienced users may not represent naïve or short-term users). The novelty and psychometric status of the FPDS_RT measure is also noted as a constraint on interpretability. The authors call for longitudinal, controlled studies—ideally randomising naïve participants to psychedelic, meditative or control conditions—to disentangle causal effects and to test the proposed attentional-training versus pharmacological hypotheses. They also suggest investigating combined pharmacological and contemplative interventions as a promising avenue. Despite limitations, the authors view the study as contributing novel, neurobiologically informed evidence about the limits and specificity of psychedelics' long-term impact on how humans process mortality at automatic perceptual levels.