Assessing the psychedelic “after-glow” in ayahuasca users: post-acute neurometabolic and functional connectivity changes are associated with enhanced mindfulness capacities

Psychedelics have regained research interest for their potential therapeutic effects in psychiatry. Ayahuasca is an Amazonian brew combining DMT, a 5-HT2A agonist, with β-carbolines that inhibit monoamine oxidase and enable oral activity of DMT. Acute ayahuasca administration produces intense introspective states and measurable neurophysiological changes in nodes of the default mode network (DMN), particularly the posterior cingulate cortex (PCC), while radiotracer studies point to increased activity in limbic and cingulate regions involved in emotion and memory. Previous clinical reports have also documented rapid and sustained antidepressant effects and post-acute increases in mindfulness-related capacities after single doses, but the neural mechanisms that might underlie these sustained or "after-glow" effects remain unclear. Sampedro and colleagues set out to characterise post-acute (within 24 hours) neurometabolic and resting-state functional connectivity changes after a single ayahuasca session in healthy volunteers, and to test whether those neural changes relate to enhancements in mindfulness capacities measured immediately and at a 2-month follow-up. They focused on the PCC and anterior cingulate cortex (ACC) as a priori regions of interest and hypothesised: (1) post-acute decreases in Glx (combined glutamate plus glutamine) and energy metabolites in PCC and ACC reflecting prior excitatory effects; (2) reduced PCC–ACC anticorrelation (increased coupling) paralleling neurometabolic changes; and (3) correlations between neural changes and increases in mindfulness measures both post-acutely and at follow-up. The study is an open-label, within-subject pre/post assessment in experienced ayahuasca users.

This was an open-label, uncontrolled within-subject MRI study in 16 healthy volunteers (6 females) with prior ayahuasca experience. Exclusion criteria included current or past psychiatric disorder, substance dependence, significant illness, and pregnancy. All participants provided written informed consent and the protocol was approved by a local ethics committee. A single, chemically analysed batch of ayahuasca was administered to all participants. Mean intake was 148 ± 29 mL, corresponding to mean doses of 45 ± 9 mg DMT and quantified amounts of harmine, tetrahydroharmine and harmaline; the reported DMT dose is approximately 0.64 mg/kg for a 70-kg person. No placebo or control condition was used; analyses compared post- vs pre-ayahuasca values within participants. Each participant underwent MRI twice: a baseline scan within 24 hours before the ayahuasca session and a post-acute scan within 24 hours after the session, scheduled at approximately the same time of day. Imaging on a 3T Siemens TIM Trio included high-resolution T1, single-voxel 1H magnetic resonance spectroscopy (MRS) in three volumes of interest (VOIs: PCC, ACC, cerebellum), and 7-minute resting-state BOLD for functional connectivity. MRS quantified absolute concentrations of Cr (creatine), Cho (choline), Ins (inositol), Glx (glutamate plus glutamine), and NAA-NAAG (N-acetylaspartate plus N-acetylaspartylglutamate). Absolute metabolite levels rather than ratios were used to avoid potential confounds with creatine. Resting-state functional connectivity analyses used CONN software. Seed-to-voxel maps were computed for three spherical seeds (10-mm radius): PCC at MNI 0,-56,28; dorsal ACC (dACC) at 5,14,42; and superior rostral ACC (srACC) at 0,15,30. Preprocessing and first-level analyses produced individual z-maps which entered paired-samples second-level t tests comparing post- vs pre-ayahuasca. Cluster-level family-wise error correction (FWE) was applied with threshold FWE < 0.05 and minimum cluster size 20 voxels. Psychological measures comprised the Spanish versions of the Five Facet Mindfulness Questionnaire (FFMQ), the Experiences Questionnaire (EQ), and a short Self-Compassion (SC) scale; a Mindsens Composite Index was computed from FFMQ and EQ items. Questionnaires were completed at baseline, within 24 hours post-intake, and at a 2-month follow-up (14/16 participants responded). Acute subjective effects were assessed retrospectively at the post-acute visit using the Hallucinogen Rating Scale (HRS) and compared to previously collected laboratory data for verification of psychoactivity. Statistical analyses: MRS concentrations were analysed with repeated-measures ANOVA (factors: ayahuasca intake and metabolite type) per VOI, followed by pair-wise t tests when appropriate; significance set at P < .05 and FDR correction reported. Connectivity changes used paired t tests on z-maps with FWE cluster correction. Mindfulness changes were tested with paired t tests; follow-up analyses were noted as exploratory due to smaller sample. Exploratory correlations used difference scores (post-minus-pre) between neural measures and questionnaire or HRS scores; no multiple comparisons correction was applied for these exploratory correlations and results were interpreted cautiously.

All 16 participants completed baseline and post-acute assessments; no drop-outs or exclusions for movement in the fMRI data were reported. MRS data availability varied due to signal-to-noise limitations: usable Glx data in the PCC for 12 participants, ACC for 13, cerebellum for 11; most other metabolites quantified in 14 participants. In the PCC VOI, repeated-measures ANOVA showed a significant metabolite-type effect and an interaction between intake and metabolite type. Pair-wise comparisons indicated significant post-acute decreases in creatine (Cr) and NAA-NAAG; Glx (glutamate+glutamine) was reduced in an initial t test (P = .041) but this effect became marginal after FDR correction (P = .068). Inositol and choline showed no changes. In the ACC and cerebellum VOIs there were no significant effects of ayahuasca intake on metabolite concentrations apart from the expected metabolite-type differences. Resting-state connectivity analyses found several post-acute changes. Prior to ayahuasca, the PCC seed showed positive correlations with DMN regions and negative correlations (anticorrelations) with ACC and task-positive networks (TPNs). Post-acutely, the negative correlation between PCC and ACC was markedly reduced, indicating decreased orthogonality of their BOLD activity. Cluster-level increases in PCC seed connectivity were observed in the ACC and a large occipital visual cluster where connectivity shifted from negative to positive; additional increases emerged in pre- and post-central gyri, superior temporal gyrus and insula. No clusters showed decreased connectivity for the PCC seed in the second-level analysis. Using the dACC seed, post-acute increases in connectivity appeared with medial parietal cortex including precuneus/PCC, while connectivity with visual association areas (BA 18/19) decreased. The srACC seed showed increased coupling with medial temporal lobe (MTL) limbic structures on the right — parahippocampal gyrus, hippocampus, amygdala — as well as left parietal and medial parietal (PCC) regions; srACC–visual cortex coupling was reduced post-acutely. On psychological measures, paired comparisons showed post-acute increases in FFMQ subscales nonjudging [t(15) = -2.92, P = .011; FDR P = .021] and nonreacting [t(15) = -2.61, P = .020; FDR P = .031]. The EQ increased post-acutely [t(15) = -3.58, P = .003; FDR P = .011], the Mindsens composite rose [t(15) = -3.63, P = .002; FDR P = .011], and self-compassion scores increased [t(15) = -3.00, P = .009; FDR P = .020]. At the 2-month follow-up (n = 14), only the nonjudging FFMQ score remained significantly higher than baseline by paired t test [t(13) = -2.22, P = .045], though this did not survive FDR correction (P = .495); other scales returned to baseline. Retrospective HRS data confirmed significant acute psychoactive effects; in comparison with prior laboratory data at 0.75 mg DMT/kg, most HRS subscales matched that medium dose except cognition which was higher in the present sample (t(26) = 4.25, FDR P < .01). Exploratory correlations linked neural and psychological changes. Greater decreases in PCC Cr and in NAA-NAAG correlated with higher HRS-Cognition scores (Cr: n = 14, r = -0.735, r2 = 0.540, P = .003; NAA-NAAG: n = 14, r = -0.622, r2 = 0.387, P = .018). Glx decreases in the PCC correlated with increases in the FFMQ nonjudging subscale post-acutely (n = 12, r = -0.589, r2 = 0.506, P = .044) and at 2 months (n = 11, r = -0.740, r2 = 0.548, P = .009). Functional connectivity increases between srACC and PCC correlated with post-acute increases in nonjudging (n = 16, r = 0.604, r2 = 0.365, P = .013) and nonreacting (n = 16, r = 0.522, r2 = 0.272, P = .038); srACC–MTL connectivity increases correlated with nonjudging (n = 16, r = 0.637, r2 = 0.406, P = .008), nonreacting (n = 16, r = 0.656, r2 = 0.430, P = .006), and self-compassion (n = 16, r = 0.514, r2 = 0.264, P = .042). Post-acute increases in srACC–PCC and ACC–MTL connectivity were also positively associated with sustained nonjudging at follow-up (srACC–PCC: n = 14, r = 0.584, r2 = 0.341, P = .028; ACC–MTL: n = 14, r = 0.566, r2 = 0.320, P = .035). The authors note these correlation analyses were exploratory and not corrected for multiple comparisons.

Sampedro and colleagues interpret their findings as evidence that ayahuasca induces measurable neurometabolic and functional connectivity changes that persist into the post-acute "after-glow" period and relate to enhanced mindfulness capacities. The post-acute reductions in Glx in the PCC (Glx = combined glutamate plus glutamine) together with decreases in Cr and NAA-NAAG are consistent with a model whereby acute 5-HT2A-driven excitatory glutamatergic activity is followed by a relative post-acute reduction in metabolite levels. The authors argue this pattern fits prior electrophysiological reports of decreased inhibitory alpha rhythms during acute psychedelic states and with other data suggesting acute glutamatergic activation. Functionally, the data showed reduced anticorrelation and increased coupling between DMN (PCC) and ACC/TPN regions, plus enhanced srACC connectivity with medial temporal limbic structures. According to the authors, these changes provide a neural substrate for the observed psychological effects: reduced self-referential, judgmental and reactive processing, increased decentering and self-compassion, and greater interplay between cognitive control and memory/emotion circuits. They note parallels with mindfulness training, which also reduces DMN–TPN anticorrelation, and propose that such neural reorganisation could help explain rapid and sustained antidepressant effects reported in clinical studies. The authors acknowledge several limitations. The sample was small and open-label with no placebo or time-control condition, so effects could reflect non-specific or expectancy influences. MRS Glx quantification suffered from signal limitations that reduced sample size for some analyses. All participants had prior ayahuasca experience and presented relatively high baseline mindfulness scores, which may both bias results and constrain the magnitude of potential increases. Correlational findings were exploratory and not corrected for multiple comparisons. Finally, only a subacute window was sampled (within 24 hours post-intake), so connectivity changes cannot be assumed to reflect persistent network reorganisation over longer timescales. For future research, the authors recommend larger, double-blind, placebo-controlled designs and inclusion of ayahuasca-naive participants to disentangle prior exposure effects. They conclude that the observed post-acute neurometabolic decreases in PCC and increased inter-network connectivity were associated with enhanced mindfulness facets, providing tentative neurobiological clues to how ayahuasca and similar psychedelics might exert therapeutic effects.