Subacute Effects of the Psychedelic Ayahuasca on the Salience and Default Mode Networks

Classic psychedelics produce altered states of consciousness with wide-ranging effects on sensory, cognitive, autonomic, interoceptive, self-referential, and emotional systems. Ayahuasca is a traditional Amazonian admixture containing N,N-dimethyltryptamine (N,N-DMT) and monoamine oxidase inhibitors; N,N-DMT has high affinity for serotonergic and sigma-1 receptors, while the monoamine oxidase inhibitors render it orally active. Recent clinical trials have reported rapid antidepressant effects of ayahuasca, including benefits observed 1 day after a single session, but the neural mechanisms underlying changes in affect, mood, and self-representation remain incompletely understood. Task-free functional MRI (tf-fMRI) can map intrinsic large-scale brain networks such as the salience network (involving the anterior cingulate cortex and anterior insula) and the default mode network (involving posterior cingulate cortex, precuneus, medial prefrontal cortex), which are implicated in interoception, socioemotional processing, and self-referential thought. Pasquini and colleagues set out to characterise subacute (24-hour) changes in intra- and inter-network functional connectivity after a single ayahuasca session in ayahuasca‑naïve healthy participants, and to relate those changes to subjective measures of the acute psychedelic experience. The primary hypothesis was that 24 hours after dosing ayahuasca would alter higher-order affective and self-referential networks (salience and default mode) but not primary sensory networks. The investigators used a seed-based connectivity approach, complemented by independent component analysis (ICA), to map salience, default mode, visual and sensorimotor networks before and 24 hours after dosing in a randomised placebo-controlled design.

This randomised, placebo-controlled study recruited 50 cognitively and physically healthy, ayahuasca‑naïve participants from local media and online advertisements. Exclusion criteria included current chronic disease; the extracted text does not list the full set of exclusions. Participants were randomised to receive a single oral dose of ayahuasca (1 mL/kg; n intended = 25) or a placebo beverage (1 mL/kg; n intended = 25). Alkaloid concentrations in the ayahuasca batch were quantified by mass spectrometry and reported as group means: N,N-DMT 0.36 ± 0.01 mg/mL, harmine 1.86 ± 0.11 mg/mL, harmaline 0.24 ± 0.03 mg/mL, and tetrahydroharmine 1.20 ± 0.05 mg/mL. The placebo was designed to mimic organoleptic properties (bitter/sour taste, brown colour) and contained water, yeast, citric acid, zinc sulfate and caramel colourant; zinc sulfate produced modest gastrointestinal discomfort to partially control for non-specific effects. Task-free fMRI scans were acquired on a 1.5 Tesla scanner approximately 1 day before dosing (baseline) and again ~24 hours after the session. EPI parameters were TR = 2000 ms, TE = 35 ms, flip angle 60°, 35 slices, 213 volumes; T1-weighted structural images used an FSPGR BRAVO sequence. Seven participants were excluded after initial inclusion (three from ayahuasca, four from placebo) for reasons including loss to follow-up, scanning artefacts, a discovered cystic lesion, and elevated blood glucose discovered post hoc; the final analysed sample was 43 (22 ayahuasca, 21 placebo). Preprocessing used CONN and included discarding the first three volumes, slice-timing correction, realignment, spatial smoothing (6 mm FWHM), coregistration to structural images, normalisation to MNI space (2 mm3 voxels), motion assessment with ART (outlier volumes defined by global signal >5 SD or framewise displacement >0.9 mm), and physiological/noise regression using CompCor. Residual motion parameters, outlier volumes, white matter and CSF signals were regressed out, and a temporal band-pass filter of 0.01–0.1 Hz was applied. Mean framewise displacement was low in the sample and near a stringent 0.25 mm threshold. For intra-network analyses, bilateral regions-of-interest from the Brainnetome Atlas were used as seeds: anterior cingulate cortex, posterior cingulate cortex, occipital cortex and precentral cortex. Voxel-wise seed-based regression generated salience, default mode, visual and sensorimotor connectivity maps. Group ICA (20 components, two-step PCA reduction, infomax algorithm) was performed as a control. Inter-network connectivity was estimated by extracting average time series from the identified network maps and computing pairwise Pearson correlations. Statistical analyses used voxel-wise one-sample t-tests to define masks, then voxel-wise two-sample t-tests on change maps (post minus pre) to compare groups; reported voxel-cluster thresholds varied (commonly joint cluster and extent probability p<0.05 and p<0.01 uncorrected, with some stricter FWE-corrected thresholds at mask derivation). The investigators controlled for head-motion differences in secondary analyses, ran nonparametric permutation testing with FSL Randomise (threshold-free cluster enhancement, p<0.05 with FWE correction), and used partial correlations (controlling for framewise displacement differences) to associate connectivity changes with Hallucinogen Rating Scale (HRS) subscales assessed ~4 hours post-dosing. Two-sample t-tests assessed demographic and HRS group differences; multiple comparison corrections were applied where stated.

Forty-three participants (22 ayahuasca, 21 placebo) passed quality checks and had tf-fMRI at both time points. Groups were comparable for age, sex distribution and head motion. Acute subjective effects measured by the Hallucinogen Rating Scale were robustly greater in the ayahuasca group across most subscales: somesthesia (p<0.0005), affect (p<0.005), perception (p<0.0005), cognition (p<0.0005), volition (p<0.009) and intensity (p<0.0005); all survived correction for multiple comparisons except volition. Comparing change maps (post minus pre) between groups, voxel-wise two-sample t-tests indicated increased functional connectivity within the salience network in the ayahuasca group relative to placebo, localised bilaterally to the anterior cingulate cortex and superior frontal gyrus (reported test statistic t = 1.3 for the seed-based finding; cluster/extent thresholds p<0.05 and p<0.01 uncorrected). Conversely, decreases in default mode network connectivity were observed in the ayahuasca group, predominantly centred on the posterior cingulate cortex. Inter-network analyses showed increased connectivity between the salience and default mode networks after ayahuasca compared to placebo (t = 2.5; p<0.02). No significant group differences were found for intra-network connectivity of primary sensory networks (visual, sensorimotor) or their inter-network couplings with salience and default mode networks. Control analyses yielded convergent results: ICA replicated increased anterior cingulate and superior frontal connectivity for the salience network and a trend-level decrease in posterior cingulate connectivity for the default mode network. Adding the difference in mean framewise displacement as a covariate did not materially change the findings. Nonparametric permutation testing (Randomise) also supported the primary results. Exploratory partial correlations (controlling for framewise displacement differences) related subacute connectivity changes to acute HRS scores within the ayahuasca group: increases in salience network connectivity correlated positively with somesthesia (partial r = 0.55, p<0.01), decreases in default mode connectivity correlated negatively with volition (partial r = -0.59, p<0.005), and increased salience–default mode connectivity correlated with affect (partial r = 0.47, p<0.03). These associations were generally absent in the placebo group. Multiple regression across all participants tested interaction effects: the slope relating salience connectivity change to somesthesia differed by group (β = -0.51, t = -3.00, p<0.005), and the slope for affect versus salience–default connectivity also differed by group (β = -0.39, t = -2.64, p<0.02). The regression for volition versus default mode connectivity change found no significant group interaction (β = -0.13, t = -0.88, p = 0.383).

Using a randomised placebo-controlled design with tf-fMRI scans before and 24 hours after dosing, the investigators report that a single ayahuasca session produced subacute increases in anterior cingulate and superior frontal connectivity within the salience network, decreases in posterior cingulate connectivity within the default mode network, and increased coupling between salience and default mode networks. No subacute changes were detected in primary sensory networks, suggesting some specificity to higher-order affective and self-referential systems. The authors interpret the salience network increases as related to interoceptive and socioemotional processing, noting that hubs such as the anterior insula and anterior cingulate play central roles in integrating autonomic and visceral information and mobilising visceromotor responses to salient stimuli. Pasquini and colleagues link their findings to pharmacology: N,N-DMT acts at 5-HT2A receptors and other ayahuasca alkaloids interact with monoamine systems, receptors whose cortical distributions overlap hubs of the salience and default mode networks. They compare the present subacute profile with prior acute and subacute studies of psychedelics and ketamine, which have reported mixed effects on salience and default mode connectivity (acute decreases, entropy changes, or post-session increases depending on compound and timing). Structural work in long-term ayahuasca users showing anterior cingulate cortical thickness increases and decreased thickness in certain default mode regions is cited as context for possible longer-term anatomical-functional relationships. The authors acknowledge several limitations. They note the use of relatively liberal voxelwise cluster‑extent thresholds in parts of the analysis and the small sample size and moderate MRI field strength (1.5 Tesla), which reduce statistical power and may favour false positives. Although they performed several control analyses—ICA replication, permutation testing and motion covariate analyses—to mitigate methodological concerns, the default mode connectivity decreases did not survive more stringent thresholds and thus should be interpreted cautiously. Correlational analyses tying connectivity changes to HRS subscales were exploratory; the volition subscale has known low internal consistency and both groups showed comparable relationships between volition and default mode change, raising the possibility of non-specific or placebo-related effects. Finally, the investigators emphasise that effects after a single session in healthy participants are expected to be modest, and recommend larger, longitudinal studies including clinical populations and multimodal autonomic measures to clarify acute and longer-term impacts on networks supporting interoception, affect and self-referential processing.

The study provides preliminary evidence that a single ayahuasca session produces subacute functional changes in large-scale brain networks involved in interoception, affect and self-referential processing: increased salience network connectivity (notably anterior cingulate), decreased default mode network connectivity (posterior cingulate), and increased salience–default coupling 24 hours post-dose, with these changes correlating with acute subjective somatic, affective and volitional experiences. Primary sensory networks did not show subacute connectivity changes. The authors conclude that ayahuasca may exert sustained effects on mood-related neural circuits, while emphasising the need for further, better‑powered and longitudinal research to confirm and extend these findings.