EXPRESS: Global Increases in Brain Glucose Metabolism Following Acute N,N-Dimethyltryptamine and Harmine Administration in Healthy Volunteers: A randomised [F]FDG-PET Study

Serotonergic psychedelics such as DMT are understood to alter perception, emotion, and self-experience largely through 5-HT2A receptor agonism. DMT is of particular interest because, when combined with a monoamine oxidase A inhibitor such as harmine, it can produce an ayahuasca-like psychedelic state lasting several hours. However, although many neuroimaging studies have examined functional connectivity, brain activity, and perfusion during psychedelic intoxication, the effects of these drugs on cerebral glucose metabolism have been far less studied. The paper notes that only two human FDG-PET studies, both with psilocybin almost three decades earlier, had investigated this question, leaving it unclear whether the metabolic findings generalise to DMT or ayahuasca-like formulations. Egger and colleagues therefore set out to test whether a novel oromucosal DMT plus harmine formulation changes brain glucose metabolism in healthy volunteers. Their primary hypothesis was that global cerebral metabolic rate for glucose consumption would increase under the active drug compared with placebo, rather than decrease. They also aimed to examine whether global metabolism related to plasma drug levels, subjective intensity of the acute experience, regional cortical changes, resting-state network effects, and 5-HT2A receptor density maps.

The researchers conducted a single-blind, placebo-controlled, within-subject crossover FDG-PET study in healthy male volunteers. Twenty men were recruited, but six withdrew before completing both sessions, leaving 14 participants for the final analysis. Participants were aged 25-45 years, had prior psychedelic experience, and had not used psychedelics in the preceding three months. The text states that full inclusion and exclusion criteria were in supplementary material. The study was approved by ethics committees, registered at ClinicalTrials.gov, and all participants gave written informed consent. Each participant completed two study days, one with placebo and one with active treatment, in random order. The active condition consisted of buccal administration of DMT plus harmine in three equal doses spaced 20 minutes apart, totalling 90 mg DMT and 120 mg harmine, given as orodispersible tablets. The scan was scheduled to capture peak effects, with PET acquisition beginning about 100 minutes after the first dose and lasting roughly 70 minutes. Participants rested with eyes closed and no music during the scan, although one of three background music playlists was played during the pre-scan and post-scan periods. Vital signs, subjective ratings of drug intensity and related experiences, and venous blood samples for pharmacokinetic analysis were collected across the session. Standard questionnaires including the Mystical Experience Questionnaire and the 5-Dimensional Altered States of Consciousness questionnaire were completed later in the day. PET and MRI preprocessing used standardised pipelines. An image-derived input function was extracted from the descending aorta rather than using arterial blood sampling. Glucose metabolism was quantified with Gjedde-Patlak graphical analysis to estimate unidirectional blood-brain clearance and then convert this into cerebral metabolic rate for glucose consumption using blood glucose measurements. The authors also ran a complementary two-tissue compartment model analysis on region-of-interest data. Surface-based analyses were performed across the cortex, and network-level analyses averaged values within the seven Yeo resting-state networks. Plasma DMT, harmine, metabolites, and serotonin were quantified by UHPLC-MS/MS, and exploratory pharmacokinetic analyses focused on area under the concentration-time curve from first to last sample. The primary outcome was tested with a one-sided paired t-test, while regional and network-level effects and exploratory correlations used paired t-tests, Pearson correlations, and false-discovery-rate or random-field-theory thresholds where applicable.

Global cerebral glucose metabolism increased under DMT plus harmine relative to placebo. The active condition produced a 12% increase in global cerebral metabolic rate for glucose consumption, with mean values of 16.3 versus 14.5 µmol hg−1 min−1 and a mean difference of 1.8. The difference was statistically significant on the one-sided paired t-test (t = 2.57, df = 13, p < 0.05), with a reported effect size of Cohen’s d = 0.64. The authors note that the individual paired data suggested a fairly consistent increase across participants. In exploratory correlational analyses within the active condition, global glucose metabolism was significantly positively correlated with harmine exposure, with r = 0.62 (p = 0.019). The correlations with DMT exposure and with mean subjective intensity during the PET window were positive but not statistically significant: r = 0.35 for DMT area under the curve and r = 0.39 for subjective intensity. The extracted text does not report a significant association between global metabolism and these latter measures. Regional analyses showed widespread increases in glucose metabolism across large parts of the cortex. Effects survived more stringent thresholds in regions belonging to the default mode network, frontoparietal network, and salience network. Network-level analyses indicated increased metabolism in attentional networks, specifically dorsal attention and salience, as well as in higher-order transmodal networks, namely the frontoparietal and default mode networks. The paper also reports no significant differences between placebo and active condition in the correlations between network-wise glucose metabolism and 5-HT2A receptor density maps. Regarding tolerability, no serious adverse events were reported. Two participants experienced transient nausea with vomiting after DMT plus harmine, but this resolved before the PET scan.

Egger and colleagues interpret the findings as first-in-human molecular imaging evidence that an ayahuasca-inspired DMT plus harmine formulation acutely increases cerebral glucose metabolism. They argue that the global rise in FDG uptake, together with the spatial pattern of stronger effects in attention and transmodal association cortices, suggests a state of heightened cerebral energy demand during the psychedelic experience. They position this as extending earlier functional imaging work on DMT and ayahuasca by providing a direct metabolic measure rather than an indirect signal of activity or perfusion. The authors compare the magnitude of the global metabolic increase, about 12%, with the approximately 20% increase reported in the historical psilocybin FDG-PET study, and with similar global increases seen in ketamine studies. They also note that the right-hemisphere predominance and frontal/transmodal pattern resembles prior findings with psilocybin, ketamine, mescaline, and ayahuasca, and they link this to theories of altered hemispheric hierarchy and increased entropy, meaning a less rigid and more flexible brain state. They suggest that the elevated glucose metabolism may reflect increased neuronal activity, but they also acknowledge an alternative explanation: altered metabolic coupling or mitochondrial uncoupling, rather than a simple rise in activity. They discuss preclinical work suggesting possible intracellular and mitochondrial mechanisms for DMT, while noting that such mechanisms do not fully explain earlier psilocybin findings. The paper also discusses the exploratory correlation with harmine exposure. The authors note that global metabolism correlated with harmine area under the curve but not with DMT exposure or subjective intensity, yet they caution against interpreting this as evidence that harmine drives the effect. They emphasise that the study was not designed for detailed pharmacokinetic modelling, that the sample was small, and that the observed correlation could reflect limited power or measurement constraints. They further note that harmine’s main role here is likely to inhibit MAO-A and enable oral or buccal DMT bioavailability, rather than act as the principal cause of the metabolic increase. Key limitations highlighted by the authors include the small sample, especially for correlational analyses; the lack of detailed pharmacokinetic sampling across the full scan window; the exclusively healthy, white, male sample, which limits generalisability; imperfect blinding because many participants correctly guessed the active condition; and the use of an image-derived input function instead of arterial sampling, although they state this approach has shown good concordance with arterial methods in recent work. They conclude that future studies should clarify the mechanism underlying the metabolic changes and better define the contribution of 5-HT2A receptor agonism to both the cerebrometabolic and subjective effects of DMT plus harmine.

The authors conclude that acute administration of the novel oromucosal DMT plus harmine formulation robustly increases cerebral glucose metabolism, particularly in attentional and higher-order transmodal networks. They suggest this may reflect a more entropic and less hierarchically constrained brain state, potentially relevant to the breakdown of rigid patterns of thought and emotion. They also state that future research should determine the causal mechanism behind the metabolic effect and better establish the role of 5-HT2A receptor agonism.