Draulio B. Araújo’s research with ayahuasca began two decades ago with questions about consciousness, not depression. Over time, his group moved from neuroimaging studies to clinical research, including a placebo-controlled trial in treatment-resistant depression. Its subsequent work with inhaled DMT has brought a different question into focus: what would a controlled trial need to establish before an early antidepressant signal could be considered evidence of efficacy?
Araújo is a professor at the Brain Institute of the Federal University of Rio Grande do Norte and directs its Centro Avançado de Medicina Psicodélica, in Natal, Brazil. He described that progression in his own words, including what a positive result would still need to prove and what would change his mind.
Where the research began
Araújo’s group began working with ayahuasca in 2006, a decade before its first randomized clinical trial in depression.
“At first, our main interest was not depression, but consciousness: we wanted to understand what happened in the brain during the acute ayahuasca experience using neuroimaging,” Araújo said.
“Brazil offered a unique setting for this work. Religious use of ayahuasca is legally permitted, which gave us access to experienced users who were already familiar with its effects. This was extremely valuable in early neuroimaging studies, because participants could remain calm and collaborate with demanding procedures such as MRI while under the acute effects of ayahuasca.
But those early studies also taught us something more fundamental. We quickly realized that we were not simply studying a psychoactive substance.”
“For many people, ayahuasca is a sacrament, and in Indigenous traditions it belongs to much broader systems of knowledge, healing, spirituality and social life.”
— Draulio B. Araújo
“That made us increasingly aware of the limits of reducing ayahuasca to a plant preparation, or to DMT and beta-carbolines,” Araújo continued. “Over time, the research moved from basic neuroscience toward clinical questions. We first conducted an open-label study in depression, which showed a rapid antidepressant signal, and later a randomized placebo-controlled trial in treatment-resistant depression.
Looking back, I think the main contribution of this early phase was broader than any single clinical result. It showed that ayahuasca could be studied rigorously with neuroscience and clinical research methods, but also made clear how difficult it is to separate pharmacology from subjective experience, expectation, context and cultural meaning.
The randomized trial strengthened the evidence for a rapid antidepressant effect, but many questions remained open: who benefits, how long the effect lasts, which components of ayahuasca are responsible, and how much of the outcome depends on the drug itself versus the broader experience in which it occurs.”
That trial, led by Fernanda Palhano-Fontes and Araújo, was published in Psychological Medicine in 2019. It gave a single dose of ayahuasca or placebo to 29 patients with treatment-resistant depression and found a rapid reduction in depression severity that held for a week, the first placebo-controlled clinical evidence for the compound in this population.
Why the work moved to DMT
“The move from ayahuasca to DMT was, in many ways, a move toward greater experimental control and less cultural complexity,” Araújo said. “DMT is a molecule and not a sacrament.
With ayahuasca, we were working with a complex preparation, a long-lasting experience and, as I mentioned before, something that is also a living sacrament embedded in cultural and spiritual traditions. From an academic perspective, that complexity is extremely rich but also difficult to control.
DMT allowed us to ask more precise questions. We could work with a defined molecule, a controlled dose, a much shorter experience, and a clearer relationship between administration, subjective effects, physiological changes and clinical outcomes. That makes it much easier to study dose, timing, safety, neurophysiology and potential antidepressant effects in a rigorous clinical-trial framework.
At the same time, something is clearly lost. Ayahuasca is not simply oral DMT. It includes other active compounds and carries a cultural, ritual and symbolic dimension that isolated DMT does not reproduce.
So I see the two approaches as complementary rather than competing. Ayahuasca taught us how complex psychedelic experiences can be; DMT gives us a way to experimentally dissect part of that complexity with much greater precision.”
What the open-label study could not show
“For me, the most informative finding was that inhaled DMT could be administered safely and feasibly to patients with treatment-resistant depression, and that we observed a very rapid clinical signal,” Araújo said.
“The speed of the response was particularly striking, including changes in depressive symptoms and suicidal ideation within a very short time after dosing. That is important because it indicates there is a signal worth pursuing.
But this was a small, open-label study, so we have to be very careful about interpreting the magnitude of the effect. Without a control group, we cannot separate the effect of DMT from expectancy, the clinical setting, the attention provided by the research team, participant selection, or natural fluctuations in depressive symptoms.
So I see Phase 2a primarily as a proof-of-concept study. It told us that the approach was feasible, appeared safe in this small sample, and produced a clinical signal strong enough to justify a more rigorous trial. It did not establish efficacy.”
The published Phase 2a trial he is describing, run by his group and appearing in Neuropsychopharmacology in 2025, enrolled 14 patients with treatment-resistant depression and gave them ascending inhaled doses of DMT, 15 milligrams followed by 60, with no control group. The published results reported a response rate of 85.71 percent (12 of 14 patients) and remission in 57.14 percent (8 of 14) seven days after dosing. By three months, response was recorded in 57.14 percent (8 of 14), and remission in 35.71 percent (5 of 14). The day-seven and three-month figures describe two different points in an open-label follow-up with no comparison group, not a stable plateau, and should not be read as showing that every participant who improved at day seven maintained that improvement unchanged through month three.
The controlled trial that will test the early signal
That more rigorous test is the subject of Araújo’s ongoing Phase 2b trial, registered as a randomized, double-blind study of inhaled DMT in major depressive disorder (NCT07562191) and, as of this writing, recruiting across five sites in Brazil.
“The main challenge in psychedelic research is that blinding is inherently difficult,” Araújo said. “Participants often know whether they have received an active psychedelic, which means expectation can become part of the treatment effect. One way to reduce this problem is to use a very low dose of DMT as an active comparator rather than an entirely inert placebo. The idea is to create at least some perceptible effects in the control group, making treatment allocation less obvious.
We also need larger samples, standardized clinical support, independent raters, and preferably more than one study site. Most importantly, improvement from baseline is not enough. A convincing trial has to show that patients receiving the active dose improve more than those receiving the control condition.
I also think the field has sometimes focused too narrowly on depression rating scales. Of course these are essential, but we should also ask whether patients are actually living better: whether their quality of life, functioning, sense of well-being and satisfaction with life improve. In some of our follow-up work, these broader measures have provided information that was not fully captured by symptom scores alone.
For me, that is the critical step. Early studies can tell us that a treatment is feasible and that there is a signal. A controlled Phase 2b study has to tell us whether that signal survives a much more rigorous test, and whether it translates into meaningful changes in patients’ lives.”
An active low dose can make guessing harder without guaranteeing that no one guesses; Araújo does not claim the design solves the blinding problem outright, only that it narrows it. The registration matches his description in its structural features: 140 participants randomized to a higher-dose regimen (15 milligrams followed by 60) against a lower-dose active comparator (1 milligram followed by 4), with change in MADRS score from baseline to day 7 as the prespecified primary outcome between the two arms. Its registered secondary outcomes are extensive, covering suicidality (including the Columbia-Suicide Severity Rating Scale and a dedicated MADRS item), safety and vital signs, anxiety and mood, subjective psychedelic effects, and quality of life and functioning (including the WHOQOL-BREF and the WHO Disability Assessment Schedule), tracked through twelve months. The registration does not include neuroimaging, fMRI, or EEG among its outcome measures. None of this is a reported result yet: the trial is enrolling, not reporting, and what it finds is exactly what the earlier open-label study could not tell him.
Why a brain scan is not a mechanism
“I have become increasingly cautious about assigning a therapeutic meaning to any single neuroimaging finding,” Araújo said.
“Our earlier studies with ayahuasca showed changes involving the default mode and salience networks, including effects that persisted into the following day. More recently, the international mega-analysis in which we participated, combining data from multiple psychedelic compounds and laboratories, provided a broader picture. One of the more reproducible observations appears to be a reorganization of communication between networks that normally occupy different levels of the cortical hierarchy, particularly increased interaction between higher-order association networks and sensory or sensorimotor systems. Interestingly, some of the widely discussed claims about a generalized ‘disintegration’ of intrinsic brain networks appear less robust when larger datasets are considered.
These observations are fascinating for understanding the psychedelic state, but they are not yet mechanisms of antidepressant action.
To make that transition, we need longitudinal experiments in patients in which neural measurements are obtained before, during and after treatment within randomized controlled designs. A candidate neural change should precede clinical improvement, covary with the treatment effect, and, ideally, statistically mediate the difference between the treatment and control groups.
Even that would establish a strong association rather than definitive causality. A stronger causal test would require experimentally manipulating the proposed mechanism. For example, pharmacologically altering the receptor system or neural process in question, and determining whether this also modifies the subsequent clinical effect.
One of the major challenges for the field is precisely to distinguish the neural correlates of having an extraordinary conscious experience from the neural processes that actually contribute to recovery.”
The international mega-analysis he refers to was published in Nature Medicine in 2026; its finding that network-disintegration claims look less robust at larger scale matches his account. The longitudinal, mechanism-testing design he describes is distinct from his currently registered Phase 2b trial, which does not include the neural measurements that sequence would require.
Who benefits, and whether it lasts
“Larger studies should move beyond the question of the average change in depression score,” Araújo said.
“We need to understand whether response varies according to baseline severity, duration and degree of treatment resistance, suicidal ideation, anxiety and other comorbidities, concomitant medication, previous psychedelic exposure and potentially biological or neurophysiological characteristics.
The outcomes should also extend beyond symptom scales. Remission of depressive symptoms is important, but so are suicidality, anxiety, quality of life, functioning, social relationships and the ability to return to ordinary activities. A patient can improve substantially on a rating scale without necessarily having recovered their life.
Duration is equally important. Measurements at one day and one week are useful for identifying rapid effects, but follow-up over months is needed to characterize relapse, persistence of remission and the need for retreatment. It is also important to document changes in concomitant treatments during follow-up, because otherwise an apparent long-term effect may actually reflect subsequent interventions.
For me, a meaningful sustained therapeutic effect would therefore involve convergence: improvement in depressive symptoms that persists over time and is accompanied by improvements in functioning, quality of life and clinically important outcomes such as suicidal ideation.”
What would change his mind
“What would most increase my confidence would be replication in adequately powered, randomized, multicenter studies showing that DMT produces a clinically meaningful advantage over a credible active comparator, assessed by independent raters, and that at least a meaningful proportion of this benefit persists beyond the immediate post-treatment period,” Araújo said.
“I would be particularly encouraged if improvement extended beyond depression rating scales to functioning, quality of life and suicidality, and if the safety profile remained favorable as the number and diversity of exposed patients increased.
Conversely, I would substantially revise my interpretation if the large effects seen in early open-label studies largely disappeared under rigorous controlled conditions. The same would be true if improvements proved to be extremely transient, were strongly explained by expectancy or treatment allocation effects, were restricted to a very narrow and highly selected population, or were accompanied by safety problems that only become apparent in larger samples.
A negative result in a well-designed trial would therefore not be a failure of the research program. It would answer exactly the question that the early studies were unable to answer.
Our current position is that inhaled DMT is a promising experimental intervention. The emphasis should remain on both words: promising, because the clinical signal is substantial enough to justify larger studies; experimental, because those larger controlled studies are precisely what we still need before drawing conclusions about its therapeutic role.”