Proposed Mechanisms vs Proven Clinical Outcomes

Three different kinds of finding

News stories about psychedelics often move from "the drug changes the brain" to "so it treats the condition". Those are separate steps, and each needs its own evidence.

A biomarker change is not a clinical outcome. A mechanism that is plausible is not a mechanism that is proven.

Evidence labels used on this page

Insufficient evidence means studies do not yet show whether something is true. Mixed means studies disagree. Preclinical means cells or animals only, which this site never presents as results in people. All labels are listed in how we label evidence.

Example 1: the default mode network

The US National Institute on Drug Abuse (NIDA) says classic psychedelics temporarily disrupt communication between brain regions, including the default mode network, which is most active during self-focused thought, and that this may account for why thoughts become less self-centered. [1] A 2023 systematic review found consistent acute disruption of connectivity within this network across LSD, psilocybin and ayahuasca studies. The same review states that it is unclear how central the network is to the therapeutic potential of these drugs. [2]

Label: the imaging change is established as an acute finding in those studies; its role in any treatment benefit is insufficient evidence.

Example 2: 5-HT2A receptor activation

A 2026 review says activation of the serotonin 5-HT2A receptor is critical for the psychopharmacological effects of psychedelics, yet it is uncertain which signaling mechanisms are necessary for rapid-acting and durable therapeutic effects. [3] A review of psilocybin and depression says the mechanisms behind its antidepressant and psychedelic effects are currently unknown. [4] See Serotonin Receptors in Plain Language.

Example 3: the "mystical experience"

A 2022 systematic review of 12 studies of psilocybin, ayahuasca or ketamine looked at whether the intensity of a mystical-type experience was linked to symptom reduction. Ten of the twelve found a significant association, whether as correlation, mediation or prediction. The authors note that most studies were small, lacked diversity, and that six of twelve were open-label and therefore open to bias. [5]

Label: promising but limited for an association. An association inside a trial does not prove that the experience causes the benefit, and it does not show that the experience is required.

Why blinding and expectancy matter for mechanisms too

A review of randomized trials argues that treatment effect sizes in psychedelic trials are likely over-estimated because participants can tell whether they received the drug and because expectations are high. [6] If the clinical outcome itself is uncertain, any mechanism proposed to explain it is uncertain too. See Context and Expectancy in Psychedelic Research.

The FDA's 2026 guidance adds that the contribution of the psychotherapy component to any efficacy seen with psychedelic drug treatment has not been characterized. [7] So "the drug" and "the therapy" are not yet cleanly separated in the evidence.

A simple checklist for any claim

1. Is the finding from cells, animals or people? Animal or cell results are preclinical.
2. Is it a mechanism, a biomarker or a clinical outcome?
3. Was the result in a randomized trial with a comparator, or a small open-label study?
4. How long were people followed, and what adverse events were recorded?
5. Is it one substance and one condition, or being generalized to "psychedelics"? See Why Different Substances Cannot Be Treated as Equivalent.

For trial-reading detail see How to Read a Psychedelic Clinical Trial. The historical and neuroscience background is in Psychedelics: History and What Neuroscience Shows.

Education only

This page explains how to read evidence. It does not recommend any substance or tell anyone to treat a condition with it. Talk with a clinician about any health condition: talking with a health professional.