Why Some Psychedelic Sessions Turn Frightening. A New Brain Study Offers a Clue
Most people who take a psychedelic under clinical supervision come through the session well. Some do not. A 2026 Stanford study in the Proceedings of the National Academy of Sciences offers a partial explanation for that difference. The clue sits in an unexpected place, namely the direction brain signals travel.
Do you work in the ketamine industry? Save thousands on overhead and other fees by joining the first GPO for ketamine clinics and practitioners. There’s no cost to join and zero obligations. Sign up here!
| Key Takeaway | Detail |
|---|---|
| What the team measured | The frame by frame direction of activity in the default mode network, not simply how active it was |
| The main finding | MDMA, psilocybin and LSD each reduced both the strength and the bottom up direction of signal flow |
| Why that surprised the field | A leading model predicted more bottom up signaling under psychedelics, not less |
| The link to experience | Larger reductions tracked with more intense dread of losing the self and loss of control |
| The animal check | Mice given LSD showed the same pattern. Two sedatives did not |
| The main limitation | Small samples and group level averages. The work cannot predict any one person’s session |
Looking for treatment? Find Spravato clinics (which is covered by insurance) and ketamine clinics closest to you as well as other psychedelic therapies in your area.
The Study Measured Direction, Not Just Activity
Brain imaging usually treats activity as a still picture. A region lights up or it quiets down. That approach says little about where a signal came from or where it goes next.
The Stanford team, led by Adam Pines with senior author Leanne Williams, borrowed a tool from computer vision called optical flow. Engineers use it to track motion between video frames. Applied to brain scans, it maps the path activity takes from one moment to the next.
The researchers ran that analysis across four independent datasets. The data covered humans and mice, three drugs, and nine separate drug versus control comparisons.
The Traffic Ran the Other Way
The default mode network is the set of regions tied to self reflection and the ordinary sense of being a person. Psychedelics reliably change how it behaves. What researchers disagreed about was the mechanism.
One influential account holds that psychedelics loosen the brain’s top down expectations. Raw sensory signal then floods upward with less filtering. That model predicts a rise in bottom up processing.
The Stanford analysis found the reverse. MDMA, psilocybin and LSD each weakened both the overall flow of signal and its bottom up directionality. As Pines put it, the team expected one result and instead found the opposite.
Worth noting for anyone following this literature: earlier imaging work had been read both ways. The authors present their result as a clarification of a contested question rather than a demolition of the older view.
The Size of the Change Tracked With Distress
The most useful part of the paper for patients is not the mechanism. It is the correlation.
Participants whose bottom up signaling dropped the most also reported the roughest experiences. They described a more intense dread of losing their sense of self. They reported stronger feelings of impaired control.
That relationship is an association. It does not establish that one caused the other. It does suggest that difficult psychedelic experiences have a measurable signature in the brain rather than arising at random.
The Mouse Data Answers an Obvious Objection
A skeptic could argue that any drug altering consciousness would produce this pattern. The researchers tested that idea directly.
Mice given LSD showed the same attenuation of bottom up flow. Mice given two sedatives did not. The effect appears specific to psychedelics rather than to sedation in general.
Finding the same pattern in mice and humans also suggests something conserved across species. That echoes earlier work showing five psychedelics share a common brain signature.
What the Study Cannot Tell You
The samples are small. The human arms included 14 participants on MDMA, 18 on LSD and 6 on psilocybin, alongside 14 mice.
The work also reanalyzes existing scans rather than testing new patients. Every result is a group level average. Nothing here predicts how a specific person will respond in a specific session.
The authors are direct about this. They measured one directional property of brain activity, and the field still lacks a full account of what bottom up signaling represents.
Why It Matters If You Are Preparing for a Session
Patients often interpret a frightening session as a personal failure. They wonder whether they resisted too much or chose the wrong moment.
This research points somewhere else. Challenging experiences appear to track a physical change in how the brain routes information. That reframing supports the parts of treatment that clinics already emphasize, including thorough screening, preparation sessions, a trained sitter and structured integration afterward.
Readers who want the experiential side of that territory can start with our guide to ego death and what it feels like. For the emotional processing angle, see our coverage of what brain scans reveal about psychedelics and emotion.
Where the Work Goes Next
The obvious next step is individual prediction. Group averages cannot tell a clinician which patient faces a harder session.
Answering that question requires larger samples and scans collected for this purpose rather than borrowed from older trials. Until that work exists, the finding belongs in the category of promising mechanism, not clinical tool.
