A Stanford University team has used a novel optical‑flow approach to track moving brain activity and found that psychedelic drugs reduce the magnitude of propagating brain waves and curb bottom‑up signals entering the brain’s default mode network. The work, conducted in humans and mice and published in the Proceedings of the National Academy of Sciences, pinpoints a consistent pattern across MDMA, psilocybin and LSD that may illuminate how these substances reshape subjective experience.
The default mode network sits at the top of a processing hierarchy in the brain, with information typically travelling from lower‑order sensory regions up to higher‑order areas before feeding back through top‑down pathways. Bottom‑up processing updates internal models with new sensory information, while top‑down processing uses past experience to interpret that input. When these flows become imbalanced, psychological distress or perceptual distortions can arise.
Traditionally, researchers studied static brain regions rather than the dynamic waves that traverse the cortex. This static view risks overlooking how signals move over time across the brain’s surface. The Stanford team adapted an analytical technique originally used for video analysis — optical flow — to measure the direction and size of propagating brain activity.
In the first human study, 14 volunteers received MDMA and had brain activity scanned afterwards. Compared with placebo or no drug, MDMA reduced the overall magnitude of propagating brain waves within the default mode network and lessened the proportion of signals arriving from bottom‑up directions.
A second small study involved six participants who took psilocybin. Their brain scans were compared with baseline and with an active placebo condition using methylphenidate. Like MDMA, psilocybin reduced the overall magnitude of moving brain waves in the default mode network, and the reduction in bottom‑up directionality was observed but did not reach statistical significance. The researchers suggest lingering drug effects may have influenced baseline measurements.
A third human study with 18 volunteers given LSD found a similar pattern: a fall in both the magnitude of propagating brain waves and the amount of bottom‑up input to the default mode network, compared with saline placebo.
To test whether the findings extend beyond humans, the team studied 14 mice using widefield calcium imaging, a method that records neuronal activity directly. The mice received LSD, the sedatives diazepam or dexmedetomidine. LSD produced the same reduction in magnitude and bottom‑up flow as seen in humans, while diazepam also reduced bottom‑up signalling to a lesser extent. Dexmedetomidine produced the opposite effect, increasing bottom‑up inputs to the network. The cross‑species results bolster the view that the optical flow measure reflects genuine neural activity rather than artefacts of blood flow.
Researchers also explored whether the observed changes related to subjective experiences. In the MDMA cohort, those showing the greatest decline in bottom‑up processing reported the strongest sense of impaired control and greater dread of ego dissolution, a fear of losing one’s sense of self. The team notes that overly strong suppression of grounding sensory input could contribute to the negative experiences sometimes associated with psychedelics.
The findings carry potential implications for therapy. By temporarily quieting intrusive bottom‑up signals entering the default mode network, psychedelics might alleviate certain depressive and trauma‑related symptoms. However, the researchers cautioned that reducing bottom‑up information too much could be problematic for some patients, particularly those at risk of psychosis who already struggle with integrating external reality with internal beliefs.
Limitations of the study include small sample sizes across the four datasets and the focus on the brain’s outer cortex, with deep structures and interregional connections not fully captured by the method. Imaging technologies used with human participants also constrain spatial and temporal resolution, potentially missing faster dynamics. All experiments involved healthy volunteers and laboratory mice, so further work is needed to understand clinical implications and which patients may benefit or be harmed.
The study, titled “Psychedelics disrupt hierarchical cortical propagations in the default mode network of humans and mice,” was led by Adam R. Pines and Leanne M. Williams of Stanford University, with collaborators including Xue Zhang, John Kochalka, Sam S. Vesuna, Isaac V. Kauvar, Divya Rajasekharan, T. Rick Reneau, Teddy J. Akiki, Laura M. Hack, Joshua S. Siegel, and Leanne M. Williams.
