Psychedelics reduce bottom-up brain activity in the default mode network

Psychedelic drugs like LSD and psilocybin reduce the amount of bottom-up brain activity flowing into a network associated with self-reflection. These altered trajectories of brain signals occur consistently across humans and mice, offering a biological explanation for how these drugs reshape the mind. The findings were published in the Proceedings of the National Academy of Sciences.

Psychedelics are gaining attention as potential psychiatric treatments for conditions like depression and trauma. To understand their therapeutic benefits and risks, researchers need to know exactly how these substances alter brain function. Prior research has identified that psychedelics affect the default mode network, a collection of brain regions responsible for introspection, daydreaming, and mental rigidity. Many psychiatric conditions involve abnormal activity in this specific network, making it a primary target for new treatments.

The default mode network sits at the top of a processing hierarchy in the brain. Brain activity constantly moves across the surface of the cortex, which is the brain’s outer layer. Information travels from lower-order sensory areas up into higher-order areas like the default mode network, a process known as bottom-up processing. The reverse movement, from higher cognitive regions down to sensory regions, is called top-down processing.

In a healthy brain, bottom-up processing updates our internal models with new information from the outside world. Top-down processing uses our past experiences and expectations to make sense of that incoming sensory data. An imbalance between these two directions of information flow can lead to psychological distress or perceptual errors.

Past imaging studies typically measured brain activity in static regions, treating the brain as a set of fixed locations. This approach ignores the continuous, wave-like movement of signals across the brain’s surface. Analyzing static brain regions is similar to measuring the total rainfall in a single county without tracking the movement of the storm system on a weather radar. By treating the brain as a collection of isolated points, traditional research methods might obscure the true dynamics of how different brain areas communicate over time.

Adam R. Pines, Leanne M. Williams, and their colleagues at Stanford University wanted to observe these moving signals directly. They adapted an analytical technique called optical flow, which tracks the frame-by-frame movement of pixels in a video. The team used this mathematical approach to track the direction and size of brain activity waves moving across the cortex.

The researchers first analyzed a small study of 14 human volunteers who received MDMA. They compared functional magnetic resonance imaging scans taken after the participants took MDMA to scans taken after a placebo pill or no drug at all. The optical flow analysis revealed a consistent pattern across the subjects.

The team found that MDMA reduced the overall magnitude of brain waves moving through the default mode network. The drug also reduced the proportion of signals traveling in a bottom-up direction into the network.

Next, the team examined a second small study involving six human participants who ingested psilocybin. The researchers compared brain scans from the psilocybin sessions to baseline scans and to an active placebo condition. For the active placebo, participants received a dose of the stimulant methylphenidate. This stimulant mimics the physical arousal of psilocybin without the psychedelic effects.

As with MDMA, psilocybin reduced the overall magnitude of propagating brain waves in the default mode network. The reduction in bottom-up directionality was also present but was not statistically significant. The researchers suspect this weaker result occurred because the effects of psilocybin lingered for days. This persistence may have artificially altered the baseline scans taken shortly after the drug sessions.

In a third small study, 18 human volunteers received an intravenous infusion of LSD. The researchers compared their brain activity to scans taken after a saline placebo infusion. Consistent with the other substances, LSD reduced both the magnitude of the moving brain waves and the proportion of bottom-up signals entering the default mode network.

To ensure these effects were not unique to humans or specific to magnetic resonance imaging, the team looked at a small study of 14 mice. They used widefield calcium imaging, a technique that directly records the physical activity of brain cells with high resolution. The mice received LSD, a sedative called diazepam, or a different sedative called dexmedetomidine.

Like the human participants, mice given LSD showed reduced magnitude and bottom-up flow of brain activity in the default mode network. Diazepam, an anti-anxiety medication, also reduced bottom-up signals, though to a lesser extent than LSD. Dexmedetomidine produced the exact opposite effect, increasing the proportion of bottom-up signals entering the network. These animal results confirmed that the optical flow technique measures actual changes in nerve cell activity rather than artifacts of blood flow.

The researchers also investigated whether these changes in moving brain activity related to the subjective experiences reported by the human volunteers. In the MDMA study, participants who experienced the greatest reduction in bottom-up processing also reported the most intense feelings of impaired control. They also reported a greater dread of ego dissolution, which is a fearful reaction to losing one’s sense of self.

This psychological correlation suggests that suppressing bottom-up information flow too much could trigger the negative experiences sometimes associated with psychedelics. Bottom-up processing normally grounds individuals in their immediate sensory environment. A severe drop in this grounding input might leave individuals feeling detached or at the mercy of their unanchored internal thoughts.

Some psychiatric conditions, such as ruminative depression, are characterized by excessive automatic thoughts and abnormal bottom-up signaling. Reducing bottom-up flow might explain why psychedelics offer relief for some patients with depression. The drugs could temporarily quiet the intrusive signals feeding into the default mode network.

Individuals at risk for psychosis already suffer from impaired bottom-up processing. Giving psychedelics to these vulnerable patients could exacerbate their symptoms. The drugs might overpower their ability to update internal beliefs with external reality.

The sample sizes across all four datasets are small, which limits the ability to account for individual differences in how people react to these drugs. Larger datasets will be necessary to map out how different psychedelics affect people with varied biological backgrounds. Expanding the number of participants could reveal more subtle alterations to brain activity flow that this initial analysis missed.

The optical flow analysis focused exclusively on the outer surface of the brain. Psychedelics also heavily influence deep brain structures, but tracking moving signals between these buried regions and the surface remains computationally out of reach for this specific technique. The imaging technologies used in human participants also have limits in their spatial and temporal resolution. These equipment limitations mean researchers might miss faster changes in brain signaling.

Finally, the studies relied entirely on healthy volunteers and laboratory mice. Future research will need to test these observations in clinical populations to determine how these changes in brain activity relate to therapeutic outcomes. Understanding the exact mechanisms of psychedelic action could help doctors identify which patients are most likely to benefit from these treatments and which patients might be harmed by them.

The study, “Psychedelics disrupt hierarchical cortical propagations in the default mode network of humans and mice,” was authored by Adam R. Pines, 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.