A treatment changes the brain. But what happens next may depend just as much on what surrounds the person receiving it: the music they hear, what they see, what they are thinking about, and the state they are already in. 

In this episode of Neural Compass with Mark Jacobstein, we explore what happens when the brain is treated not as a static collection of regions, but as an adaptive system whose trajectory can be perturbed, measured, and potentially guided. 

Mark speaks with Adeel Razi, whose work at Monash University brings together engineering, machine learning, computational neuroscience, and neurobiology. His recent research examined brain activity under psychedelics and challenged the idea that these compounds simply introduce greater disorder. Using methods that track changes within individual brains over time, his team found structured trajectories beneath that apparent disorder, with the organization of those trajectories related to the depth and quality of the psychedelic experience.

This conversation examines why “set and setting” can be observed in brain dynamics rather than treated only as a therapeutic convention; why perturbing a pathological brain state is not enough unless there is some way to influence where the system goes next; how closed-loop treatment could move psychiatry toward adjusting dose and intervention according to an individual’s response; and why Razi believes psychiatric disorders may be better understood as trajectories involving the brain, body, and environment rather than fixed states.

Neural Compass is a podcast from Jimini Health, which builds clinically embedded AI for behavioral health.

If you found this conversation valuable, please leave a rating or review on Apple Podcasts.

About the Guest: 

Adeel Razi is a computational neuroscientist at Monash University, where he leads research combining engineering, physics, machine learning, and neurobiology to understand how brain networks behave and reorganize. His work spans dynamic causal modeling, psychedelic-induced brain dynamics, biologically inspired artificial intelligence, and the study of adaptive neural systems.

Before moving into neuroscience, Razi worked in wireless systems and chip design at Broadcom, applying concepts from information theory and multiple-input, multiple-output systems. He later brought many of those same quantitative ideas into neuroscience, using them to study causal information flow between brain regions and how biological systems respond to perturbation. 

[00:16:26] Finding Structure Beneath Psychedelic Disorder
Dr. Adeel Razi explains that psychedelic brain activity is not simply random or disorganized. His research found structured neural trajectories beneath the apparent disorder, with stronger organization linked to deeper subjective experiences. The findings also provide evidence that factors such as music, meditation, and other elements of set and setting actively shape brain dynamics.

[00:20:46] How Psychedelics Could Free The Brain From Rigid States
Dr. Razi describes psychiatric conditions as rigid brain trajectories, using the analogy of a ball trapped in a deep valley. Psychedelics can act as a perturbation that temporarily flattens that landscape, allowing the brain to escape an unhealthy pattern. Therapy and context may then help guide it toward healthier, more stable trajectories.

[00:41:00] What A Dish Of Neurons Can Teach Us About AI
Dr. Razi discusses DishBrain, a closed-loop biological system capable of learning to play Pong. His team is now studying whether these neurons can learn multiple tasks without forgetting earlier ones - a challenge artificial neural networks often face. By observing biological neurons during continual learning, the researchers hope to uncover principles that could eventually inform new AI models.