Adam Gazzaley — Brain Optimization and The Future of Psychedelic Medicine

How Gazzaley builds security into high-risk psychedelic research at UCSF

Gazzaley's lab moved from video game-based cognitive interventions to psychedelic research remarkably fast. He builds systematic security around risky protocols so the science can proceed. Psychedelics were never once presented during his full neurology residency at UCSF.

"How come never during all of that training was psychedelics ever presented in an academic manner as being a potential tool? Not once, never."

Psychedelic therapy protocol: eye mask, curated music, internal focus

The standard clinical psychedelic protocol involves ingesting the substance, wearing an eye mask for internal experience, and listening to curated music that shifts across the session. The research raises challenging questions about how to measure and optimize these subjective experiences.

"And you would then ingest this and have usually a blindfold on, an eye mask, so it was more internal, you weren't seeing things, and you'd be listening to music. And that music would vary across sight. It may not even be music that you actually like, but that is the design right now. And then after, you would have the support during the session. And after the session, you would have several other therapist-led meetings to help process and integrate what you went through and then follow up. And that's the basic design. And it's a good one. And it's one that we've been really using for decades, even in the 60s, similar designs to this. And it's good methodology. And it's getting better and better. What I see missing there in understanding the middle part is the first thing is we don't really know what's going on with a participant during the treatment itself. You're sort of a black box. Now, an extremely good therapist that has done this for 40 years is picking up subtle verbalizations or body movements or changes in your facial expression. And they're able to protect you if you need it or help just sit with you as you go through some challenges. And that's as good as it gets. But that's not really scalable at any level because you can't have as many experts as there are people in need of a treatment like this. So that's the first opportunity. And I'm going to talk about that in a moment. And the other opportunity is a blindfold and a mask. The best option is music really ideal? And what type of music? And what about all the other senses that we have, how we smell and feel and other aspects of hearing that are not being used during these treatments. Unknown. So used in the real world of therapy, but not really studied. We don't know who they're best in. So that's the second domain. So let's split those off. So the first domain is recording. So we have a system at UCSF we call multimodal biosensing. And my hypothesis is that with enough sensor technologies, not just looking at brain activity, but looking across the whole spectrum of physiology and behavior. So high density EEG, facial expression, electrodermal activity, heart rate and its variability, looking at facial expressions, body movements, doing experience sampling, either through vocalizations or using joysticks so that there is some degree of communication. My hypothesis is that not one single of those signals are going to tell us about the state of that individual during a treatment. But if we can collect them rapidly in real time and integrate them, use machine learning approaches so that we can look for the patterns of what's the most meaningful data, we will be able to plot the experience landscape of an individual traveling through a treatment like this. We won't know the content. We won't necessarily know that they're picturing the details of a traumatic event in their lives, but I believe that we will be able to tell their level of arousal, the valence of it, is it positive or negative? So putting that together, their stress, the state of stress, their attention, whether their attention is internally directed or externally directed, as well as hopefully even a sense of their awareness of what's going on. So real-time state recording as implemented through multimodal biosensing. That's a very high-tech endeavor, but it is possible right now. It takes a real multidisciplinary team. So you're talking neuroscientists, as well as people on the clinical side, biosensor technology experts, signal processing, machine learning. It is a real challenge and it has not been accomplished yet anywhere in the world at any level that allows us to really track state. So that's the first challenge that we want to really tackle is using technology to understand the state of the person while they go through a treatment. The other side of it is starting to manipulate all of these levers of contexts of the setting and the set. So how do we position them through different information that presented different environments before the treatment? And then during the treatment, not just think about music with eyes closed. Sure, we will do that. That is the standard approach right now. But what is it like to look around you, to see either light or abstract views or views of nature, to smell a forest when you see it and when you hear it, what we call multi-sensory integration, to feel low-frequency vibrations through your bodies as you enter into these different environments, all of that is unexplored. And so by putting these things together, we start understanding the journey, the journey within the treatment itself, how these little, not little to the individual, but these discrete moments in time, these micro experiences along the way, how they're influenced by the context of the environment that we're able to control around the participant, and then how all of these elements add up to an outcome, to a sense of well-being or joy or mystical perspective that occurs after. So how do we connect the events that occur during the treatment with the outcomes? All of that is unexplored, and that's basically what we're going to be focusing on. There is so much room for experimentation. So many variables. There's a lifetime there. I always say to Robin, I was like, we will die before all that that you just heard is figured out. There's so much. It's incredibly exciting to me for many reasons. And just to give a snapshot of personal experience here that raises questions, I think, related to those that you'll be exploring in a very methodical, tech-enabled way so that you can actually measure, capture, and quantify these discrete events, discrete changes. I recall the first time I did a five-session series of infusions. This is intravenous infusions of ketamine. And in this case, I wasn't using it because I was suffering from acute depression nor chronic pain. Those are two indications. But I wanted to be able to speak to the effects and possible side effects, after effects of ketamine treatment if I were asked, which I knew I would be asked. And what really stuck out to me, among other things, in this ketamine treatment, that it was the first time I had been seated in a chair and that video had been put on for me. And I was asked to select a DVD. So it was very Yeah, retro in that sense. But, And I was able to choose a DVD that I then used as a constant throughout my sessions. I didn't change the video because I didn't want that to be a sort of an uncontrolled variable. And I'm not pretending that these are really tightly controlled, but I was trying to minimize the number of variables. So I kept this constant as this video, and I'd never in a deep psychedelic experience, which you can certainly have with ketamine at sufficient enough doses, even though it is a bit bizarre as a dissociative anesthetic, I never had the experience of attempting, even attempting to watch video in these states. So to observe how the imagery, and this isn't, I suppose, on some level surprising because internally generated imagery can certainly have a tremendous effect, but to observe how the imagery affected the experience, how the music on the video affected the experience, how closing my eyes or opening my eyes while being exposed to this external stimulus affected my experience was very new, right? It was very novel to me. That's not what usually happens. If you are in a supervised setting, certainly within a university context, right? It's more or less what you described. It's like, okay, I've hold on, lay down. We have pretty good evidence to suggest that this at least doesn't interfere with reasonably good outcomes, but there's so much fine-tuning and assumption testing that can be done and that you will be doing. It's incredibly exciting to me. That's great to hear. Yeah. When I first started putting the meat on the bones and saying, okay, this is something exciting. Robin's interested. I have my leadership at UCSF is behind me creating this. I felt nervous because this is a new field to me, and it has a lot of history, right? There's a lot of therapists that have been doing this for decades, and obviously the indigenous cultures around the world have used these as not just mental health tools, but as religious and spiritual. And I wanted to dive into that part and determine if anything that I have just said to you is offensive to them. You know, I just was curious how it would be received because it is very, very different than how these experiences are often delivered throughout history. And I was really surprised to find that, especially on the multisensory stimulation side, therapists that I talk with were really receptive to it, that they thought that this is actually what they mean when they say shamanism. It is this act of environment tuning and creating to help guide that journey. And they feel that they're maybe pretty good at it because they've been doing it for a long time, but they were certainly open to having data that they could reach for that would give them more fine tuning and also recognizing how hard it is to train people that are new to this. And that if there was a library of stimuli and evidence of who it works and how it works, and also their ability to know what's really happening inside an individual's mind, they were very, very activated by that. Now, I'm sure that won't be true for everyone, but that was really reassuring. And the last point I want to make on this was my big goal here is not to say, we don't need therapists anymore. We don't need humans in the loop. We have enough technological advancements in terms of sensor technology and machine learning and artificial intelligence that we're good. We just need you to take this step into the box and come out fixed. And I really actually do not believe that that is the future we will or should have. I think that what we're doing here is building a set of tools, really sophisticated, informed, data-driven tools that will allow a therapist or any practitioner in the space to be more effective at helping their patient. That is what I hope and believe will come from this long journey and that technology will be what it always should be. Is it just another human tool? So that's where I hope and do believe that we will end when we go through this research approach. Looking at Akili and Endeavor R RX for a moment, because I'm tying these things together in my own mind, I'd love for you to stress test what I'm about to say in treating not just children with ADHD, but patient populations or populations with autism, with depression, these different conditions, and talking about how it is possible that a single intervention can be efficacious across multiple conditions. This conversation or related questions translates quite well to the study of psychedelics. And certainly Robin Carr Harris talks a lot about the default mode network and how and why it appears that some of these compounds are therapeutically effective, or at least appear to be. The data suggest that they are effective across a pretty wide range of conditions that most or many people would assume are not related at face value, right? So you have different types of addiction, including alcohol use disorder, opiate use disorder. That's opiate use disorder forthcoming at Hopkins. And then you have PTSD. You have treatment-resistant depression as well as major depressive disorder. There are people looking at OCD. And the assumptions of which, or one of the assumptions underlying the cross-efficacy is that perhaps there's this commonality, this common kind of neurological correlate with the default mode network, that there's a rigidity associated with these various behaviors that is attenuated with psychedelics."