#608 AI Won’t Replace Therapists. It May Replace Guesswork | Dr. Steve Rondeau

[00:00:00] 

Mehmet: Hello, and welcome back to a new episode of The CTO Show with Mehmet. Today, I'm very pleased, joining me from Colorado in the US, Dr. Steve Rondeau. Dr. Steve, he, of course, he's, he's, uh, coming from a medical background, but also he's done a lot of work studying the human brain. Um, he, he published, uh, papers, books, and today we're gonna talk about this important topic.

But I don't like to steal much from my guests' time. As the audience know by now, uh, I keep it to themselves, introduce themselves. So Dr. Steve, again, thank you very much for being here with me today. Let us know a little more about you, your background, your journey, and what you're currently up to, and then we can start right away the discussion from there.

Steve: Yeah. Thanks. Thanks, Mehmet. Um, yeah, so I have a long journey that sort of got me to where I'm at, but I think ultimately, you know, where we're at now is, is trying to answer the question of how two people with the same [00:01:00] diagnosis, mental health diagnosis, can respond so differently to treatment. And I think everything I did up until this point has got me to a place where that's our, our obsession right now, is how do we, how do we answer that question?

So rewind back about 20-some years or so. Um, I was working mostly in autism and developmental pediatrics. Um, I had just finished medical school, did my residency training, and, um, worked, uh, almost exclusively in pediatrics and developmental disorders, like I was saying. And one thing that really, um, was profoundly obvious to me was how objective data can guide treatment so, um, you know, sort of intuitively in, in that field, just like every other field of medicine.

But when we started our mental health practice in 2009 here in Colorado, um, it was a totally different story. And so talking to the psychiatrist and the other mental health providers, [00:02:00] it became apparent to me that it was, uh, a lot more clinical judgment, a lot more experience, a lot more, uh, you know, intuition and feeling than it was based on, on testing.

And I think when I came to that place, it, you know, my, my brain didn't know how to handle that, you know? It's like, "How do we do this without data?" You know? And so that's the question we've been trying to answer now for about the last 20 years. 

Mehmet: So this is what you def- you usually define it or, like, you describe it as the guessing game, I believe, right?

Um, um, so, so what is broken, Steve? Like, why, why, you know, people, they didn't think about it for a long time, in your opinion? 

Steve: Yeah. You mean in terms of, uh, the system and how it works? 

Mehmet: Yes. 

Steve: Yeah. You know, it's-- there, there's an illusion of, of objectivity, I think, that's come in the field over the years. Um, and maybe I'll [00:03:00] a-answer that question with a question for you, which is, you know, when you have-- I mean, how many people do you know that say have depression, but when you, you, you meet them look very different?

Or you could substitute if you don't know depression, you know, anxiety or ADHD, and how different they look individually. So I mean, I'm, I'm sort of asking you that question. Do you have-- do you-- can you think of two people like that and, and how different they may seem in real life? 

Mehmet: Mm-hmm. Yeah, yeah. Uh, and by the way, I can think-- uh, uh, I'm, I'm not a, uh, an, an expert, but- Mm-hmm

sometime I think I saw also people who denies- 

Steve: Ah ... 

Mehmet: like have, uh, you know, having something. But, you know, I think they need to go and-- and not, not because it's a sickness, of course everyone knows- Right ... like it's something for our, uh, well-being. So I think they need to go and, and, and check themselves. But yeah, I know also people who thinks that they have something, but yeah.

Um, a-a-and you know, maybe you can enlighten us more, Steve, about it. Like, and I [00:04:00] think, you know, the way of living that we have today, especially with this fast-paced- Oh ... technology changes is one of the main factor. 

Steve: Undoubtedly. 

Mehmet: Um, uh, so, so, you know, just-- I, I know you're gonna answer and put that, but, you know, also I want to, to, to try to understand from you within the same topic, as a clinician, you know, um, like how-- you know, like what kept, you know, repeating in, in patient care that made you start looking for objective biological measurements instead of symptom-based diagnosis?

Because, you know, i-if you go today, people, they get out a f- a, a set of questions, fixated questions I call them, um, and then they say, "Okay, based on this, you answered that." So what, what, uh, patterns you saw, like you, you also, I would say, uh, like it, it encouraged you to know that there is another way of doing this?

Steve: Yeah. Yeah, those are good questions. I mean, coming back to where we're at, [00:05:00] um, I think what has-- what I've seen, um, in my years of doing this and working with many, many, uh, mental health providers is that we've relied a lot on a diagnostic system that, um, does a very, very good job at describing people's suffering, okay?

We have a manual that describes it very well. Um, largely it's, it's done with the DSM or the Diagnostic and Statistical Manual. It's been through a number of revisions in which a new one is, is on the horizon. But it, it doesn't give us, um a full picture of what's gonna actually help that person. And I think that's what's uniquely different about mental health versus the other specialties in medicine.

Whereas you have, um, this illusion that the diagnosis is going to get us to a, an answer, it actually [00:06:00] doesn't. And that's been published in a number of studies that the, the correlation between an, an, um, accurate diagnosis and treatment outcome is actually, or positive treatment outcome is actually quite low.

And so it, it speaks to this underlying sort of physiology that makes us all unique. And so, you know, your flavor of anxiety and my flavor may look totally different if we both, you know, had the same, you know, hypothetical diagnosis here. And so that's what we're trying to, to understand. And so the book that we use for diagnosing does a very good job of describing how we, you know, how we, um, um, m- may, may converse, how me and you may talk about a condition, about a patient, a shared patient, but it doesn't help us understand how we got to that point.

And so when you, when you start to answer that question, you start to see the variability within, um, individuals and why perhaps, uh, that can explain [00:07:00] how people respond so differently to all these interventions. You know, we have, we have, you know, say for ADHD, for example, we have brain-- we have medications that speed our brain up.

We have medications that slow our brain down, and everything sort of in between. And so it, it really speaks to the, the variability. But if you have a eight-year-old child in front of you who's diagnosed with ADHD or struggling in school, you wanna know which thing is gonna help them, you know, as opposed to just trying a bunch of, um, you know, a bunch of treatments and hoping one of them sticks, so.

Mehmet: Right. 

Steve: Yeah. 

Mehmet: Um, and in order to, you know, have this better way of doing things, I know, like, you've built a database of more than 50,000 brain scans. 

Steve: Mm-hmm. 

Mehmet: Um, i- i- it's not an easy task to do, so I'm sure, like, there were a lot of challenges, uh, in collecting the data. And then of course, um, you know... So people thinks, yeah, collecting data is [00:08:00] easy.

What, you know, it's, you know, it's simple. They just put, you know, the, the patient, you know, in the machine, you know, and then they take the scan, and that's it. But of course, there are, like, a lot of, um, missing parts, I would say. Of course, without going too much technical maybe, but, like- How is that process and what are, like, the most difficult part about it, especially because you have to standardize the data and of course there is, you know, the, the, the, the other, I would say, challenges around, you know, how to correlate this data together after 50,000 subjects?

Steve: Yeah. Uh, that's a good question. I'm, I'm sort of having flashbacks to about 2016 or so when we started the project, and, um, the idea at that point was to gather all the brain scans that we had done that time, you know, hundreds, and compile them into a repository where we could just have them. So say you, you know, I wanna come out and do a presentation to your group about, oh, whatever, [00:09:00] stress or ADHD or something.

And so I would have a repository of brains where I could go and look up pictures and cases and, you know, do interesting work with it. Said, "Hey, we have so much data, let's do something interesting with it." At that point, I was doing much more clinical work and, um, and so these scans were all used, uh, by me clinically and our team, uh, of individuals here.

And, uh, so I go to my tech team and they're, they're building this essentially sort of primitive file cabinet for me, okay? This sort of basic project, a few thousand dollars. And they said, "Hey, we can add, you know, AI tools to this and, and really take this to the next level if you'd like." And I, you know, my ears sort of perked up and I was like, "Well, that sounds interesting.

I, I like tech and I like data. Let's, let's see if we can pull these things together." And so what turned into probably a two or three-month project has now evolved into a, you know, seemingly never-ending project now. Um , and we're [00:10:00] constantly refining and constantly making it better. But the idea was, is to take those scans and, and to be able to, you know, make some meaningful, um, conclusions and so forth from it.

And now it's, it's e- it's evolved into a much more elaborate process of looking well beyond what we could ever see with our, you know, sort of naive human, human eye. And so, um, when we gather EEG data, which is actually quite easy to do. You're, you know... I think a lot of times when we talk about brain scans pe- people picture a, you know, a tube and sedation, and we're gonna sit in a scary thing with a lead vest or something.

It's... No, it's quite easy. In fact, you could do brain scans in just about any kind of context even, um, functionally with a, you know, with an athlete or something, for example, okay? So it's very- 

Mehmet: Mm-hmm ... 

Steve: very easy and sort of practical to do these days, especially with the, the hardware evolving. Um, so we record these, these, um, these data, and then we have it, um, processed through our, through our system.

So, um, in doing [00:11:00] that, we've, we've piqued the interest of a lot of other groups, and so they started sending us data and we started processing their data. Said, "Hey, can you process ours? Can you process ours?" And we started getting data from all over the place, from law enforcement to first responders to, to groups in Hollywood with, you know, uh, peak performers and, and, you know, actors and, uh, everything in between.

So we have all these little niche-y groups of, of, of people where we can, um, um, you know, process their data and help them understand their brain better and help them perform better. So, um, I think to answer your question specifically around the, the challenges, I mean, I think the challenges are limited by, you know, what do we do with all the data that we have?

How do we process it into meaningful, you know, ways and make, make it understandable? And how do we bring something to the masses that don't quite understand, you know, what it is we do or, or don't know that they need it yet? And I think for me, that's one of the, the bigger [00:12:00] challenges is, is changing, um You know, essentially sort of a dinosaur and, and getting that to change from, from where we are now with, you know, subjective questionnaires to, to making this routine before we do any intervention to understand, hey, what's going on with this person?

So for example, I mean, I, I ultimately wanna answer the question of like, you know, someone comes in and they say, "I'm depressed." Well, is it a, is it a biological hardwiring pattern that we can measure, okay? 

Mehmet: Mm-hmm. 

Steve: Or is it circumstantial based on, you know, trauma, grief, something that happened to you? Or is it, is it both?

You know, and likely there's a, there's a, a middle ground that includes both. But if you include the biological data That the range of possibilities is, is enormous. So we did a presentation for providers the other day, and I had one of my middle school kids do the math for me because this was [00:13:00] a complicated math problem.

I said, "Hey, son, come here." This is my advanced math seventh grader. I said, "I have 12 different brain patterns we measure." Might not sound like a lot. We actually measure thousands, but we report on a dozen different patterns, okay? And I said, "If we have these 12 patterns and any one of these can show up like depression," and again, this was a presentation to providers about depression.

I said, "What's the possible combinations of patterns that you could have in one individual?" And so there, there's over 4,096 different combinations of patterns that you can have in that brain scan. So if someone comes in with depression, okay? And you're a provider without objective data, okay? We just have a questionnaire and symptoms.

There's over 4,000 different combinations, okay? 

Mehmet: Wow. 

Steve: That, so now your job as a provider is to, is to, and, and I use the word guess nicely, I mean, but it's an educated guess, to guess which treatment is gonna work [00:14:00] for that patient, okay? Now, if you really want your mind blown, we're adding a new biomarker to our, to our panel here this week.

Now you've taken 4,000 and you've doubled it. It's gonna be over 8,000 combinations of patterns that could show up, all of which respond differently to different treatments. And imagine the trial and error for that patient. But not only that, imagine the exhaustion for, on the provider's behalf of having to, you know, having to help those people.

'Cause we all go into it for the same reason. We all go in to help people. But if you're trying, I mean, one in 8,000- Yeah ... that's, that's like playing the lottery, you know? That's tough. 

Mehmet: Yeah. 

Steve: That's very tough. Yeah. And it wears on you after a while, you know? 

Mehmet: Right. Now, w- when you, when you started to, of course, as you said to your point, like, of course, collecting the data is the first challenge, and then, you know, what, what to do with this data.

But once you started it to, to do the analysis, um, can you tell us, [00:15:00] for example, about one example where, you know, maybe for years, uh, conventional psychiatric thinking was telling something while your data proved the opposite? Mm-hmm. And of course, the data, uh, you know, like when you see a lot of patterns like, "Oh my God," like we, we might say, "Okay, we, we knew all this, uh, um, information wrong all these years.

We hold on it." And yeah, here we go. Here's the scientific proof and, you know, because we did all this analysis. So was there any- anything similar maybe, or maybe close to? 

Steve: Yeah. I, I know what you're saying. Um, there's two things that come to mind. Um, one of them has to do with suicidality, which, um, I, I think, you know, they're, they're probably in the bigger scheme of life, and this probably isn't the, the topic of the podcast, but I, I don't think there are coincidences in, in that way.

I think this was sort of given to me to, to [00:16:00] figure out. So I think the suicidality part is one, and the other was a case that, um, we talk about in the book of a girl who, um, and maybe I'll start there, was in her 20s And I, at that time I was working in residential addiction substance use center, um, with these, these kids 17 to 25, and they would come in, um, after having been sober, and this was a transitional living arrangement that we have, um, for these people to go back into, into life.

And so we do brain scans on these individuals, um, on everyone that comes through the center. Okay? And a lot of times we don't have history, which is another sort of challenge within, um, uh, sort of disjointed care within the system, and, um, uh, which is getting better, but it, it still can be a challenge of getting information.

These are people from all over the world, all over the country. So this girl comes in and she's... I said, I look at her brain and I said, "Why are you here? You don't [00:17:00] look like a lot of the other people that we see at this center." I said, "This looks quite different. This looks like you have, like, ADHD or s- maybe some learning issues.

Like, what, what's going on?" 

Mehmet: Mm-hmm. 

Steve: She says, "Well, I, that's... No, I have panic attacks. I can't leave the house. I'm, you know, I'm, I've been in three residential addiction centers, um, fi- or three different residential centers. I'm finally in an addiction center. I don't have an addiction. I just am struggling," and they didn't know what else to do with me.

So they have her in this center, and these are... I mean, you, you're not just hanging out here for fun. I mean, they're very expensive and, um, um, you know, this is, um, you know, someone who's, who's deeply struggling. So I said, "Well, why, why don't we treat you like you have ADHD? Because that's what your brain is, is telling me.

That's what the pictures look like. Instead of treating you as if you have panic, because looking at your list here, you've been on 40 different medications." She's in her 20s, 40 different medications. I said, "Why don't we treat you as if you have ADHD?" So I go back to the psychiatrist and I tell him. He says, [00:18:00] "Well, I mean, she's in an addiction- she's in a treatment center.

We've got, you know, safe, uh, you know, environment. Let's, let's try it." And so we did, and within a couple days, I mean, that was, that was it. These, these short-acting ADHD medications transformed her life. Panic attacks went away and, you know, the rest is history, happy ending kind of story, you know? And, uh, but, but the thing, the point of that is, is that if we had only treated her story and only treated her label, she would have continued down these paths of medications that were targeting the label, not what her brain was saying, not what her brain was saying.

And so by treating what the brain was saying, we understood, you know, her words were, were, were sort of the manifestation of those patterns. And so it was really a transformative moment for us, and I think once the, the, the psychiatric team saw that, it became sort of a light bulb moment of like, "Hey, we need to be doing this, like, routinely on people."

And so actually, that's one of the key moments that, [00:19:00] that really, um, springboarded this work to be done routinely on, on all of the mental health patients that came into that, to that center, and then subsequently just, just grew from there. But once you see that, I mean, you just, you just can't unsee how that, how that is, you know?

But, but to the y- just to backpedal for a second, to put a, a, a woman like this on the treatment that we did based on her diagnosis would largely be considered like, you know, borderline malpractice because you would never do that to someone who's, who's having a panic attack, right? Someone who's anxious- Mm-hmm

and you say, "Hey, I'm gonna give you like three Red Bulls or 12 cups of coffee." They- it would be like, "No way. Like, that's crazy. You're gonna make it worse." Right. But in reality, her brain was going too slow, not too fast. So I think one of the key learning, you know, moments from that was understanding that, hey, there's, there's anxious brains that are going too slow, and there's anxious brains that are going too fast, and we can't treat them the same.[00:20:00] 

And so that was really the, the start of it, um, you know, from there. And so, yeah, like I said, the suicide stuff too really transformed us a lot too in, in understanding that people can largely be struggling, um, with suicidal thoughts and, and suicidal ideation. And, um, and we know even just from behavioral research that most of that goes unspoken.

Most first side- first- Right ... time suicide attemptors, um, don't tell anybody. And so if we can see that in the brain data- It can open up conversations. It can open up conversations with law enforcement officer or a pilot or a teacher or, you know, a, a guy who's suffering at home with a family but you don't know it because he hasn't told anybody, you know?

And so having that data, it doesn't have to be a label, it doesn't have to be a sentence. It can just be a conversation starter to say, "Hey, I see these patterns that are [00:21:00] associated with perfectionism and negative self-talk. Tell me about that. How does that show up in you?" You know? And then it starts the conversation.

So that's been a lot of fun work for us is, is, is, um, you know, in those two areas for sure. 

Mehmet: Yeah. Uh, just one thing, uh, which I'm by-- you know, I'm curious, uh, to, to- Please ... understand more 'cause you mentioned about the biomarkers also as well. So yeah, got the brain, uh, scan part of it and, you know, this-- the example that you just gave is, is, is great one, you know, for, for, uh, anyone to understand, you know, uh, how that works.

What I'm curious about, you know- Again, I'm not the expert you are, Steve. 

Steve: Mm-hmm. 

Mehmet: Um, how much, you know, you know, the genetics, the DNA plays role? Because we always hear sometimes, you know, like, okay, um, this boy is angry like his father, or like this girl is like behaves a certain way because, you know, her mother.[00:22:00] 

Is this, is this purely psychological and brain-related, or like really is there something, you know, also in the DNA or like m- maybe in, in, in our, you know, other part of the body that might be responsible for certain, you know, uh, symptoms related to, to these kinds of, uh, uh, mental health issues? 

Steve: Yeah. No, it's a really good question.

I get variations of that question quite a bit. And certainly we, you know, we don't have all the answers. What we do have- Mm ... is a starting place to have a conversation about data, um, taking the subjectivity out, okay? And that's what we wanna try to do, is to, to take that subjectivity out of the equation and, um, you know, language barriers, cultural barriers, and these sorts of things, and just look at the data.

But yeah, look, you're, you're absolutely right. There's, there's a neck down component we call it, that, that plays a large role in [00:23:00] what, how people experience the world and how their brains, you know, respond to, say, stressors, for example. So, you know, as far as genes playing a role, I mean, absolutely there's, there's likely a component.

But if it were genes alone, of course, we would probably have an answer for a lot of these complex behavioral and psychiatric types of conditions. And so what we're trying to do is see how those, um, those measures, those sort of like static measures show up functionally in a, in a brain, okay? And so I think that's the key is to say, okay, yes, you have this gene, okay?

But is it being expressed in a biological way that can be manipulated or supported, you know? And I think that's one of the, the main, um, um, you know, things that, that we're trying to do. But, but also understanding, and, and this is a big part of why we wrote the, the book, the Think Like a Brain book, um, just, just recently, is that- We want people to [00:24:00] understand that these labels aren't just, um, deficits or problems, but they can be strengths and gifts when they're supported.

And so once you see the patterns, you can start to create environments that support those patterns. So we started talking about this before the call actually, um, about environments. I didn't know we'd get into this, but we were talking about Arizona, which I just left. And when we take our Huskies, we have two Huskies.

They're, you know, Huskies, big furry dogs, two coats. 

Mehmet: Yeah. 

Steve: Okay? When we take them to Arizona, they're absolutely miserable, okay? Those dogs are miserable. They don't wanna go outside. They don't wanna play. Um, it's just not a good environment for Huskies, okay? But when we bring them back to Colorado, and we take them up to the mountains, and they're playing in the snow, those dogs are happy as can be.

Now, if I were to evaluate those dogs in Arizona, they would be depressed. They may be moody. You know, pick a label, okay? But if I evaluate them playing in [00:25:00] the snow with the pack and, you know, uh, I can't even get them to come inside. They're having so much fun. And so that's what we wanna try to show with these brain scans too, is that people aren't broken.

They're functioning with brain patterns that aren't being supported in their environment. And so I think that if we can start to create those, those environments, we can show where people are, um, where their gifts are. So, so back to your question about these other sort of physiological and, and biological components is that, um, um, you know, having those static measures doesn't necessarily show a manifestation.

That's what we wanna try to do. So we wanna try to show, A, what's going on functionally with this person. I'll tell you a quick story actually about this. Sure. Really interesting, um, work. So as my world started to collide, I was doing all this sort of biological work on autism and research and biomarkers and so forth, but I was also doing brain scans over here on the...

I don't wanna say on the side, but it was sort of peripherally related to what I was doing [00:26:00] And we had a woman come in, uh, out of town, did this sort of elaborate workup, and, uh, we did a brain scan on her, and I said, "What's going on with this woman? Her brain scan looks normal. What's up?" And, uh, they said, "Oh, no, no, no.

This is... This woman's struggling immensely. She's got schizophrenia. She's hospitalized, you know, struggling." I said, "Well, I don't see that in the brain scan." And so we went back to the team. We had our meeting. They talked to her, and I said, "Do we need to look somewhere else?" And so we did. Uh, we looked at her immune system and found these, um, sort of brain antibodies that were, um...

Her brain was sort of attacking itself, okay? And when we treated those, the lady's life completely transformed, okay? She went from schizophrenic, you know, dys- dysfunctional, hospitalized, um, you know, highly medicated, to married, has children, totally normal life now, okay? But my point of that is, is that, you know, if we look at genetic data or we look at these static markers, it doesn't necessarily tell us [00:27:00] functionally what's going on with the brain.

And in this individual, the brain scan looked largely unremarkable, which told us that there must be something else going on. So the point is, is that even in the absence of data, or I'm sorry, in the absence of, um, aberrant findings, of abnormal findings, we still can create a, a, a treatment plan or a path to help individuals that doesn't revolve around, you know, trial and error and guesswork.

So it's really been an interesting journey to see how these things play out, not only in, you know, these deviant sort of patterns, we'll call them, but also, you know, just purely in the absence of, of findings. So that was an interesting one that sticks out. 

Mehmet: Yeah. Now, I want to go back a little bit to, to the AI part.

Uh- Sure ... and of course not, not from the technical perspective, but on a high level. Um, and actually it's good because, uh, currently I'm reading, you know, um, you know, uh, a book about... Not technical, it's like more philosophical book about AI. Mm-hmm. It's called The Nexus, uh, by Yuval Harari. [00:28:00] He's, he's a, he, he, he's behind the Sapiens and many great books also as well.

And he comes to the part which is talks about the explainability, right? So, so- AI systems are black boxes, right? So- 

Steve: Yeah ... 

Mehmet: sometimes we don't know why they decide to do something, right? And, um, we don't know even the bias that based on the data that they were trained on. So how, how we can balance, you know, the algorithms', uh, recommendations, I would say, with clinical judgment here to make sure that, yes, we know that the traditional way is not giving us the full picture, or sometimes it's even misguiding us.

We agree on this. But how we make sure, like also when we rely on the data and the algorithms, we can still justify it from clinical perspective. 

Steve: Yeah. That's a fantastic question. I think that, um, what we wanna be [00:29:00] mindful of and, and clear about is that we're not trying to create a tool to replace the provider, okay?

What we're trying to do is add a level of objective data to a provider's clinical judgment, which is essential, okay? Actually, when I, I mean, in full disclosure, when I first started, I thought that we might be able to do that. We might be able to m- make, um, make a system that could fill some of the gaps where care isn't being delivered, you know, under, under, um, served populations and rural areas and so forth.

Um, but what we've really learned over the years is that it, it, it definitely requires a human in the loop sort of scenario where we can, um, take into account, uh, history and, and cultural differences and things like that, okay? Um, so back to your, your question, you know, kind of on a A technical level. What we're trying to do [00:30:00] is look at the bigger picture of brain patterns on a data side, almost like a, almost like a data scientist in your pocket, okay?

To say, "Hey, I have a theory." Okay, and this is how it actually came about with our, our theory of, um, uh, of suicidality. You know, all these groups were publishing papers on suicide. Uh, it's a very niche-y group in EEG, okay, in brain scans like we do. They were doing single biomarkers, okay? Single biomarkers.

This biomarker is associated with suicidality in this population, and this biomarker... And so different groups had different biomarkers. And, you know, uh, uh, from my perspective, clini- as a clinician at the time, it was like there's no way. Suicidality is way too complicated to be narrowed down and boiled down to just one biomarker, okay?

And we had different theories of, like, maybe four or five different biomarkers, okay? And so we bring a human data scientist into our, into our program, and we say, "Hey, look at our data and tell us what you, what you see." And [00:31:00] so the idea was, was to come in there and, and see which of the biomarkers seem to hold water in our bigger pool, you know, of data.

So I think ultimately from the AI, what we're trying to do is not necessarily Create new, um, necessarily anything that would r- you know, replace the clinician or tell them what to do necessarily, but to actually pull different pockets of data together. So, so for example, one of the things that I'm most interested in is predicting how people respond to treatment, particularly novel treatments.

Like here there's a lot of talk about psychedelics now and, and neuromodulation techniques where we stimulate the brain and do different types of things like that. I want to know, and I think our patients want to know, and definitely our providers do, who's gonna respond to that treatment before you even start?

Like, that's the question I want to answer. So not necessarily pulling some new novel thing out of the air, although that's great. [00:32:00] I want to be able to say, "Hey, there was 1,000 people before you who did that treatment, and they all did well. What do they have in common?" Okay? "And do you have that marker?" Okay?

So those are the questions we're trying to answer. So basically pulling our data, understanding, you know, what people have taken, what have they tried, what hasn't worked, and then also pulling the data of all these brilliant people who are publishing research every day into the system so we can start to see, hey, do these biomarkers hold water in our system as well?

Okay? You know, they see this biomarker in their cohort. Do we see that in ours? Okay? So that's really what we're trying to do is, is to really take all the evidence that's been out there and then compile it with the evidence that we have, and put it together and say, "Here, provider, Mr. and Mrs. Provider, here's, here's, um, a guide that may help you to understand what to do, you know, with this patient that's sitting in front of you."

So- 

Mehmet: Right. H- H- How, how people are... You know, because there was a lot of discussion [00:33:00] about these, um, you know, new trials. Um- Mm-hmm ... because people still when they hear the name, you know, um, their mind might go somewhere else, right? And, um, so are you seeing, like, kind of pushback on, on adopting these, uh, these, uh, psychedelics, you know, mainly in, in, in treatments?

Steve: Oh, oh, psychedelics you're saying? 

Mehmet: Yes. 

Steve: Yeah. Um, it's funny. I was a, a, a lot of the pushback. Um, I, I mean, I... See, the thing, the is- the issue, you know, at the core of it is that I, I really am uncomfortable with indiscriminate use of treatments. So something new and novel comes out, we're gonna do it with everybody, you know?

And we learned that so It was so ingrained to us working in autism, okay? When you see-- There's this common sort of saying in autism, "If you've seen one autistic kid, you've seen one autistic kid." Like, that's it. [00:34:00] And it's like when you understand that at your core that, like, every kid that comes in with autism or, you know, now, now most of them are adults 'cause this was 25 years ago.

But they-- when you see one, they're, they're vastly different, and what res- what helped that kid is likely not gonna help the other kid, even if it's their sibling. I remember a case, we had two autistic kids. We did the exact same treatment with them at home, and one of the kids did profoundly well. The other kid had no response at all.

And I mean, they're siblings doing the exact same treatment at home. You'd think at least they would, you know, maybe a little bit and a little bit more, but it was, like, no response at all, and then another kid doing profoundly well. And so when you understand that at your core that, like, the label just tells you the, the cluster of symptoms, but it doesn't tell you what's going on, you know, kind of under the hood.

It's basically a check engine light, but it doesn't tell you what's going on under the hood. So, um, to your question, sorry, pushback on the psychedelics. Um- 

Mehmet: Yeah, 

Steve: yeah. Yeah. I [00:35:00] had a lot of reservation because it was being used so, um, just willy-nilly. I mean, it was like everyone who had-- I mean, the, the criteria was treatment-resistant depression.

People could get ketamine. Ketamine isn't a classic psychedelic by the, the definition, but it's, it's categorized that way. Um- 

Mehmet: Yeah. 

Steve: And so everyone's getting ketamine And I'm like, "Well, the definition is, is that you've, two treatments have failed you, uh, two first line, um, psychiatric treatments, uh, uh, medicines."

And it's like, well, if you look at this through the brain scan, I just told you 49 or four- 4,000 different combinations, that's easy. I can make someone fail two treatments like piece of cake, you know? And so, um, and then you go do, do, go do ketamine. So I was... I had a lot of pushback to that myself. But when you start to see, "Oh, here's a group of people that respond really well, let's put them in a group and do that treatment," your outcomes are remarkably better.

And so that's what we're trying to look for. So I think [00:36:00] that as far as, um, um, psychedelic treatments go, um, I think we're gonna continue to see resistance within, uh, some of the, the, the conventional field if we continue to use it like we continue to use our current psychiatric medications. But psychedelics have done one thing that's been great, I think for, for, for what I do, is that it's created interest outside of medicine to be able to see- Yeah

what's happening in the brain. You know, what's happening to my brain when I take this thing, or which thing should I do? Should I go to the jungle and do ayahuasca, or should I go do- 

Mehmet: Yes ... you 

Steve: know, ketamine here or whatever. And it's like, "Oh, I saw, you know, Aaron Rogers do, you know, psilocybin on Netflix," or like, whatever it is, you know what I'm saying?

And so people see these things and they're like, "I think, well, I should do that. I have PTSD. I should go do DMT." But, um, a- and so what it's done is it's taken people from these other fields, particularly people who are interested in tech like us, and people who are, um, maybe have, uh, resources or a team [00:37:00] of people with resources to, to start to answer these questions of like, "Hey, we wanna know what's happening in the brain."

But the r- the real question I think we should, should get to is I wanna know who's gonna respond before they do the treatment. Because if you're on your last leg, you're a war veteran, you're someone who's experienced deep grief or trauma, and this is your last hope, and you're gonna go spend 6K to go do, you know, psychedelics in the jungle, and if it doesn't work, you're done.

Mehmet: Yes. 

Steve: I wanna know who that is before they do that, you know? I wanna know who that is, at least with some level of objectivity, not just, you know, "I saw it on Netflix." 

Mehmet: I, I, I think, you know, Steve, this is the, the beauty of, you know, um, marrying the experiment with the, with, with the, with the power of, of data and, you know, being able to analyze.

So, so I think, you know- Work similar to what you've, you've done and you're, you're still doing is- 

Steve: Mm-hmm ... 

Mehmet: uh, I call it killing the myth, right? So, [00:38:00] you know, the, the myth is, would become busted. Like, because, you know, people when... You need to try at least, you need, you need to test. And without doing that, we will never move forward.

Sorry. And I'm happy, you know, like, because, uh, you know, I, I shared with that you w- w- before, like even personally, right? So even myself- Mm ... right? So, so sometime I, I sit and think, "Okay, do I have ADHD?" Mm-hmm. Um, you know, do, do I have this? Do I have, do I have that? And then my problem is, uh, it's not my problem, but my, uh, you know, my expectation is, you know, when I...

If I go to, to, to someone to see me, they're gonna follow a book, right? I'm not against following books, don't get me wrong. No. Like, I mean, when doing diagnostics. But I mean, y- you know, this person there, you know, I don't know their current mental status, right? 

Steve: Mm-hmm. 

Mehmet: Maybe I came to them today and they were in a very bad mood because maybe they had some issues at home in the morning or [00:39:00] maybe- Yeah

you know, they, they had a fight with someone else. I, you know? Yes. So, so they would be doing this and they would be asking the questions and they would be taking notes when I'm talking to them. I've seen it, you know. I went like maybe three or four, four years ago, um, to, to a mental health clinic just, you know, uh, it's just a checkup.

It's not, it's not like something, you know, uh, I wanted to, to know about. I did it also with my daughter, by the way. Like, you know, like w- because someone told us something and then, you know, I start to do the research and, and then it ended up, you know, what they mentioned is it, it didn't match. So a- a- again, you know, a- a- and back to, to the point you just mentioned, like we need to test, we need to go- Right

to see where, where we can go forward. And, you know, mentioning this, tell me more also about the book. Like people, what can expect from the book? Other- Mm ... something that we didn't maybe discuss today yet. 

Steve: Yeah. I mean, I think you touched on something that's m- maybe becoming more and more intuitive to consumers, is that, you know, this, this level of [00:40:00] subjectivity is no longer acceptable, especially in an era of the, the, the data analysis that we have with, with AI and these other, these other tools.

It just is not, it's not acceptable anymore to do that. And so in any other field in medicine, we test the organ that we're treating. Every other field, okay? Mental health is the only one that doesn't do that. And so, you know, I think the time has come that we need to, you know, adopt some of these tools. I think some of the challenges, and this goes back to what you were asking before, is- And, and, and maybe one of the even larger challenges that, that we didn't mention is, is this sort of fallacy of normalcy, okay?

So like what is normal? So what, what are you gonna compare me to? How do you, how do you judge whether what I'm showing in my scan is normal compared to, to people who don't have symptoms and so forth? So there's some of the challenges around that. Um, and, and that's [00:41:00] been going on for, for a while, and I have some answers and thoughts of course to, to that.

Um, so I think that's a, an area that we, um- That, that as consumers we're starting to see. I mean, we would never accept that in, in cardiology. I mean, imagine, for example, you go into cardiology and you say, "Hey, my, my chest hurts." Okay? Mm-hmm. And they say, "Well, tell me about your chest." "Well, it kinda hurts.

I've been stressed lately and, you know, when I exert myself, I get, you know, I get these chest pains." Okay. Well, that could be, you know, you could have a rib out of place. You could have a lung infection. You could have heartburn. You could have, be having a, a heart attack. I mean, there's a number of different things that could be all causing it.

But if they said, "You know what? Just go try this thing. Go take this, this, this pill. It's a heart medicine, and we'll see how it does." Like, that would never fly. Never, you know? And, and the same with a runny nose. I mean, a runny nose could be 10 different things. You could be crying, you could be sick, you could eat something spicy.

But if we just give everybody antihistamines that has a [00:42:00] runny nose, and then if antihistamines doesn't work, we do another med, and then we do another med, we do another med. It's like, what, what are we doing here? I mean, we don't even do that with our cars, you know? It's like you plug it in, the light's on, you plug it in.

You figure out what's going on before you start ripping parts out, you know? So that's what we wanna try to do here, you know, with this. So I think ultimately, you know, in your question about, about the book and the papers, I mean, the papers, the papers ultimately are a group of colleagues of mine, um, you know, wanna answer this question around why do people respond so differently to mental health treatments?

And so we gathered all the information on lots and lots of studies to show the variability within these common conditions: anxiety, depression, ADHD, trauma, so forth. And, and trying to give clinicians a guide around bringing brain scans, EEGs, to the table with their clinical evaluations, something they're already used to, okay?

Pairing them together, okay? To show them scientifically there's, [00:43:00] there's evidence for that. The book, on the other hand, is, is more geared towards, um, providers, therapists, but, but also it's geared towards, um, you know, consumers as well, so people understand that, like, hey, my brain isn't, isn't necessarily broken.

I think the problem is that we just haven't, you know, looked under the hood. So we tell some stories about people who have gone through these experiences. We talk about, um, how the field's evolved. We talk about, you know, why objective data, you know, can make a difference, and it's, um, it's really, uh, you know, a compilation of the last, you know, 20-some years of, of doing this.

So I think it's a good entry point for people to start to have that thought and that conversation, whether it's someone who's trying to perform at the highest level of their game, or it's someone with a kid who's struggling or whatever. But you, you know, you take that book and you go to your provider and say, "Hey, I wanna do, I wanna do this.

I wanna figure out how we can do, you know, do a scan and, and bring this, this to the, to our conversation." So, so ultimately, that's where we're at with the, [00:44:00] with the work that we're doing. 

Mehmet: Um, before I wrap it up, just, you know, out of curiosity, when, Steve, we can expect this to be, um, you know, widely used worldwide and get rid of You know, the-- I, I'm, I'm sure, like, the traditional way is not going anywhere- Mm-hmm

don't get me wrong. 

Steve: Mm-hmm. 

Mehmet: But, you know, to, to have these se- I-- it's kind of a second opinion. Yeah. It's kind of rectifying maybe a, a, a wrong treatment, like similar to the, to the example of, of the girl you just, you know, discussed earlier. Uh, I, I gave you, like, my own ex-experience also as well. So what, what can we do, you know, to, to change how...

A-and, you know, this is in the world of tech. People might ask- Mm ... while listening, "Oh, but this is CTO Show." Like, like, what Steve and Mehmet are discussing is maybe not related directly to tech, but it is related. And people- Mm ... who follow the show, they know that I had, like, uh, you know, [00:45:00] people talking about mental health issues for founders.

Yeah. I had my f- my, my buddy, uh, James Oliver a couple of, uh, months ago. He came and- Yeah ... talked about it. Uh- Yeah ... I had, like, several, several, you know, um, doctors who came, and, and, and they shared their experience, and, and also yourself. So yeah, we-- talking is good, but how we can make it, like, more, um, spread, I would say.

Like, how we can enhance it worldwide, not- Right ... you know, I know, like, you're, you're based in the US. So what, what are we doing to change this, Steve? 

Steve: Yeah. I mean, the interest is definitely there. I talked to somebody f- you know, from your country just a few months ago, and, and there's definitely a, uh, an interest in, in changing, um, the paradigm.

Um, it's a really good question. I mean, I, I-- ultimately I hope in my lifetime it becomes, uh, you know, standard. Um, I think that there's resistance in some ways around, you know, territory and who takes this on. So is it [00:46:00] neurologist? Is it psychiatrist, or is it a whole new profession that comes on and does this?

I mean, ultimately a provider can't, um... And I say can't, I mean, they probably could if they tried really hard and, and studied. But, um, they can't bring this to their practice without having another person in the loop, and so that's the role that we fill, to analyze the scans and, and, and do that so that the providers don't have to go and, and learn all this stuff in, in-- while seeing patients and in their practice.

So, um, so yeah, I think larger adoption, um, validation, um, third-party payment becomes a big part of the conversation. When insurance parties start paying, people tend and doctors tend to start listening a little more. Um, so I think that that can be, um, you know, a big, a big entry point. But, um, but ultimately, I mean, on, on our front, uh, it comes with, you know, with funding and support and getting the word out.

So it's, it's definitely changing. We're seeing more and more [00:47:00] pairing of technology. So, like in, for instance, the TMS world, the The magnetic stimulation world is now, um, coupling EEG scans before the, the treatment. So again, it's not a one-size-fits-all, it's more about personalized medicine. So we're seeing a really big push towards personalized medicine with, um, with some of these technological-based, um, treatments as well, and then monitoring that, seeing how your treatment is responding.

So I think there's definitely an interest. It's about being able to, to bring this at a large level on a platform where people can see it, doctors can experience it. Um, you know, most of my most of our partner clinics, I'll tell you, are- were first patients. They were-- Their kids were patients. Their- 

Mehmet: Mm-hmm.

Steve: You know, they had a struggling patient, and then they, they, they sent them over, and then they experienced that firsthand, and then they became, you know, customers of ours. And so [00:48:00] I think ultimately it comes down to, you know, to, to seeing that at... But, you know, I think back to, oh, maybe when I was in college or something, you know, and, and not to compare myself to Jobs, but Jobs comes out and he says, "Hey, you guys need this, this iPod in your pocket to carry, you know, a million songs."

And we're like, "That's dumb. Why do I need that? I got a CD player." Like, "Oh, you know, I don't need that." And so, um, you know, I think and l- and a large part of it is that we're, we're delivering something to people that they don't necessarily know that they need, but there's something telling them like, "Hey, we, we need to change."

Yeah. We need to change. And so it's, it's about, you know, building the team, getting the word out, getting on these platforms so people can see what's going on. We're not trying to take anybody's jobs. Robots aren't, aren't using AI to replace psychiatrists. It's more about just bringing objective data to the table.

Um, and, and really, I, I mean, ultimately, Matt, is, is anyone with a brain, this applies to, okay? So whether it's your struggling, you [00:49:00] know, neighbor, or whether it's your kid going to school and understanding should they go to art school? Should they go to music school? Should they go to sports? Like what-- engineering?

What should they be doing? And if you try to get the, the, the, the wrong brain pattern to fit in the wrong environment, you're gonna have friction. And so, man, wouldn't that be cool to do that on every 10-year-old and know like, "Hey, this is the right environment," or, you know, maybe we need to... May- maybe your kid isn't, isn't wired for, for science, but they're wired for music.

Let's, let's support them over there. So I think ultimately it comes down to anyone with the brain that we, we need to be, we need to be doing these scans routinely. 

Mehmet: Yeah, I'm trying to get a, the name of a book because, uh, I, I like to, to, to mention the book, uh, um, by the head of... It's by the head of, a previous head of, uh, Google China.

Now, now I, I will, I will get the name very shortly. And the reason is, uh, to your point about, you know, [00:50:00] um, jobs and, and, you know, I mean, the, the, like replacing jobs and this- Yeah, yeah. 

Steve: Yeah ... 

Mehmet: yeah. So, so the, yeah, so I found the book name. It's AI Superpowers by, uh, Kai-Fu Lee. 

Steve: Mm-hmm. 

Mehmet: And, you know, I like the book and, you know, the, the author mentioned that, you know, probably one of the things that cannot stay, because AI, uh, I believe in, in a long time will not be able to do it, the empathy.

And, you know- Yeah ... especially when it comes to, to be working in the field of psychology, I think empathy- Yeah ... is, is much needed. It's great. So yeah. Yeah, but, but yeah, you, you, you need, you need, you need a, something that help you to do better diagnostics, and this is the whole, you know, topic maybe we discuss today.

And just if you allow me also, Steve, to add something regarding- Please ... why this is important. Um- Of course, like it's not new. It's been with us for a long time, you know, the pressures of the life and, you know, the way, you know, how, you know, i- if my grandfather, you know, comes [00:51:00] alive now and then see, I'm sure, you know, he, he, he, his mind will get blown and maybe he will get a, a second heart attack, whatever.

Because, you know, you know, how can you digest all this information? So just, you know, when I sit down and I think only and only about the amount of information I have to deal with daily- 

Steve: Yeah ... 

Mehmet: um, whether emails, messages, um, a- a- and forget work, by the way. I'm, I'm talking about normal life. Like s- still- Mm-hmm

I didn't, I, I didn't do anything related maybe to work. I didn't, I didn't have to put my brain on, on something, and it's, um, it's massive. Plus the messaging we get and, you know, this continuous, um, FOMO effect that we are exposed to. 

Steve: Mm-hmm. 

Mehmet: And, and you feel like you are always into this race. Mm-hmm. So g- you know, like this is why we need, we need to spread the word about it, and I'm, I'm very happy that today, um, y- you, you, you shared with us, um, what, what you've been up to, Steve.

And I think we- Yeah ... need to talk more about it, and I'm happy, [00:52:00] you know, about, you know, the, um, the efforts that, uh, you, you're working on and, you know, in, in everywhere in the world. So that's really encouraging. And I encourage everyone, you know, to take it easy. Um, yeah, I know like there's a lot of, uh, push outside about AI replacing jobs and- Mm-hmm

this and that. Take it easy, guys. Like I'm not-- I know it's not easy. Don't get me wrong. My, my job one day might be replaced. I'm not sure. But I mean, be assured that, you know, um, support, uh, and human support, human empathy is gonna be staying. And I think, you know, uh, Steve, what you're doing is, is in the heart of this because, you know, you need to show a lot of empathy to understand also your, uh, uh, your patient.

And again, you know, this is something really, really I appreciate that you took the time today to share with us. Before, you know, we close, something I always ask my, my, my guests, where people can get in touch and find out more? 

Steve: Yeah. So definitely most active on LinkedIn. Um, so you can find [00:53:00] me there and all of our work and, um, you know, spreading the word and conversing and so forth.

Um, the other area is thinklikeabrain.com. That's our new-- my new book. Um, that's probably a good, you know, entry point to grab that and, uh, and check out what we're, what we're up to. And, um, you know, the company, our, our company is AxonEG Solutions, so you can find their, our information there, our peer review papers, uh, anything we're working on, and get in touch with us through, um, through that, through that, uh, website too.

Mehmet: Great. I'll make sure that all the links will be available in the, in the show notes. And again, I can't thank you enough, Steve, for the time and, you know, all this information you shared with us. Also, you know, uh, m- not only the information, I think, you know, this enlightenments that you, you, you shared with me personally and with my audience.

Um, so thank you very much for your time. Thank you. And this is how I close my, uh, episodes as usual. If you just discovered this podcast by luck, thank you for passing by. I hope you [00:54:00] enjoyed. If you did so, give me a favor, share it with your friends and colleagues. Share it with as many people as you can, because as you see today, topic is very, very important.

It's a tech, uh, it's a tech podcast, yes, but we try also to cover all the aspects of the tech, and mental health, anything related to psychology of, of founders, of tech executives is, is, uh, is, is, is, you know, very important. Mm-hmm. So share it with as much people as you can. And if you are one of the people who keep coming again and again, um, thank you for passing by.

It's like the... You know, thank you for keep listening. Uh, I appreciate your support, your feedback, and I appreciate also, you know, you take the podcast always to new levels, so we, we see the podcast, um, you know, trending in some countries, uh, recently, which we didn't been, um, lucky enough to be there yet. So thank you again for the support.

And as I say, always stay tuned for a new episode very soon. Thank you. Bye bye.