On Precision Psychiatry: How Recent Neurobiological Breakthroughs Can Improve Psychiatric Practice

In this episode, Dr. Daniel Knoepflmacher is joined by Dr. Conor Liston to describe recent neuroscientific breakthroughs that are transforming how we diagnose and treat common psychiatric conditions, like depression. Learn how research advances in brain imaging and neuromodulation investigated at Weill Cornell Medicine are leading to novel approaches for providing targeted and effective mental healthcare. 

 

Conor Liston, M.D, Ph.D., is a Professor of Psychiatry and of Neuroscience at Weill Cornell Medicine. Dr. Liston’s research aims to define mechanisms by which prefrontal cortical brain circuits support learning, memory and motivation, and to understand how these functions are disrupted in depression and other stress-related psychiatric disorders.

On Precision Psychiatry: How Recent Neurobiological Breakthroughs Can Improve Psychiatric Practice
Featured Speaker:
Conor Liston, M.D., Ph.D.

Conor Liston, M.D, Ph.D., is a Professor of Psychiatry and of Neuroscience at Weill Cornell Medicine. Dr. Liston’s research aims to define mechanisms by which prefrontal cortical brain circuits support learning, memory and motivation, and to understand how these functions are disrupted in depression and other stress-related psychiatric disorders

Transcription:
On Precision Psychiatry: How Recent Neurobiological Breakthroughs Can Improve Psychiatric Practice

Dr. Daniel Knoepflmacher (Host): Hello, and welcome to On the Mind, the official podcast of the Weill Cornell Medicine Department of Psychiatry. I'm your host, Dr. Daniel Knoepflmacher. In each episode, I speak with experts in various aspects of psychiatry, psychotherapy, neuroscientific research, and other important topics on the mind.


Our topic today is precision psychiatry. In many medical specialties, doctors employ a combination of clinical examination, medical history-taking, and objective tests to determine the correct diagnosis and choose an appropriate treatment. For instance, to definitively diagnose diabetes, physicians often use a simple blood test called the hemoglobin A1c. In psychiatry, we use a comprehensive, structured interview that includes clinical observation and extensive review of the patient's history. But when it comes to diagnoses like major depressive disorder or bipolar disorder, we lack the types of objective biological markers that are widely available in other medical specialties. For years, many research efforts in psychiatry have been focused on identifying biological measures that can be used to make definitive diagnoses and offer patients more efficacious and personalized treatments. This is the goal of what is called precision psychiatry. But the fact is, we aren't there yet.


Today, we'll try to understand exactly how far are we from reaching this promised land as we discuss some of the exciting new scientific breakthroughs in the ongoing quest for precision psychiatry. To guide us on our journey, I'm thrilled to be joined by Dr. Conor Liston, who is a Professor of Neuroscience and Psychiatry here at Weill Cornell Medicine and winner of the 2023 Thomas Williams Salmon Medal in Psychiatry from the New York Academy of Medicine. He leads a research lab focused on the neurobiology underlying various psychiatric disorders, including important work that could revolutionize the diagnosis and treatment of depression. Conor, thank you so much for joining me on the podcast today.


Conor Liston: Thanks so much for having me. I'm really excited to be here.


Dr. Daniel Knoepflmacher: Well, we have a lot to talk about. And I want to begin by asking you, what led you to specialize in this area of research? Were there specific reasons why you gravitated towards neurobiological research of psychiatric disorders?


Conor Liston: Yeah, you know, that's a great question. I think for me, like, many physician scientists as you mentioned, I'm a neuroscientist and a psychiatrist. And when I first came to MD PhD training, which was actually here at Weill Cornell and at the Rockefeller University in our Tri-I program, psychiatry wasn't really on my mind, to be honest. But I also came in with an open mind about where my interests might lead me. And it was really through research work I did with the late Bruce McEwen at Rockefeller, a pioneer in neuroendocrinology and our understanding of how stress affects the brain, and with BJ Casey, a professor at the time here at Weill Cornell, who's a leader in developmental cognitive neuroscience and understanding how developments in the brain give rise to psychiatric conditions. It was really my experiences in their labs that kind of piqued my interest in psychiatry.


And a couple of things were happening around that time that drew me further. One, as I was thinking about specialties, the field of neuroscience and its interface with psychiatry was just beginning to be transformed by new technologies that were allowing us to ask questions that we couldn't really ask before. One of those is optogenetics. We could talk at length about optogenetics if you're interested, but the upshot is developed by leaders at Stanford and many other institutions across the country. These tools use light to control the activity of brain cells in a really specific and precise kind of way and allow us to ask questions about how brain circuits support behavior, cognition, functions that are of interest to us in psychiatry. And for the first time, so many questions in neuroscience that were pertinent to psychiatry seemed experimentally tractable. And that seemed really appealing to me as a researcher considering different clinical fields.


And at the same time, I also learned that I really liked psychiatry, the practice of psychiatry, in ways that I didn't necessarily expect. I think I went into my clinical clerkships as a medical student with the misconception that our incomplete understanding of the neurobiology of psychiatric conditions would be a real hindrance to helping patients. And my experience from day one was that that was not at all the case. I think, you're a clinician, you can relate to this, there's many ways that we wish we could do better in psychiatry. But the fact is there's so many people were really able to help. I was able to witness that on my very first day in psychiatry and throughout my experience. And it really drew me to the clinical practice with psychiatry as well, being able to help people with the promise of research potentially down the road, transforming the way we think about diagnosis and treatment in psychiatry. And that really kind of sealed the decision for me.


Dr. Daniel Knoepflmacher: Well, it sounds like it set you up perfectly for research, because you saw treatments where empirically people were getting better. Yet when you asked your senior psychiatrist exactly what was going on neurobiologically, they may not have been able to tell you.


Conor Liston: That's exactly right. Yeah, we had hints, right? And we still do. There's a solid foundation to build on, but there's so much to learn. And in a way, it's really encouraging and gratifying that the treatments that we have work as well as they do for so many individuals even though we don't necessarily understand everything about the biology of mental illness, and how could we? The brain is an extraordinarily complex organ. This applies not just to psychiatry, but to neurology and to many other medical specialties. We're just beginning to scratch the surface in our understanding of the mechanisms that give rise to these conditions. And I, for one, am really excited for what the future holds for psychiatry and how we might be able to help even a larger number of people more effectively.


Dr. Daniel Knoepflmacher: That really relates to our topic today about the promise of precision psychiatry, and it remains a promise for the future right now, not exactly reality on the ground. So, could you tell us a little bit about how you see the status quo, and then where diagnosis and treatment in psychiatry can be markedly improved?


Conor Liston: Absolutely, yeah. And that was of the questions that drew me to psychiatry in the first place. My interest in this topic actually dates back to my experience as a medical student. I bet you probably had a similar experience. I remember being a little puzzled about how and why we diagnose patients the way we do.


Depression is a great example. It happens to be the one condition that we study a lot in my team. We diagnose depression today when a patient presents with five or more of nine symptoms. Doing some quick math, that means there's at least 256 unique symptom combinations that a patient can present with and still get this diagnosis. And that's not even accounting for the fact that some of the criteria contain opposites of themselves like sleeping too little or sleeping too much, weight loss or weight gain. And I think it's sort of intuitive that someone who's lost 40 pounds, sleeping four hours a night, intensely anxious, agitated, low mood, this person may not have the exact same biological problem as someone who's gained 30 pounds, sleeping 20 hours a day, not at all anxious, just kind of physically slowed, can't get out of bed, profoundly anhedonic. These people, if they were sitting in your office together and we were going through an inventory of symptoms, they would look almost like opposites of one another, and yet they get the same diagnosis, and often they get similar treatments. And again, it's really a miracle that those same treatments work as well as they do for these very different people. But I think it's intuitive as well that they may not have Identical biological substrates that give rise to the symptoms that they're suffering from.


And so, one kind of promise of precision psychiatry is to rethink how we diagnose depression and other conditions. Identify perhaps subgroups of depression, or even down the road, wholly new diagnostic categories that are more anchored in biology that might open new avenues for exploring mechanisms that might be operative in one group of patients with depression, but not in another group of patients. And eventually, kind of matching treatments to subgroups of patients who are most likely to benefit from them and hopefully designing fundamentally new treatments informed by this kind of new mechanistic understanding. That's where I think precision psychiatry is going.


But you're absolutely right, we're not there yet. There's a lot of exciting developments in the field. Every year, I feel like the pace is accelerating. More and more work from groups here at Cornell, but all over the country are pushing this field forward. But in practice, precision psychiatry isn't something that's really ready or seeped into the actual clinical practice of psychiatry, and for good reason, because we really need to test precision psychiatry approaches, make sure they work, understand for whom they work well, understand under what circumstances they don't work as well, before we consider deploying them at wide scale. But we are making progress. We have trials underway. I think there's a lot of reason for hope and for excitement in this area.


Dr. Daniel Knoepflmacher: There's so many conditions within psychiatry and perhaps the precision can begin in a certain area and then extend to other conditions. I know you've done a lot of work on depression and you alluded to subtypes of depression. One question I have before I want to get into your actual important work subtyping depression is that, historically, before those criteria that you mentioned where you go for five out of nine, in the old days, there were psychiatrists who sat and really looked at a lot of different patients and actually came up with subtypes based on what they were seeing empirically. I'm wondering how much that has been helpful for you as you're doing work that's going looking into the biology that's perhaps underlying these different subtypes.


Conor Liston: It's been a huge help. And you're absolutely right, this is a really rich field that dates back decades. Arguably, if you go back to like Kraepelin in the 19th century, like the same idea of subtyping patients presenting with superficially similar symptoms, it's been around for a long time. I think, painting with broad strokes, the general approach has been to look for groups of patients who tend to present with similar sets of clinical symptoms and subgroup them in that way and then ask whether those subgroups correspond to different kinds of biological markers, something you might measure in the blood or in a brain scan or perhaps a different likelihood of responding to a different kind of treatment. That work has been profoundly influential to take a couple of examples like atypical depression and melancholic depression, our clinical subtypes where we've learned a lot about biological processes that might underlie the conditions about on average tendencies of patients, for example, atypical depression patients may respond better to atypical antidepressant strategies like monoamine oxidase inhibitors and other approaches. Seasonal depression is one especially around this time of year in northern latitudes that we're all aware of. It's seeped into the public lexicon. It's been very influential work.


But one of the challenges with those approaches is that it's been difficult to map any of them onto a biomarker that is useful in widespread clinical practice, right? And I think that's not controversial and pretty intuitive, like as you said, we don't really use biomarkers much in psychiatry because there's no evidence that really supports widespread use. And there are exceptions like melatonin measurements may be useful in a subgroup of people with seasonal depression and melatonin-targeting treatments may be useful for those individuals. But speaking broadly as a general rule, it's been challenging to translate that kind of clinical subtyping approach into biomarker discovery and, fundamentally, new treatments.


But I will say, as you alluded to, the approaches that investigators in this area have been taking for decades have provided kind of like a roadmap for how we can think about these problems and how we might make continued progress on them in future kind of building on that foundation.


Dr. Daniel Knoepflmacher: Well, let's talk about your work then. So, you specifically done work using neuroimaging to subtype depression. So, let's hear from you about that.


Conor Liston: Yeah, absolutely. So, we're very excited about this work. And it was directly informed as you alluded to by kind of earlier approaches to clinical subtyping. Our approach, instead of grouping people by symptoms and asking whether we can identify biological markers associated with those subgroups, has been kind of to flip that upside down and ask, "Let's try to group people by biology." And then, let's ask whether those biological subgroups are associated with different kinds of clinical symptoms, different kinds of treatment outcomes and whether we can use that information to kind of inform decision-making in psychiatry down the road with prospective testing.


You can imagine using many different kinds of biological markers. Friends and colleagues up at Sinai are doing great work with inflammatory markers, for example. In my lab, we're particularly interested in resting state fMRI measures of functional connectivity in the brain. This technology is similar to the kind of MRI scan one might get if you went to an emergency room with signs of a stroke or a severe head injury. But this kind of MRI scan has been sensitized to neuronal activity essentially. And based on seminal discoveries by Bharat Biswal and Mark Raichle and others 25 years ago, we've discovered that the brain at rest exhibits these spontaneous fluctuations in the fMRI, what we call the bold signal, the signal that we're measuring from an fMRI scan. And regions of the brain that are strongly connected to one another tend to fluctuate together. And so, you can measure how correlated those fluctuations are to create what is essentially a map of how connected every brain region is with every other brain region. And it's a little bit of an oversimplification, but I think it's a useful way of thinking about it.


One way of kind of thinking of an analogy to this approach is our nation's airport map. Our airports are organized into kind of hubs and spokes and we think the brain networks may be organized in a similar kind of hub-and-spoke-like fashion. And this hub-based organization for networks is associated with all sorts of kind of predictable properties. And one of them, like in an airport network, we can explain why on a beautiful day, it's sunny today in New York City, you could be sitting on the runway at JFK endlessly waiting for your flight to take off because there's a blizzard in Chicago O'Hare and the chaos in that hub percolates out into the rest of the network and can affect many airports all over the place. We think, in some ways, something similar might be happening in depression and other psychiatric conditions. It's not like the gross kind of anatomical structure of the brain is dramatically altered. If you get most brain scans, their brain is going to look perfectly normal with depression or schizophrenia or autism. But the way the different brain regions are connected has been changed and dysfunction in one kind of hub can percolate out into the rest of the network and cause dysfunction all over the place. This is something we think we can map with fMRI. And our basic approach has been to try to take those maps. And ask whether different subgroups of patients have particular kinds of connectivity abnormalities or changes that are associated with particular kinds of symptoms. And if so, can we cluster people, can we group them on that basis using machine learning methods and a kind of data driven-approach to just tell us really what the data are revealing to us about how people group together. And then having identified those subgroups, we can try to ask questions like, are they associated with different kinds of clinical symptoms? Do they respond differently to different treatments? And we're finding that they do. Happy to dig into that in greater detail, but that's the general idea. By defining these subgroups, we can build biomarkers, statistical classifiers for diagnosing them in individual patients. And hopefully, down the road, matching them to treatments that are most likely to work for one subtype, but not for another subtype.


Dr. Daniel Knoepflmacher: I want to talk about the treatments piece in a moment, but sticking with the subtyping. So just so I understand, I'll give two different types of depression. Somebody who has a more agitated, anxious state, not sleeping; somebody else, you mentioned atypical depression who has what we call hypersomnia, sleeping excessively; maybe eating a lot of carbohydrate-rich foods, the other person who is not hardly eating at all. These, you're saying that looking at a functional MRI of these two types of depression, that you're going to see a different pattern of connections in the brain in this hub and spoke model that sort of is similar among different individuals within each of those types of depression.


Conor Liston: That's right. Except one what's kind of interesting is that the people with atypical depression and the people with other forms of other clinical subtypes of depression, their brain maps actually aren't that different. These differences only become clear if you start by grouping them based on their brain maps. And then when you have them grouped in that way, you can ask whether they differ with respect to clinical symptoms, and they do. But not in a straightforward way that maps on to atypical depression or seasonal depression, for example. I can squint at our results, and I can sort of see findings reminiscent of atypical depression, for example, in some of the subtypes or a more typical, like agitated anxious depression, like you said, in another subtype.


But the fact is that you couldn't really diagnose these subtypes based solely on asking about symptoms like anxiety and appetite, that the brain scan and its quantification of connections tells you something special that you can't get from symptoms. And we're getting a little bit into the weeds here, but I think it's kind of interesting. What we've learned as we've developed the approach is what these subtypes define, we think, with respect to symptoms, is a propensity to express a certain combination of symptoms when you get sick. But they don't necessarily dictate in a one to one, 100% kind of correspondence exactly what symptoms a particular person is going to experience, which I think is why that mapping of, say, anxious-agitated depression to a subtype isn't so clear. Does that make sense?


Dr. Daniel Knoepflmacher: Yeah. Yeah. So, it's not just like a fingerprint for agitated depression that really differentiates out from a melancholic depression.


Conor Liston: That's right.


Dr. Daniel Knoepflmacher: And I'm imagining you were talking about treatment, having something that shows a pattern of connectivity then allows you to tailor treatment.


Conor Liston: That's exactly right. And the first way that we've embarked on that approach uses a treatment called transcranial magnetic stimulation, which many of your listeners are probably familiar with. I think it's been around for decades now. It's only more recently, I think, that it's become a more widespread option that patients have access to, and that doctors are more likely to consider. Transcranial magnetic stimulation, or TMS, works by delivering high-frequency, high- potency magnetic field pulses across the scalp, which elicits an electric field in the brain that activates neurons.


And we've known for some time that functional connectivity measured by fMRI in the way I described before mapped from the target in the brain that you're trying to stimulate, the most common one being the left dorsolateral prefrontal cortex, kind of on the left side above the ear. Another target people often use is the dorsomedial prefrontal cortex, which lies like right along the midline where the scalp meets the forehead. We've known that functional connectivity mapped from those targets is at least modestly predictive of your likelihood of responding to that treatment. And I think that kind of makes sense intuitively, because if you imagine stimulating the dorsolateral prefrontal cortex, that activates the neurons at the targets. And then, those neurons will in turn activate other neurons that they're connected to, and the degree to which they're connected to neurons in regions A, B, or C might differ in different individuals and might determine differences in kind of the network level effects of the stimulation. That's the hypothesis anyway.


And so, we reasoned one of the observations that kind of stood out in our subtyping approach, the original one published now six years ago, was that the dorsomedial prefrontal cortex, this area that I mentioned along the midline where the scalp meets the forehead, that this area has very different connections across the different subtypes. And we hypothesized that the four subtypes that we identified might respond differently to that treatment. And that's what we found that one of the subtypes was very responsive, that two of the subtypes were not very responsive. And that the fourth subtype showed kind of an intermediate level of responsiveness. And so, that got us very excited. It was an encouraging form of external validation that the subtyping approach that we were trying had legs, had some utility and was likely to lead to something interesting. But it was also potential practical clinical implications. If you can predict who's going to respond to the TMS that this targets, that's useful information. It's useful if you know a person's going to respond, maybe you would consider that treatment for them earlier. If you know a patient isn't going to respond, maybe you would not consider that treatment and consider other alternatives for that individual.


And so, we're now embarking on a trial to test how well that really works. We have algorithms for predicting response to that dorsomedial target, as well as a left dorsolateral prefrontal target. And my colleagues and I in the interventional psychiatry program led by Ben Zebley and Emanuel Elbau and Lindsay Victoria are playing a big role in leading those studies as well, along with Faith Gunning, along with many others, I don't have time to name all of them. But we are embarking on a trial to test how well that performs, and it's going to tell us a lot of useful information, not just about how accurate the biomarkers are, you know, that's critical, obviously. But there are scenarios in which even a 100% accurate biomarker might not be clinically useful. If, for example, everyone in our sample is highly likely to respond to both treatments, you don't need a biomarker, right? Like, you can just roll the dice. And so, what you really want is a biomarker that differentiates people who respond to one and not the other. And we'll see how well they perform. Any of your listeners out there, we are currently recruiting for that trial. The good news is that everybody gets an active treatment. There is no sham or placebo control in this study because the goal is just to test the biomarker. So, everyone gets an active treatment. It's free of charge to all of the participants who are eligible. And we'd love to hear from you if you think you might have good candidates for that.


Dr. Daniel Knoepflmacher: So, you're recruiting candidates for this. And the treatment, just so I understand, people are coming in with depression and then you're evaluating them. You are going to be doing functional MRIs. And at that point, there are going to be subtypes. So, you're not trying to recruit for a single subtype, you're recruiting for depression. And then, you're doing both targets, or one of those targets when you're doing the TMS?


Conor Liston: That's right. So, everybody, all people with depression are welcome. We're not recruiting for a particular subtype. Everybody who comes in gets a deeply phenotyped clinical assessment, evaluated by our team of psychiatrists and clinical psychologists. They get a brain scan. They get subtyped. We apply our biomarkers, our algorithms for predicting response. And then, every participant gets randomized to receive one of the two treatments. Both of them active, both of them roughly equal in their efficacy rates in large scale studies. In that way, we'll be able to tell how well the biomarker performs and how useful it is for improving outcomes in the whole population based on its performance.


Dr. Daniel Knoepflmacher: So, there could be a future where people with depression have a structured interview, get a functional MRI, and then get a specific TMS treatment that is going to have a much higher efficacy rate than a TMS treatment that isn't using a subtype and trying to link the two together is the current practice.


Conor Liston: That's exactly right. And we think we might not be that far away from that. We're excited about it. We'll know for sure how well it works. We're blinded, so we won't know for sure until the study ends and that's the way it should be. But if it works out in the way that our preliminary data suggests, it could be really helpful. Mainly, there are some people who are going to get better in response to either treatment and there are others who are going to not get better in response to either treatment. But there's, we think, a solid group of people who are likely to respond just to one of them. And those are the people we think we're really going to be able to help. And down the road, if you know the biomarker performs well, you can imagine scenarios where TMS might be considered earlier for some patients. That's a big issue as well. Access to TMS is typically reserved for patients who've tried and not responded well to at least two, sometimes three, four, or five trials of an antidepressant or structured psychotherapy of adequate dose and duration. And I think there's a world in the future potentially where biomarkers might allow patients to be stratified to the treatments most likely to work for them more quickly without this kind of trial and error approach that, as you know, as a clinician, is frustrating for patients and for doctors too.


Dr. Daniel Knoepflmacher: To make our practice more precise, which is the focus of today's discussion. Well, I want to ask you about the TMS specifically, because TMS itself is undergoing some advances in terms of how it's delivered. And I know your research is also looking into that as well. So, could you speak to that please?


Conor Liston: Absolutely. So, there's a couple of exciting developments there. I'll talk about at least two of them. One of them is based on really important work from a friend and colleague at Stanford, Nolan Williams, who developed a new protocol for TMS. It's an accelerated intensive form of TMS. That basically works by kind of tweaking the protocol, the parameters by which TMS is delivered based on kind of first principles derived from biology. And one of the key insights is that you can deliver a lot of treatment in a shorter period of time in this accelerated intensive form of treatment that Nolan and his team call SAINT. Patients get essentially five, six weeks of treatment in a single day. They come in the morning. They get treated on the hour every hour for ten sessions. They leave at the end of the day, so it's an outpatient, they go home. At 1165 York, we have intersession rooms. They can even lounge, relax, do some reading, they have time in between treatments. But the key point is they get treated on the hour, every hour, for ten hours. They do five or six weeks of treatment in a single day, and then they repeat that Monday through Friday. So, they get a big dose of TMS. And Nolan and his colleagues have observed really dramatic response rates in a short period of time, sometimes in as little as a day, sometimes by the end of the five-day course of treatment. And in some cases, you see part of the effect at the end of the five days, but continued kind of added benefits in the second week after treatment.


And we and the Stanford team led by Nolan and another team at UCSD led by Jeff Daskalakis, we're developing new approaches to predicting who is most likely to benefit from this particular kind of treatment? Which particular symptom domains like anhedonia or anxiety are going to be most responsive? Who's going to be able to obtain a prolonged sustained response, whereas others perhaps might need more frequent treatments. So, we're building better models basically that will allow us to deliver this accelerated form of TMS more rapidly and more efficiently. And again, hopefully, with the right modeling and the right perspective kind of validation data, make it more available and accessible to people sooner in their care.


So, that's one approach we're testing, a similar approach now, kind of at an earlier stage in development for OCD, obsessive compulsive disorder, with very encouraging results. It looks really quite interesting. OCD, as you know, is also very difficult to treat for many patients and getting the right treatment can be a challenge and we're seeing really very encouraging results in, at least, a sizable subset of the patients coming through our trial. We'll be building models around that as well.


And then a kind of second area, in addition to kind of like accelerating the treatment, my colleagues and I in the Interventional Psychiatry Program, in particular I'd highlight Chuck Lynch's work. We're developing new approaches for what the field calls precision functional mapping that could be used to inform TMS and deliver TMS in a totally different way in the future. The basic idea here with precision functional mapping is that those networks that I described earlier, like our airport network, it turns out there's another twist in how they're organized. The field discovered this just a few years ago. There are commonalities that we all share, kind of like the human face. We all have two eyes and a nose and a mouth in approximately the same location. Those kind of group level kind of shared characteristics is what the field has focused on for decades. But we've now learned that there are also these important individual differences. And so, you can average a bunch of people's faces together, you can get something that looks readily recognizable as a face. But that average glosses over all of the individual differences in our faces that are so important in kind of recognizing, distinguishing a stranger from a family member or friend, determining whether your friend is happy or sad or angry. These individual differences are so informative and it's only just recently that the field is beginning to appreciate that they exist at all and that perhaps they can be kind of reliably mapped and quantified and might even be used to inform the way that we deliver treatment. So, we're developing in the interventional psychiatry program, new ways of of delivering TMS based on these kind of precision functional mapping approaches, targeting the precise location of a particular brain network in your brain, which might be quite different than the brain of three other people. And that's something that we're very excited about as well.


Dr. Daniel Knoepflmacher: That is exciting. So, there's a personalization for each individual and you can make the treatment more precise based on those variations. Something that we haven't really looked into when we've been delivering TMS in the past.


Conor Liston: That's exactly right. Yeah. And it looks quite promising. Chuck and his colleagues have found, for example, that this network, which we call the salience network, no one really knew to look at the data in this way, which makes Chuck's approach so special. But he found by mapping this network in the brains of several hundred people with depression and scanning them repeatedly over time, Chuck has scanned some of these individuals 40, 60, 70 times over the course of a couple of years, he's found that this network is expanded in individuals with depression to the extent that you can actually see it with the naked eye. It's like nothing I've ever seen in psychiatric neuroscience.


You don't need fancy statistics, you can just notice this is bigger than this, sometimes two to three-fold bigger than what you typically see in people without depression. And so far, analyses suggest that this may be a marker of risk for becoming depressed. Chuck has also found that this expansion of this network is detectable in kids, which is really interesting, who go on to become depressed later in life compared to kids who don't get depressed later in life. So, you can scan them at ages 10 and 12 in kids who've never been depressed before. And then, looking at those individuals who go on to become depressed at ages 13 and 14, it turns out that their salience networks were bigger than normal before they ever got depressed. So, we think it might be a marker of risk and that's something we're looking into. And you can imagine if that's true, there's all sorts of ways you might imagine using tools like TMS, perhaps also tools like ketamine to engage these network differences and perhaps even do in ways that might prevent the recurrence of depression in individuals at risk for kind of recurrent episodes.


Dr. Daniel Knoepflmacher: You just brought up ketamine and we've been talking a lot about TMS. Can you speak a bit about potentially for ketamine where, again, there can be biomarkers involved or other more tailored treatments?


Conor Liston: Absolutely. We're very excited about this as well. So, we're launching both a research program and an outpatients clinical treatment program focused on ketamine and S ketamine, an enantiomer version of ketamine in the interventional psychiatry program. And people who are interested in that should contact Ben Zebley. That will be launching soon. But the basic idea there and the reason we were sort of drawn to it from a research perspective is that we know that ketamine and TMS are often offered to the same kinds of depression patients. And I'm sure that's probably been your experience too. You'd probably consider SSRIs and CBT and perhaps bupropion, MAOIs. But once patients kind of haven't responded to several trials of those kind of first line agents, that's when you begin to consider a treatment like TMS. I would say with the FDA approval of esketamine or Spravato a few years ago now, ketamine is kind of the main alternative for treatment-resistant depression in patients who might not need ECT yet, but clearly aren't getting better in response to first-line agents.


But a big challenge is how do you deliver ketamine effectively and who needs ketamine versus who needs TMS? And so, we've built models that we think can begin to predict who will respond most effectively to ketamine and who will respond most effectively to TMS. These are in early stages. We are going to be developing these models further. My colleague, Logan Grosenick, is big part of that, a computational neuroscientist in Psychiatry here at Weill Cornell who has been working with us to build these models. And hopefully, if some grant applications go well, we'll be testing how well those models perform, tweaking them a little further with the goal ultimately of having tools that doctors could use to map connectivity in the brain subtype their patients, and in addition to a quantification of clinical symptoms, get some information that tells them rapidly, we think the likelihood of your patient responding to an SSRI is low. We think the likelihood of responding to ketamine is very high. We think the likelihood of responding to TMS is moderate. And therefore, we recommend to try ketamine first. If that doesn't work, consider TMS. And we think we're getting there. Lots of testing to do still, but we're on our way.


Dr. Daniel Knoepflmacher: Wow. Lots of exciting developments that you're working in so many areas and trying to push this quest to precision psychiatry forward. I guess I'm going to ask you one last somewhat annoying question, because predictions are never accurate or easy to make, but thinking about some of these developments, when do you think they would be ready for prime time? When here at Weill Cornell and elsewhere will psychiatrists be able to do what you just described in terms of choosing a treatment for depression or even some of the more rapid TMS being more widely available? What are your predictions for that?


Conor Liston: Yeah. So, one way to answer that is that the biomarkers in our group that are kind of in the furthest stage of development are the ones I described first for predicting response to two different brain targets in patients undergoing conventional TMS. And we're currently doing a randomized controlled trial, a prospective test of the biomarkers. We're about halfway done with that. And so, we will have an answer to how well they perform in probably two years or so, two to three years. And if they perform as well as we hope, we've already been in contact with NIH, various regulatory agencies, the FDA about how to design things in a way that we could seek approval for these approaches and make them more widely available. So, I would say, we'll have a sense in a couple of years as to your question of timeline as to how promising this approach is in the near term. And if all goes well, on the order of five years perhaps, they might be more widely available to patients.


As for the accelerated TMS approaches, that will be available much sooner. So, it's those treatments, the SAINTS Accelerated TMS for Depression, the version that we're testing for OCD, those are available in research settings here at Cornell right now. For any patient who wants, again, reach out to us. We're always accepting referrals. And the clinical service, which will launch in the first quarter of 2024, will definitely be offering those services for depression. I think a little more data will be needed before accelerated TMS is offered for OCD, but that will be clinically available for depression next year.


Dr. Daniel Knoepflmacher: Wow. Well, Conor, I could spend another hour delving into this with you, but we've run out of time. I'm just so grateful to you for sharing this amazing work you're doing. I don't know how you're able to do so many things at once. But just today, thank you for making the complexity and the promise of all of this neurobiological research easier to understand for me and for our audience.


Your research is really exciting. I mean, I think it provides hope for a day, and maybe not that far away, when we're going to be offering rapidly effective treatments to people who are struggling with treatment-refractory depression with OCD. I know that I'd love to have you back to continue this conversation, maybe talk about some other things like psychedelics, which I think we could have a really good conversation about. So, just thank you so much for joining me today.


Conor Liston: Well, thanks, Daniel. Thanks for inviting me. It's been really fun and I'd love to come back anytime. So, I look forward to future conversations.


Dr. Daniel Knoepflmacher: Awesome. We're going to do that. And thank you to all who listened to this episode of On The Mind, the official podcast of the Weill Cornell Medicine Department of Psychiatry. Our podcast is available on all major audio streaming platforms, including Spotify, Apple Podcasts, and iHeart Radio. If you like what you heard today, tell your friends, give us a rating and subscribe so you can stay up-to-tdate with all of our latest topics and fantastic guests. We'll be back soon with another episode.


Promo: Back To Health is your source for the latest in health, wellness and medical care for the whole family. Our team of world-renowned physicians at Weill Cornell Medicine are having in-depth conversations, covering trending health topics, wellness tips and medical breakthroughs. With a spotlight on our collaborative approach to patient care, this series will present cutting-edge treatments, innovative therapies, as well as real life stories that will answer common questions for both patients and their caregivers. Subscribe wherever you listen to podcasts. Also, don't forget to rate us five stars.


disclaimer: All information contained in this podcast is intended for informational and educational purposes. The information is not intended nor suited to be a replacement or substitute for professional medical treatment or for professional medical advice relative to a specific medical question or condition. We urge you to always seek the advice of your physician or medical professional with respect to your medical condition or questions. Weill Cornell Medicine makes no warranty, guarantee, or representation as to the accuracy or sufficiency of the information featured in this podcast. And any reliance on such information is done at your own risk.


Participants may have consulting, equity, board membership, or other relationships with pharmaceutical, biotech, or device companies unrelated to their role in this podcast. No payments have been made by any company to endorse any treatments, devices or procedures. And Weill Cornell Medicine does not endorse, approve, or recommend any product, service, or entity mentioned in this podcast.


Opinions expressed in this podcast are those of the speaker and do not represent the perspectives of Weill Cornell Medicine as an institution.