Science Friday - Why Aren’t There Biomarkers For Mental Illness?

Episode Date: February 23, 2026

Despite major advances in our understanding of the biology of mental health disorders,  there’s no blood test or brain scan that will confirm if you have depression, anxiety, PTSD, or any other psychiatric illness.  And yet, the American Psychiatric Association recently announced that it will be including biomarkers for mental conditions in the next edition of the Diagnostic and Statistical Manual of Mental Disorders (DSM), which guides diagnosis and treatment of mental illness. So how close are we to pinpointing the biological markers of mental illness, and what does that mean for diagnosis? It’s complicated.  Host Flora Lichtman untangles some of this science with psychiatry researcher John Krystal. Guest: Dr. John Krystal is a professor of psychiatry, neuroscience, and psychology at the Yale School of Medicine. Transcripts for each episode are available within 1-3 days at sciencefriday.com. Subscribe to this podcast. Plus, to stay updated on all things science, sign up for Science Friday's newsletters.

Hey, it's Flora Lichten, and you're listening to Science Friday. Today in the show, despite major advances in understanding the biology of mental health disorders, there's no blood test or brain scan that will confirm if you have depression, anxiety, PTSD, or other psychiatric illnesses. And yet, the next version of our Diagnostic Bible, the DSM, the Diagnostic and Statistical Manual of Mental Disorders, will include biomarkers. So how close are we to pinpointing the biological markers of mental illness and what does that mean for diagnosis? It's complicated. Here to untangle some of the science for us is Dr. John Crystal, Professor of Psychiatry, Neuroscience, and Psychology at the Yale School of Medicine based in New Haven, Connecticut. John, thanks for being here.

Well, my pleasure, Flore. Okay. John, how hard is this problem, the biomarker for mental illness problem? It's an enormously challenging problem. You know, some people say that the brain is the most complicated structure in the universe, and we're trying to decipher the codes that underlie the symptoms and challenges that people experience in their lives. And therefore, it's an enormously challenging problem that we've been working on for a long time. Well, why is defining biomarkers for mental illness so difficult compared to other illnesses? Well, one of the reasons is because we are usually not analyzing the exact tissue that we're interested in.

In other words, if you have a tumor, they would do a biopsy and analyze your cells to find out exactly what the molecular signatures of your illness are. And that helps to make sure that you get the treatment that you individually need. That's what we'd like to go for in psychiatry to apply something that people call precision medicine. But we have to do that without sampling your brain tissue because you probably don't want us to do that. I mean, if you could sample brain tissue, would you be able to pick out biomarkers? Like, is that even the way that we would be able to understand a psychiatric disorder? Believe it or not, we are analyzing.

the brain tissue. And it's an enormously complicated task, but we're not biopsying people's brains. What we're doing are that there are some people who have been extremely generous in donating their brains to science, and we're analyzing their brain tissue and identifying the signatures of individual cells, and progressively the molecular signatures of individual cells in particular circuits, in particular parts of the brain to help us unravel what is, I think, emerging as the first true molecular understanding of the biological basis of psychiatric disorders. And what disorders are we talking about? I mean, even for something like depression or anxiety? Absolutely. In fact, here in Connecticut, at the VA Connecticut Health Center in West Haven,

we've done the first molecular analysis of post-traumatic stress disorder and comparing it to a disorder like depression. Both illnesses have depression as a symptom, but it turns out that the underlying molecular signatures of depression and PTSD are somewhat, although they overlap a fair amount, there's still in important ways differ. Really? I mean, I wondered if one of the challenges is that mental health disorders vary so much, even within one condition and, you know, across conditions, you have these overlapping symptoms, does that make the search for biomarkers more difficult? Yes. I think that is one of the enormous challenges that we face. You know, I think that the DSM-5 has something like 270 or so different ways that you can make the criteria for major depression.

And that includes people who are sad all the time, or people who are generally in good mood but react strongly to bad news, or people who just lack energy and don't feel necessarily sad or depressed. So there are a lot of different ways to meet a given diagnostic criteria and probably many, many different, biologically different subtypes of problems like depression, or anxiety. Are there mental disorders that we have really reliable biomarkers for that we can use for diagnosis? Well, if you think of psychiatry broadly, as I do, then one of the disorders that psychiatrists are often involved in treating is Alzheimer's disease.

And there is a real success story where there is a pet scan you can do that helps to identify a level of a chemical that's accumulated in the brain associated with Alzheimer's disease to help to make the diagnosis of Alzheimer's disease in a very specific way. And what's important about this is that when you're developing memory problems and you have this emerging diagnosis of Alzheimer's disease, there's something that we can do for you now to slow the progress of your memory. memory problems. So in the ideal world, what we're trying to get to is that kind of precision in our diagnosis process so that we can get people the best help that we can.

Because it helps with treatment. Because it helps with treatment, exactly. And it also helps to frame expectations. When we know that someone is evolving Alzheimer's disease, that's important information that people need to help prepare their estate and their families and things like that. So understanding what you're going through is very empowering. And it helps people to cope as well. For other conditions, what biomarkers are you looking at? Is it like signals on fMRI or blood tests? And what seems promising? You know, I tell the applicants to our residency program that this is the the most exciting moment in the history of the field of psychiatry, literally because all of the

areas that you just mentioned, blood tests, genetic analyses, functional magnetic resonance imaging, structural brain scans, PET scans to measure the level of various proteins in the brain, etc., etc. We're using every kind of information that we can collect to try to better understand and subtype the various psychiatric diagnoses. And we're on the cusp of using data that people collect, like their daily pattern of activity, where they go, the level of social interaction, much more detailed, granular behavioral information

to integrate with all this other information to help us to better inform diagnosis and treatment. The genetics must be so difficult, though, right? Because if I understand correctly, there's so many genetic variations that can play a role in mental illness. It's not like a gene, right? You're absolutely right. In fact, there are many different ways that genetics can increase the risk for psychiatric disorder. Probably the one that we've made the most progress on so far is schizophrenia. So about three to five percent of the people who have schizophrenia have a rare, rare, rare mutation in a gene that increases their vulnerability for schizophrenia, 50 or 100-fold even. But that's a very small percentage of the people with schizophrenia. Most people who develop schizophrenia have a whole collection of common variants that individually have a trivial contribution to their risk for developing schizophrenia. And there are about 450 different common variants in this way that contribute to the risk for

schizophrenia. 450. About 450, right, right. And so different people who are developing schizophrenia may have a collection of these various risk variants, but they may not have the same ones. And so that contributes to heterogeneity as well. Is the difficulty finding specific biomarkers for specific conditions, does that suggest that maybe the way that we categorize mental illness isn't quite right?

Yes. I think that we in psychiatry always view our diagnostic manual as a work in progress. You know, if we think of psychiatric disorders, the way you might think about breast cancer. There was a time when we diagnosed breast cancer by finding a lump and trying to do something. Now we get the tissue, genotype the tissue, find the mutations, and characterize other aspects of the tumor, and so we can really target in a very specific way. Right. There's not one kind of breast cancer we now know. Exactly. Exactly. And it makes a big difference in terms of the kind of treatment you get for breast cancer. In psychiatry, because we're

diagnosing today based on behavior, it's analogous to diagnosing breast cancer based on the lump. And so what we're trying to do is to get as much information to help us to get to the point where we can make diagnoses with the same precision for depression that we currently do for breast cancer. And I think that's going to come. I mean, given what we've been talking about, do you think the science is at a stage where it's ready to be incorporated into the next DSM? Or is that sort of aspirational and helps push the field along? It's mostly aspirational in the sense that few of the biological

findings that we have are really ready to implement in the average doctor's office in their everyday work with patients. I think, though, that one of the things that I admire about the process that the American Psychiatric Association is engaged in, is they're trying to set the stage for the emergence of many of these biomarkers as they matured, as they matured, so that psychiatric practice is anticipating their arrival and is ready to incorporate them in the diagnostic scheme. Dr. John Crystal, Professor of Psychiatry, Neuroscience, and Psychology at the Yale School of Medicine based in New Haven, Connecticut.

John, thank you so much for being here. My pleasure. Thanks for having me. Today's episode was produced by Shoshana Bucksbaum. And if stories like this help feed your brain, please leave us a review on Apple Podcasts or Spotify or wherever you listen. It helps other folks find the show. Thanks for listening. I'm Flora Lichtman.