From Our Neurons to Yours

Depression's distinctive fingerprints in the brain | Leanne Williams, Stanford University

• Wu Tsai Neurosciences Institute at Stanford University, Nicholas Weiler, Leanne Williams • Season 5 • Episode 6

Share episode

Getting help for depression can be like purgatory. Setting aside for a moment the stigma and other barriers to seeking treatment in the first place, finding the right combination of medication and/or therapy can be a months- or years-long process of trial and error. And for about one third of people, nothing seems to work.

Today we're talking with Dr. Leanne Williams, the founding director of the Stanford Center for Precision Mental Health and Wellness and Vincent V.C. Woo Professor in the Stanford Department of Psychiatry and Behavioral Sciences.

Williams and her team have recently used brain imaging and machine learning techniques to identify six distinct "biotypes" of depression — each of which may require a different approach to treatment. Beyond setting the stage for more targeted therapies, better understanding the biology behind the disease could finally cut through the stigma of one of the world's most common brain disorders.

Learn more

Episode Credits
This episode was produced by Michael Osborne, with production assistance by Morgan Honaker, and hosted by Nicholas Weiler. Art by Aimee Garza.


Send us a text!

Thanks for listening! If you're enjoying our show, please take a moment to give us a review on your podcast app of choice and share this episode with your friends. That's how we grow as a show and bring the stories of the frontiers of neuroscience to a wider audience.

Learn more about the Wu Tsai Neurosciences Institute at Stanford and follow us on Twitter, Facebook, and LinkedIn.

Nicholas Weiler:

This is From Our Neurons to Yours, a podcast from the Wu Tsai Neurosciences Institute at Stanford University, bringing you to the frontiers of brain science. Over the last few years, we have in many ways gotten a lot better when it comes to how we talk about mental health and depression. To some extent, we've normalized these conversations and that's a really good thing. At the same time, the word depression gets used a little casually, which can contribute to both ambiguity and stigma. All of us experience periods of sadness and grief in our lives. That's a normal part of being human. But depression is a brain disease, one whose precise definition at a biological and electrochemical level is only now coming into focus for researchers.

Today we're talking with Dr. Leanne Williams, the Vincent V.C. Woo Professor of Psychiatry and Behavioral Sciences at Stanford Medicine and Founding Director of the Stanford Center for Precision Mental Health and Wellness. Williams and her team have recently used brain imaging and machine learning techniques to identify six distinct biotypes of depression. Understanding these distinctions is a critical step to moving away from today's trial and error approach to depression treatment, a process that can take years, and towards a biology-first data-driven approach. I began our conversation by asking her to help us understand what is precision mental health.

Leanne Williams:

To me, the essence of precision mental health is about measurement, meaning we want to get more precise ways of understanding what it is that an individual experiencing depression, what are the underlying factors that we can measure? That's one aspect of the precision approach. As many people would know, typically, in psychiatry, we do not use measurement. We do sometimes take assessments of self-reported symptoms, but not typically measurement like might occur in other fields of medicine. The second aspect is about being more precise about the types of depression that individuals can experience rather than thinking of major depression as a one-size-fits-all umbrella category, and then having done that to be able to more precisely match their type to what treatment might be of most benefit. So it's those kind of three dimensions or factors that come together to make precision mental health.

Nicholas Weiler:

I also wanted to start by laying some groundwork here. I think a valuable question is to really think about what is depression as a mental health disorder? How is it different from the grief we may experience at the loss of a loved one or the sadness we may experience just at difficult times in our life? You were in a recent panel conversation with a Oakland comedian W. Kamau Bell, and I thought he made a very astute observation that recently in the COVID-19 lockdowns, many of us felt depressed, and in fact, I think I read a study that said that rates of depression in the U.S. may have tripled early on in COVID. But yeah, what's the difference between being depressed by circumstance and having major depression?

Leanne Williams:

Such an important question because the term depression itself doesn't really convey the nature of depression as a disorder or disease as opposed to an everyday emotion. Depression as a disorder or disease is differentiated by a couple of things. One is beyond persistent sadness. So we think of depression most commonly as persistent sadness and feeling down, but there are many other defining features, including the absence of being able to feel pleasure. So it's a kind of a numbness. And then other aspects of difficulty with cognition or thinking, appetite, sleep changes, and then the related crucial aspect that these experiences disrupt function. So they make it very difficult for the individual to function either at work or school or with their relationships.

The third piece is that it's typically persistent. So to qualify for a diagnosis of depression as a disorder, you would have those persistent experiences for at least two weeks, and many people have it for much longer and in a recurrent fashion. So it's commonly a chronic condition. If you're thinking about everyday depression, that's something that fluctuates. It's more of a mood. It doesn't tend to persist over time. It may interfere with our function temporarily, but not in a persistent way. You can then think of other experiences like grief. They may be more severe, but again, the defining aspect is the sadness. If you think of times when you may have experienced grief, within the waves of grief you can have moments of feeling pleasure, whereas in depression as an illness, that doesn't tend to happen. And we can talk about the brain regions. There are kind of different brain regions you see disrupted when it's depression as a disorder.

Nicholas Weiler:

Well, that's exactly what I wanted to ask you next. You have been very focused in your lab on understanding how do we measure depression? How do we see depression in the brain? So I'd love to speak for a little bit about how depression looks in the brain. What are you measuring that hasn't been measured before?

Leanne Williams:

We are measuring a few unique aspects of brain function that haven't been measured before. The aspects that are similar are basically the equipment we're using. So we're using fMRI, and that gives us a good picture of how the brain is effectively activated, like the amount of oxygen it's using to fuel what it's doing in different regions of the brain. And those different regions of the brain form what I'll call circuits. So these are kind of superhighways of connected regions that enable us to process and regulate emotions, to think, to self-reflect all of these aspects of being human that are really essential to our experience. In depression, these are the circuits that get disrupted, they get stuck, or they become underactive.

The unique aspects of what I'm doing are designed to overcome some of the barriers in using fMRI for clinical translation. So not only do we want to get fantastic quality data for neuroscience discovery, but we want to think of how do we develop a way of imaging the brain that could ultimately be used in practice. So to do that, I've developed a technology. It's the Stanford [inaudible 00:07:24] technology. What it does is takes that extraordinary complexity of the brain and all the multiple thousands of connections and focuses in on connections that are a manageable number of connections and that we know over and over are being shown to be disrupted in depression. So it gives us a manageable interpretable set. The second innovation that we are very proud of is developing a way to interpret the data for each individual patient or person. We can identify their profile of activity and connectivity in these circuits we are looking at relative to a healthy reference dataset. So we get back a number expressed as how different are their brain circuits from the healthy range.

Nicholas Weiler:

So you're focusing in on the particular networks that are known to be often misregulated in depression. And you're looking specifically at how are activity in these circuits differing from people who are not experiencing depression.

Leanne Williams:

Exactly. I measure six circuits that have been established in the field more broadly, within the imaging field using fMRI, there's consensus about these circuits, and they are also the ones most consistently implicated in depression and its treatment. So we measure those. I'm realizing it's been quite a while now, but for nearly two decades I've been measuring these circuits using standard imaging sequences, which has enabled me to pool data across multiple studies, and that's what's given us confidence about how accurately we can measure these aspects of the circuits at the individual level, how consistent they are, how generalizable and so on.

Nicholas Weiler:

And so one of the key takeaways that your team has published recently is identifying six, I think, distinct what you're calling biotypes of depression. Can you tell us a little bit about how are you defining these different biotypes? What do they represent?

Leanne Williams:

Great question. The biotype is, it's an increasingly used term in the field to define subtypes of depression that are distinct in terms of these circuits, meaning one subgroup will have something like underactivity of the circuit we need to be able to think properly and inhibit interfering information. Another will have too much activity in the circuit that responds to negative emotion. So these are distinctive profiles.

And how we went about it is we generated for our field a very large dataset, which is 801 people experiencing depression and anxiety and 137 healthy people. We quantified their brain circuits to get individual scores for their circuits. So each person has a set of scores and how much each of their scores differs from the healthy reference dataset. Those scores were then inputs to machine learning algorithms for clustering, and we used all the checks and verifications that we need to use to ensure that the solution we got was valid. The solution delivered was that there are six subtypes of these circuit profiles that are distinct from each other. Importantly though, we were able to also map them to a theoretical taxonomy that I'd proposed several years ago. So based on the knowledge in the field, I proposed how these circuits might form subtypes. So when we went to the data-driven approach, we looked at the correspondence and we found good correspondence between six of these biotypes and the eight that we'd theoretically proposed.

Nicholas Weiler:

These biotypes, so you mentioned that one might involve being overly sensitive to negative emotions. One might involve lack of regulation of intrusive thoughts and emotions, potentially. I'd love to hear more about the other types. And those two that you mentioned, I could sort of imagine how those would map on to how someone might experience depression. And so I'm curious, for the other types, do they also map on to how people might experience their depression at a cognitive and emotional level?

Leanne Williams:

Right. It's a really important question. They do map onto primarily the cognitive aspects of how you experience depression, and by cognitive I mean the emotional aspects of cognition as well. So we find mapping between the biotypes and symptoms on the one hand. So these are self-rated symptoms like you would get, say, in a clinic or a clinical trial. However, it's not a nice neat one-to-one mapping. For example, the first biotype we describe is where there's too much connectivity in the resting circuits of the brain. So there's one that's a famous circuit called the default mode network, and it's been commonly implicated in depression and in anxiety, and it's kind of too connected.

Nicholas Weiler:

Is that sort of a rumination circuit almost?

Leanne Williams:

Right. So it gives kind of more focus on negative thoughts and that circuit as well as a related one called the salient circuit, which detects emotionally salient information both internally and externally, those two map onto general severity, also onto symptoms of anxiety and worry. You then see that other types, and I can go through them, they more map onto measures that we've assessed behaviorally, and these are measures that you may not typically have in a clinic, but you often have in a trial where you're actually asking people to do different kinds of tasks on a computer.

So the one where there's an excessive response to negative emotion, you also see more brooding, but you see this hyperresponse to negative emotion expressed in reaction times as well. And then for the biotypes we see that are disrupted in the circuits involved in cognition, so either suppressing negative information or suppressing interfering information, it's like the executive of the brain circuit, those types map onto performance and reaction time on doing executive function type tasks like asking someone to stop when a red color is on and go when it's green, which is a fairly straightforward task, but it's one that will be more difficult to do with these types of depression.

Nicholas Weiler:

So is the idea with these biotypes that you could then get people more targeted treatment? Would you expect different treatments to be effective for people who, say, have more intrusive negative thoughts versus maybe a deficit in executive function?

Leanne Williams:

Absolutely. That is the goal and the focus. There's evidence from our studies and from other labs that show that using the circuit information, the biotypes can be very effective in either ruling out a treatment or ruling one in and at the same time considering more targeted treatments, as you highlighted.

Nicholas Weiler:

And just to contrast that, maybe you could speak for a moment about how we have been diagnosing and treating depression in the past and today, and I guess how that's working out for people.

Leanne Williams:

Right. The current way that we treat depression, which I imagine many people are touched by this, they've either gone through it themselves or loved ones, family members, it is effectively a trial and error process. Clearly the clinicians and doctors and therapists are so dedicated to being able to get good solutions. But in the absence of tools and the tests that I mentioned, the process is one in which we start with that broad category of depression, the umbrella diagnosis. That diagnosis in and of itself does not inform which treatment might be the most effective. So what happens is we start off with the same treatment options for everybody. We try one antidepressant perhaps, or one cognitive therapy, and then wait for usually eight to 12 weeks to see if that will work. And seeing if it works is reliant on the self-reported symptoms. I feel better or maybe I don't feel better.

So if it doesn't work, another is tried. If that doesn't work, another. Sometimes you start to have medications added, so you'll have people on multiple treatments, and for many, that process can go on for months if not years. So a lot of the people I see in our studies, they will have tried up to 10, maybe more, different treatments, and that's really hard. So the goal of having tests is to get the right one to the right person sooner.

The outcome of this trial and error process is that, and this is a consensus in our field, that only one-third of people will get better with the first couple they try. After that, it's essentially, unfortunately, diminishing returns. The odds don't really improve that much as you go along. So another one-third of people tend to form what we call treatment-resistant depression, and currently new treatments being developed are really targeting that treatment-resistant form of depression.

Nicholas Weiler:

But even that category of treatment-resistant depression doesn't really tell you why it's treatment-resistant.

Leanne Williams:

Right.

Nicholas Weiler:

Targeting treatments to that category, it may help some people but not others still. From your work so far on these biotypes, do you have any evidence about how effective it will be to get better treatment on the first try?

Leanne Williams:

I do have evidence. Very thrilled in the findings that have been coming out of the studies I've been doing. So this really is definitely a team effort so far. They're of two types. One is we randomize individuals to currently available treatments to mimic the trial and error process, and that will give us information about what is the base rate of getting better if it's just by chance, and that's usually about one third of people. Then we run models either where we go back into the data blinded to who was given what treatment, and we look at predictive models to see how much better we could do if we had have used the imaging. In a separate set of trials we do this prospectively. So we recruit people based on their imaging biotype, and we match them to a treatment.

In that first type of trial, we found that we can boost the accuracy of predicting someone can get better by at least 20%. On its own, that may not seem like a huge number, but that is over and above the chance level based on symptoms alone. So that's getting us to sensitivities and accuracies of around 80%, which are cross-validated. If I ask physicians and clinicians at the front lines, "What would make a difference to you using a new test?" They typically tell me, "If it could do better by 15%, I would use it because that's a big shift." That's getting something in the order of one person better for every five.

Nicholas Weiler:

So that would make a big difference.

Leanne Williams:

Yes.

Nicholas Weiler:

This is in the trials where people are treated normally, and then you look back and say, "If we had used the biotype information, here's how much better we think we could have done." So yeah. Tell me, in the prospective trials where you actually did use the biotype information, how did patients do?

Leanne Williams:

So this is really exciting to us. We've completed a trial recently where we were focusing in on this cognitive biotype, which is where there's a big impact on being able to function at work and so on, and that circuit implicates a particular type of targeted treatment that is used off-label right now for depression. It's called guanfacine. It's FDA-approved for hypertension, but it targets very specifically this part of the brain and boosts the plasticity in the part of the brain we're interested in. And in that trial, we were able to boost remission up to 63%. So remembering that it's normally around 33.

Nicholas Weiler:

That's doubling. That's amazing.

Leanne Williams:

Yeah. So that's what gives me optimism that we can double the rate. That was the target I set for our center. Many colleagues would, rightly so, say that's very optimistic, but I think it's a target we should aim for.

Nicholas Weiler:

That is incredibly impressive. And that was focusing on one of these biotypes, so obviously each biotype then needs to be matched with a set of treatments that are most effective for that biotype. It reminds me of a conversation we had with Kathleen Poston in neurology about Parkinson's disease. One of the things she said was, "You have to remember, it's not the symptoms that are the disease. It's the biology that's the disease and it's the biology that we need to treat." This seems like in some ways a very similar approach. What is the biology, what is the thing that's not working properly in the brain circuit, and how can we bump that back into the right zone? I think we talked with Nolan Williams about something very similar using transcranial magnetic stimulation to find the right circuit, bump it back into the right zone. So it's very exciting to be at this frontier in psychiatry where for so long we've sort of said, "Well, one in three people we may be able to help, and everyone else, we're sorry, but we don't have much."

Leanne Williams:

Right. I really like her quote about it is the biology we're treating, and once you treat the biology, you're also implementing new behaviors. If you can restore the brain function, you're able to have someone be able to feel pleasure or be able to function after a long time of not being able to. So it's that connection of your brain does connect to your behaviors and thoughts, and emotions, of course. Yeah.

Nicholas Weiler:

Well, I want to ask, we started off talking about precision mental health. I wonder whether you see precision mental health changing not only how we treat depression, but also how we think about it and how we talk about it, because of course, the stigma around mental health is so challenging. It's something that makes it very difficult for people to seek treatment, and so many people have experienced depression, and you yourself have been very open that your partner died from suicide after struggling from depression for some time. I just wonder whether, thinking of depression as a biological entity, do you see that as helping with this issue of stigma?

Leanne Williams:

Absolutely. I think it's key. It's been one of the most motivating results of the study to see the impact it can have on stigma and for individuals to literally change the narrative about how they can talk about their own depression. You see that transformative moment where it becomes very tangible that this is not their fault. They can get that sense of, it's not that I wasn't trying hard enough, or it wasn't that I'm not good enough. Once they understand that this is coming from a root cause that could be similar to if I had high blood pressure or I broke my leg. Of course, the way that they live their life will make an impact, but it's deeply transformative in that sense of this is about someone's character or their weakness or inability to cope, and it gives a different language to talk about it because depression is a very abstract concept. It doesn't explain what is it that's going on for me in the same way that Parkinson's disease, you can say, this is part of the brain that maybe is not getting enough dopamine. You can give a kind of tangible explanation.

Nicholas Weiler:

Is that something that you saw in your partner's case that it was difficult to separate it from himself?

Leanne Williams:

Absolutely. I think that was the, well, I know it was the key barrier because he expressed it. He was an emergency physician, and so in his mind he'd say, "I'm a doctor, I should be able to cope," was one aspect, and the other was knowing from his medical expertise, knowing that depression is a disorder that can be treated, there was still that concern that it will go into his medical record and then the stigma around how can I be an emergency physician if this is in my record. People will see me as not being able to cope in that role, or they won't have faith in me. It was really about that. Like I have to find a way to manage it without it being known.

Nicholas Weiler:

Well, I hope that this approach to really looking at the brain, looking at the biology behind the disease and getting better at treating that biology will help in allowing us to talk about it like we talk about Parkinson's disease or like we talk about other disorders where it doesn't feel like there's quite such a stigma.

Leanne Williams:

I'm confident that will happen. I mean, hopefully sooner rather than later. I think back to, I can remember as a child that it was stigmatized to talk about cancer. It was kind of the silent killer, and you didn't really talk about it. So I do think that if we get a new narrative around being able to say, "Oh, my amygdala is too active, and that's what's accounting for the experiences I'm having and I'm able to treat it with this approach." Targeting the amygdala is such a different language than saying, "Oh, I have depression."

Nicholas Weiler:

Well, I think that's all the time we have today. Thank you so much, Leanne, Dr. Williams, for coming on the show and for sharing your research with us and for sharing your personal story.

Leanne Williams:

Absolutely. I really enjoyed it, and thank you again for covering this topic and for inviting me on.

Nicholas Weiler:

Yeah. Well, we look forward to seeing what comes next. Thanks again so much to our guest, Leanne Williams. To read more about her work, check out the links in the show notes. If you're enjoying the show, please subscribe on Apple Podcasts, Spotify, or your podcast platform of choice and share the show with your friends. That helps us grow as a show and bring more listeners to the frontiers of neuroscience. We'd also love to hear from you. Tell us what you love or what you hate in a comment on your favorite podcast platform, or send us an email at neuronspodcast@stanford.edu. From Our Neurons to Yours is produced by Michael Osborne at 14th Street Studios, with production assistance from Morgan Honaker. I'm Nicholas Weiler. Until next time.