- [Jeff] Welcome to "Healthy Minds."

I'm Dr. Jeff Borenstein.

Everyone is touched by psychiatric conditions, either themselves or a loved one.

Do not suffer in silence.

With help, there is hope.

(lulling music) Today on "Healthy Minds".

Just like in other conditions, typically medical conditions, there may be lab tests that along with taking a history and seeing the symptoms helps us make a diagnosis and decide on treatment.

Perhaps we can do the same for depression and other psychiatric conditions.

- One of our goals is to kind of rethink how we go about making those diagnoses, and hopefully come up with a better system, perhaps one more anchored in biology, that might be useful for guiding treatment decisions, and perhaps one day discovering new treatments.

- That's today on "Healthy Minds."

This program is brought to you in part by the American Psychiatric Association Foundation and the John & Polly Sparks Foundation.

(lulling music continues) Welcome to "Healthy Minds."

I'm Dr. Jeff Borenstein.

Depression, we all know someone who's experienced the depth of a clinical depression.

But depression can look different in different people with different sets of symptoms.

Today I speak with leading expert Dr. Conor Liston, who's working on developing subtypes of depression.

Conor, thank you for joining us today.

- Thanks for having me, glad to be here.

- I want to jump right in, and ask you to speak about the work that you're doing in looking at the subtypes of depression.

- This is a topic we've been interested in now for quite a while in my lab.

It's actually a question that had been kind of puzzling me for many years now, dating back to when I was a medical student.

You probably had a similar experience learning about the way that we diagnose psychiatric conditions, and depression in particular.

A patient meets criteria for a diagnosis of depression today when you have five or more of these nine symptoms, and that means doing some simple math that there are at least 256 unique combinations of symptoms that someone can present with and still get this diagnosis.

And that's not even accounting for the fact that some of these criteria contain opposites of themselves.

Like sleeping too much or sleeping too little.

Weight gain or weight loss.

And so, I remember kind of being puzzled by that, how it is that we came to this consensus, why we diagnose depression in this way.

And perhaps, might there better ways of doing it in the future.

And before we talk about some of the work that we're doing on discovering new subtypes of depression, I always hasten to add that the diagnostic system we have has done so much good for so many patients.

In a way, it's kind of miraculous that our treatments work as well as they do for these very different kinds of symptoms.

And so, the fact that we hope to do better shouldn't deter people from seeking treatments.

But I think it's not kind of controversial to assume that someone who's suffering from profound weight loss, sleeping four hours a night, intensely anxious, and kind of physically agitated, this person might not have the exact same biological problem as someone who's sleeping 20 hours a day, has gained 20 pounds, with intense carbohydrate craving, not at all anxious or agitated, just profound lack of enjoyment in things they used to enjoy, and can't kind of get out of bed or move much.

In many ways, these patients would look like opposites of one another if they were in your office.

So one of our goals is to kind of rethink how we go about making those diagnoses, and hopefully come up with a better system, perhaps one more anchored in biology that might be useful for guiding treatment decisions and perhaps one day discovering new treatments.

- In many ways, as we know, sort of the common thread is depressed, and the person being overwhelmed with that depression.

But then the other symptoms along with it, as you just said, could be very variable.

And we all always assumed that what we call depression is more than one illness, or subtype as you're describing it.

And that's what your work is really looking at.

- That's exactly right.

And you raise a great point, which is that we're not the first to try to tackle this problem of what we call heterogeneity in depression.

How patients present with different symptoms.

There's really pioneering work going back decades looking at subtypes of depression defined mainly on the kinds of symptoms that patients present with.

And we've learned a lot from that work.

Seasonal depression is one every winter that a lot of people can relate to.

It's seeped into the public lexicon.

Other kinds of depression, melancholic depression, atypical depression.

There's been groundbreaking work defining these subtypes and trying to identify distinct kind of biological substrates of these different types of depression.

But it's also been challenging to link any of those kind of symptom-based subtypes to biological markers.

Something you can measure in the blood, something you can detect on a brain scan, that might be useful for diagnosing the subtype or for guiding treatment decisions.

I think it's, we've learned a lot from that work, but again, with some rare exceptions notwithstanding, there really aren't biomarkers for diagnosing depression or guiding treatment decisions in widespread clinical practice today.

And that was sort of one of the motivations for what we're doing.

- Just like in other conditions, typically medical conditions, there may be lab tests that along with taking a history and seeing the symptoms helps us make a diagnosis and decide on treatment.

Perhaps we can do the same for depression and other psychiatric conditions.

- That's exactly right.

And that's what we've been trying to do instead of grouping people based on their symptoms and then looking for biomarkers that might be useful for diagnosing those subtypes, perhaps we can group people based on biology and then ask whether those kind of biological subtypes of depression are associated with different kinds of clinical symptoms and different treatment outcomes in a way that might help doctors and patients figure out the right antidepressant strategy more efficiently.

- One of the areas that you've looked at is using brain scans, tell us about what you found with that.

- That's exactly right, yeah.

So, we thought long and hard about what would be the best kind of biological measure to start with if we wanted to group people on this basis.

And there's no one right answer.

We envision as this field both in our team and other teams across the country and across the world as it progresses, additional measures might be useful.

But we started with a particular kind of brain scan called a functional magnetic resonance imaging brain scan.

These FMRI scans are similar to the kind of scan you might get if you went to an emergency room with a head injury or a concern of a stroke, but they've been sensitized to brain activity.

Basically, they detect the amount of oxygen in your blood in different regions of the brain, and that information can be used to infer how active that part of the brain is.

And it turns out, based on kind of seminal discoveries by Mark Rakle, Barat Bizwal, others decades ago now, the brain at rest exhibits fluctuations in this kind of MRI, what we call a bold signal, this MRI signal.

And regions of the brain that are strongly connected to one another tend to fluctuate together.

And that kind of makes sense.

If one region becomes active and it's connected to a second region, that region is going to tend to become active as well.

And so, you can use these scans to create maps of how different regions of the brain are connected to one another.

And I think of those maps kind of like analogy to our nation's airport network.

The airports in our network are arranged into hubs and kind of spokes, and we think our brains are organized in a kind of similar hub-like fashion.

And this hub organization explains why you can be sitting on a runway at LaGuardia on a beautiful Spring day and there's a huge backup waiting to take off.

And it has nothing to do with weather where you're located, it's due to bad weather maybe in Chicago O'Hare.

And because Chicago is also a hub, that disruption kind of percolates out into the network.

And we think that something similar might be happening in depression.

But importantly we think that the exact pattern of disrupted connections might be different in different people.

And that's what we've been trying to discover.

If we gather a lot of data, a lot of these brain scans from people all across the country and put them together, and then use machine-learning methods to just allow the data to tell us how people group together based on those connections, perhaps we can learn something about subtyping in that way.

- So you've come up with a couple of subtypes.

Describe clinically what those subtypes are like.

- The initial solution that we identified grouped people based on connectivity in the brain, that explained individual differences in two different symptom domains.

One being anhedonia, a loss of pleasure in things that a person used to enjoy.

And the second being anxiety, which is not a core feature of depression as you know, but often presents together with depression in many individuals, not all individuals, but many individuals.

And we can then group people based on those two dimensions.

And in our initial work, we identified four subtypes of depression based on a person's kind of position in that two-dimensional space.

One of them, for example, is associated with high levels of anhedonia and anxiety.

Another was associated with high levels of anhedonia but not really much anxiety at all.

And the other two subtypes were associated with different combinations of symptoms again in that 2D space.

And then in more recent work, we kind of went on to tweak these methods to make them more robust essentially, more useful we think for diagnosing subtypes in kind of community clinical practice.

That's obviously the goal that we're moving towards.

Work from Katie Dunlap, who was a postdoc in my lab and now has her own lab at the University of Toronto.

Katie showed that there's actually a third dimension that is associated with insomnia.

And you can group people in this kind of 3D space.

And the subtypes actually look pretty similar, but it's maybe a kind of more refined grouping.

And kind of importantly, Katie and others on my team have shown that this way of subtyping people can also be used to predict how they're going to respond to different kinds of antidepressant treatments.

- Basically, you could, if you know the symptoms, you can pretty much estimate what the brain is going to look like on the scans, and vice versa.

That really there's a relationship between the two, an association between the two, that helps in this differentiation of the subtypes.

- That's exactly right.

And the goal is that one day, you raise a really important point that I think a lot of people miss, which is that not only can you use the brain scans to kind of predict what kind of symptoms a person is going to experience, but if you have a person sitting in your office telling you about their symptoms, you can use that information to predict to some degree what the brain scan is going to look like.

So the vision in the long term is that if we can refine these kind of AI methods further, we might be able to do a pretty good job of diagnosing the subtypes in a majority of people without the brain scan.

But currently we need the brain scan to make the diagnosis accurate.

- So you've looked at various forms of treatment, and we know that people often respond to treatment ultimately, but not necessarily the first choice of treatment.

Which is one of the challenges that we have for depression.

But it's that case in other conditions as well.

For high blood pressure, the person may not respond to the first anti-hypertensive medicine that they take.

Or for seizure disorder, they may not respond to the first anti-seizure medicine they take.

This is really going to help guide us as psychiatrists to help make a choice that's more likely to be the right choice for that individual.

What are you finding, what are you seeing at least at this point in the research?

- That's exactly right.

We're really excited about that approach.

Can we use these brain scans and this kind of approach to subtyping to reach decisions more efficiently and more quickly about what the best antidepressant is for a given individual?

The antidepressant treatment that we focused on first is one called transcranial magnetic stimulation, or TMS.

TMS has been around for quite a while, decades now.

And it's been FDA approved for depression for quite a while as well, but only recently really has it become more widely available.

And I think many of your listeners may have heard about it at this point.

So it's a non-invasive brain stimulation treatment.

You could get it in an office just like the one I'm sitting in or the one you're sitting in right now.

The person would experience basically just kind of a scratching sensation on their scalp.

And that magnetic field activates the brain at the position that the stimulator is targeting.

And we've known for some time that if you map out the brain's connections using FMRI like I described, that connectivity at the target sites, which we call in this case the lateral prefrontal cortex, that connectivity at the target site is at least modestly correlated with your likelihood of responding to treatment.

And I think that kind of makes sense.

If you stimulate this brain region, the degree to which it's connected to other brain regions might be related to how well a person responds to the treatments.

And so we reasoned that these four subtypes that we had identified that have very differing patterns of connectivity might respond differently to this treatment that depends on brain connections.

And we found that they did.

We found that one of the subtypes was very responsive to TMS targeting a particular prefrontal brain target.

And two of the subtypes were not very responsive.

And the fourth subtype showed kind of an intermediate level of responsiveness.

And that was a great kind of first step towards proof of principle, that this approach could be useful for guiding treatment decisions.

In more recent work, what we've gone on to do is build new models, which which we think of as kind of FMRI biomarkers for predicting your likelihood of responding to TMS targeting two different brain regions.

And that's important because when a person with severe depression comes to your office, as you know, you're generally not thinking about whether I should treat this person or not treat the person.

You're generally thinking about which treatment should I give them, how do I choose between treatments A and B?

And so what we wanted to try to do was develop a biomarker that could help us choose between two different treatments.

Some of this work was inspired by pioneering work from Helen Mayberg and others.

I know Helen I think was on your show in the past.

Helen had shown that you could use FMRI to predict who would respond to a medication versus a psychotherapy.

And so kind of inspired by that, we wanted to try to develop an approach for picking the best TMS targets based on those brain scans.

And we're now testing that.

We have a kind of perspective-blinded randomized control trial where people are getting brain scans.

We subtype them, we formulate predictions about how well they'll respond to each treatment.

Everyone in the trial gets an active treatment.

That's important for anyone considering participation in the trial.

The goal isn't to test whether TMS works 'cause we know TMS works already.

The goal is to test how well our biomarkers perform.

And so that trial is underway, we're about halfway done.

And when that's finished, we'll really have a much better sense of not just how accurate the biomarkers are, but how useful they really are for improving treatment outcomes, which is of course what we care about as psychiatrists.

So we're excited about that too.

- As this work continues, if there's somebody right now watching, either they themselves or a loved one has depression, is there any guidance that you could give as of right now to them in terms of how to decide what might be the best treatment along with their psychiatrist?

- I get that question a lot.

So I think, you know, a really important part of making progress in this field is testing our biomarkers rigorously and making sure that they work before we distribute them widely to the community.

And that's why we have this trial underway.

We have a couple of other trials as well looking at biomarkers for guiding treatment decisions in OCD, and looking at a new accelerated form of TMS developed by a colleague and friend of mine at Stanford, Nolan Williams.

So I think what we really need is trials to test how well they work.

And I think in the not too distant future, it might be the case that if these biomarkers are proven to work and to actually help people, that patients might be able to access them through their psychiatrist.

In the meantime, I think exactly what you said is what patients need to do.

They need to work with their psychiatrists to figure out what is the most likely treatment to work for them.

It is to some degree, as you know, a trial and error approach, but we do have some clues about which antidepressants might work most effectively for which patients.

And that's something an individual can figure out in conjunction with their psychiatrist, usually based on the symptoms they're experiencing, their history of what antidepressants worked or didn't work for them in the past.

And sometimes also family history of which antidepressants were effective and which weren't for relatives can be useful.

- And that all of those are very helpful in terms of making that decision as to what is the best treatment.

- That's right.

- Conor, I want to shift gears a little bit, and speak to you about some of the work you're doing with ketamine.

And ketamine has been found to be a rapid-acting antidepressant medication.

A form of ketamine was approved by the FDA couple of years ago for refractory depression.

Depression that doesn't respond to other treatments.

And while it's rapid-acting, it also, the effect dissipates over a short period of time.

I'd like you to speak about the work that you are doing on this important issue.

- Great question.

We're really excited about this work as well.

You know, as you said, ketamine has been around for a while.

It took a lot of work from a lot of amazing investigators to take it from a research project to a drug that might have potential for treating depression, to one that's actually widespread and available, and used today.

And one of the things that makes ketamine really special is that it can elicit these really potent and rapid antidepressant responses often after a single treatment, which as you know is very different from almost any other antidepressant that we have in use today.

It doesn't always produce lasting effects for everyone.

In the early clinical trials, it was shown that for most people, if you take one dose of ketamine, many of those people, even those who didn't respond well to antidepressants in the past, will get rapidly better, but absent some other kind of intervention after that single dose, many of those people will get depressed again.

So the question is how do we keep them well.

- Right, which is one of the key questions about the use of ketamine, because we know that it can work right away, but then it dissipates.

So getting it to continue would be a tremendous step forward in the use of rapid-acting ketamine to treat depression.

- That's right, that's exactly right.

And you know, we and others are finding that simple things like repeated dosing might be useful for extending the durability of, kind of the survival of those new connections and that kind of comports with what we know as clinicians, that a lot of people might need more than one treatment to get better and stay better.

And perhaps there are other kind of fundamentally new ways that don't involve repeated dosing of ketamine to achieve the same goals.

And that's one of the things we're trying to chase down in the lab.

- I want to ask you, as we finish up our conversation, where do you see all of this going?

What do you envision in five, 10, 15 years from now in terms of how the work you and others are doing will impact the clinical care of people who have depression?

- Yeah, so I'll touch on kind of three themes there.

The first I think is efforts from my lab and from many others to kind of rethink how we go about diagnosing depression.

If we could identify those likely non-responders more quickly, we might be able to kind of triage them to other treatments that might be more likely to work for them in a more efficient way.

And hopefully save a lot of time and frustration for patients and their families, in kind of reaching the right treatment decisions more quickly.

So that's one theme.

I think another area that we're really excited about is figuring out how to deliver TMS more effectively for individual patients.

And then third, in the longer run, we hope that the studies, like the one that I described with ketamine, that they might lead to new insights about how antidepressants work.

And there are many labs doing amazing work with ketamine.

And we think that, what we're more focused on is not necessarily understanding the molecular biology of ketamine.

Many great teams are already working in that space.

But rather using ketamine kind of as a probe to initiate, to induce a mood state switch.

And trying to discover what are the kind of fundamental mechanisms that regulate those switches.

And if we knew more about that, we might be able to develop totally new treatments.

Or perhaps new treatments that could be combined with existing treatments to achieve much more potent responses with fewer side effects.

So that's a third area that we're really interested in.

- All of those bring great hope for where we're going in the future, and also there is a hope for people currently now in terms of current treatments and ways of diagnosing and figuring out what to do for people.

Conor, I want to thank you so much for joining us today.

And more importantly, this extraordinary work that you and your colleagues in your lab are doing that really does provide great hope for the future of treatment, thank you.

- Well, thanks so much for having me.

It was really great to be here.

And I hope we can do it again soon.

(relaxing music) - If you or a loved one are experiencing depression, don't suffer in silence.

There are many treatment options, including talk therapy, some of the more traditional medicines such as SSRIs, the newer medicines that are rapid-acting, and brain stimulation such as TMS.

Don't suffer in silence.

Seek help because with help, there is hope.

(relaxing music continues) Do not suffer in silence.

With help, there is hope.

This program is brought to you in part by the American Psychiatric Association Foundation and the John & Polly Sparks Foundation.

(relaxing music continues) (relaxing music continues) (relaxing music continues)