Skip to main content
Loading Explore guides…
Structure of Psychopathology · Acedra
Home
Explore
Economics
Structure of Psychopathology
Sign in to save
Chat
If YouTube asks you to sign in, open the video on YouTube.
Watch on YouTube
Reload player
Transcript
114 segments
Copy
Hide video
0:00
Okay, thank you all for coming. I'm Mary Miller and I'm a professor of economics and the interim director in the Institute for Social Sciences here, and I'm happy to welcome you to the first noon lecture of winter quarter. These lectures illuminate topical cutting-edge social science research that took place on our campus and we're trying to really give an opportunity for faculty, especially new faculty, which we have a ton of in social sciences this year, a chance to
0:24
present their work to an interdisciplinary audience. These are just one of many events that are hosted or supported by ISS, and we have a great website and Facebook, Twitter, YouTube, etc. Before I introduce our speaker today we have another noon lecture coming up later this month, on Monday February 26, assistant professor of economics Diana Moreira will present a talk on political turnover, bureaucratic turnover, and the
0:47
quality of public services in Brazil. So, you can write that one down, if that sounds interesting to you, and then presented for us this afternoon is associate professor of anthropology, Chris Hopwood. Psychology. Sorry, it says anthropology here. Dr. Hopwood earned his PhD in clinical psychology at Texas A&M and joined UC Davis from Michigan State this past Fall. He conducts research on personality assessment, personality disorders, and interpersonal processes
1:17
he's an associate editor for The Journal of Personality and Social Psychology, and The Journal of Personality Disorders, and a board member of the North American Society for the Study of Personality Disorders, and the Society for Personality Assessment. Please join me in welcoming Chris Hopwood. *Clapping* So, I'm happy to be here, and I'm thrilled to be at Davis, for all sorts of reasons. Not least of which is the weather, of course.
1:45
So, I understand the audience is a bit mixed, in terms of background, so I thought I would talk about something very big picture: the structure of psychopathology. Epidemiologists have tried to quantify the burden of mental health in a variety of different ways in terms of its impact in the criminal justice system, or education, healthcare system, lost work, etc., and by all those metrics it's a massive burden. And that's before you even get to the distress and misery that a person with
2:18
mental health problems suffers from. One of the sort of ironic pieces is that any epidemiologist trying to calculate the burden runs into the problem that it's very difficult to measure; and the reason it's very difficult to measure is because we don't understand, or have been using, evidence regarding the structure of psychopathology. So that's going to be my theme, the structure of psychology, to foreshadow there's a sort of revolution happening in the field, where we're going
2:46
from one understanding it's a way psychopathology variables are arranged to another. And, a lot of my work has related to that transition. I'll start with an analogy, but psychiatry and contemporary times is operating in a a way that's similar to the chemists operating before this picture right. We don't really understand how different psychiatric phenotypes are arranged with one another, because we don't have a good understanding of the
3:14
basic elements of human individual differences of neuroscience, etc. And, so, the way it typically works is people will present to clinics or hospitals with certain kinds of symptoms. And, if you've had a practice long enough, you start to notice patterns. And then you write down these patterns and, well, some people seem to be more this way and some people more that way, and you kind of bootstrap up from that sort of descriptive presentation and inference
3:39
to some kind of model of how people with different kinds of extensible mental disorders differ from one another. Ultimately, what we'd like to be able to do is to go from that kind of model to one in which you'll understand the basic elements. So, to go from water to hydrogen, oxygen, etc., right. And we're still quite a ways off. Luckily, we're in a position where we could take a major advance and that's kind of where psychiatry is at right now. It's in the middle of taking this to
4:04
a major step. The book in which psychopathology is categorized is the Diagnostic and Statistical Manual of Mental Disorders. This book has wide-ranging influence on law, on funding, on research, on practice. It's the book in which all of the mental disorders are listed, all the criteria for us that those disorders are listed, with thresholds, are given, etc. This book is also tied directly to the International classification of diseases, which is the
4:39
scheme that physicians throughout the world use to classify mental disorder, and it's used by all of the major mental health professions, not just psychiatry, but psychology, Social Work, counseling, etc. So, it's really the kind of nexus of psychopathology for a variety of researchers, practitioners, etcetera. It's also the way that the public largely understands psychopathology, so the assumptions of the DSM are also the way we tend to think about mental disorders.
5:06
And, the most important, probably, version of the DSM was the third edition which was published in 1980. 1980 and the 70s, which preceded it, were kind of an interesting time in psychiatry. The influence of psychoanalysis and psychiatric thinking was waning, and psychiatry was a bit under attack in the sphere of public opinion, because there was this view that mental health diagnosis and treatment were a way for the powerful to oppress the powerless.
5:38
And, so, there's a lot of concerns about medication, about diagnosis, about classification of a hospitalization. And, so, psychiatry was in a bit of a defensive position, both retreating within its ranks from psychoanalysis and, from outside, from people who are largely skeptic, skeptical about the endeavor in general. And so the idea of the third edition of the DSM, the first two were largely ignored, but the third edition ended up having a lot of importance in the
6:05
variety of ways I just described a moment ago. The idea of the third edition was: we don't really know what's going on underneath the hood, but the one thing we should be able to do is reliably agree, when we say something like depression, that we're saying the same thing. And, so the goal was to get away from inference, to get away from trying to come up with some explanation of why a person might be depressed and just focus on the readily observed presentation so that we
6:34
could all sort of agree, like these are the things that go along with depression or whatever, and distinguish depression from anxiety or whatever. That sort of model or way of understanding psychopathology really hasn't changed. There's been some minor changes around the edges, but for our purposes the same underlying assumptions have persisted for the thirty-seven years since the publication of DSM. Thirty-eight years.
6:55
I'm still getting used to 2018. And so that's going to be the foil in today's story, is the DSM or medical model of psychopathology. I want to start by illustrating a little bit more of an experience near example. So imagine a person who feels like this person looks. This could could be you, it could be someone that you care about, someone that you love. Psychopathology is sufficiently common in that it shouldn't be difficult for you to imagine at least one person that
7:26
feels like this person looks. And, ask yourself, is this person sad, miserable, hopeless, depressed? Is this person ruminating about how other people probably think about her or him? Afraid of going out into public, thinking they're probably lesser than other people? Does this person tend to see the world maybe through glasses that distort reality in some ways, maybe not see things the same way that other people see? Are they detached from meaningful goals? Most
7:59
likely many of the people that you sort of brought to mind would have met all of those criteria, as it were, right? That many people who feel the way this person looks can be described along all of those lines, and the basic problem is that from the DSM perspective those three sets of descriptors are the core features of three completely discrete categories of illness; in fact, not only discrete from one another, but in entirely different chapters at the
8:25
moment. Right. So, being sad and hopeless and depressed, they are the core features of depression, which is one of the disorders in the mood section of DSM. Being preventive and fearful of social engagement are core features of anxiety, generalized anxiety, being one of the disorders in the anxiety chapter, and having disconnections from reality and from meaningful goals are core features of schizophrenia, one of the disorders in this the current chapter. So it's this basic
8:52
disconnection between how we're asked to understand people, from the perspective of the DSM, and how people really seem to work in practice that's the root of the problem. So, I want to be a little bit more specific about that problem. The first symptom of the problem, as it were, is comorbidity which is what I just kind of described that is that there's an assumption that if you're in one of these boxes, it shouldn't tell you anything up about being in one of these
9:21
other boxes, right? And this is the kind of medical model assumption, if I do a diagnostic test that suggests you have tuberculosis, I shouldn't then think well then you probably also have breast cancer, for example, right? Because those things are discrete, they have separate etiologies, separate treatments, and so you would try to do a differential diagnosis to identify one as opposed to the other so you can engage in the appropriate treatment. In fact,
9:43
if I know, in practice, that you're anxious, I'm betting a pinkie that you're also depressed because it is rare, actually, for people to be severely anxious and not also depressed. If you get anxious enough, then you probably start distorting reality too. And, so, in practice, these things are extraordinarily difficult, indistinguishable one from the other. So that's the problem of so-called "comorbidity" that is people who have one disorder tend to not just have
10:05
one, they tend to have several. You can imagine how difficult it is to treat a person that has several disorders, if the idea is that every disorder ought to have its own treatment. The DSM somehow manages to also simultaneously have exactly the opposite problem as well, which is heterogeneity. That is, there's an assumption that if you're inside of one of these boxes you should be the same as any other person inside one of those boxes, right. So, again, tuberculosis should
10:32
be similar across people who have that diagnosis, which is why an effective treatment for that condition should work across different people who have two diagnoses. If we use the case of depression there's nine symptoms and the DSM, and you need five to be above the diagnostic threshold, that means two people could have four unshared symptoms and only one common symptom and yet achieve the same diagnosis. So, now again you're the clinician you got two people,
10:58
same diagnosis, but radically different presentations. What makes you think you should apply the same treatment to that person? A third problem has to do with the assumption that there are boxes in the first place, that is this idea that there's a difference between the well and the unwell. Some people are sick and some people are healthy. My guess is that it's not hard for many people in here to think of a time when they felt like this, right, because all of us, if
11:25
we're human, feel like that sometimes and really would be concerned about a person who never did, right. That is, some of us jump in and out of these boxes all the time. I do so daily, it seems like sometimes, right. So, the idea that some people should be classified as reliably unwell and some people should be classified as reliably well, sort of falls on its face. What this leads to is these arbitrary cut-offs, like five out of nine for depression, there isn't any
11:54
evidence, there's not a shred of evidence, that five is a meaningful cutoff for depression. And, in fact, the rule in the DSM was people should have roughly more than half, right. So five is more than half nine and that's why you need five depression symptoms to have a diagnosis. That shouldn't be very satisfying to us, right, from a scientific perspective, and you have to remember these epidemiologists that we talked about in the very beginning are going around
12:20
counting up the number of people in our population who are above these cut-offs and then saying, well and therefore this is the prevalence of these disorders and the cost of these disorders is whatever the lost work and health care burden is associated with these people. It's all based on this house of cards, which is the idea that there are sick and nonsick and that more than half of the symptoms are the reason to qualify people as sick. So, I already mentioned treatment
12:48
specificity is one of the problems that derives from the categorical model. You could also imagine that people who are interested in what's the underlying causes of psychopathology have a hell of a time trying to figure that out, when they're starting with a system that doesn't correspond very closely with nature. The good news is that there's a different way of thinking about it, and that different way of thinking about it, in part because the DSM is so well
13:14
established and has been a target for research for nearly four decades, that different approach has started to build up a stable body from which a better system could be drawn. So, I'd like to distinguish two symptoms of systems, the medical model, a system that is used to frame the current version of the DSM, and the model that I'm going to turn to next. The way that the DSM works is, it's published by the American Psychiatric Association and
13:45
the Board of Trustees of the Association pick an editor or editors to oversee the publication of the manual. Those editors then pick people that they regard as experts in certain classes of disease or disorder, they say you go get a committee together and tell us what we should do with that chapter, right, and so they then go pick the people they think that are sort of the experts in whatever area, the mood, or anxiety, or psychotic, or whatever section of the manual they've
14:11
been charged with reviewing and so all these committees sort of make proposals this is what we think should happen with this diagnosis, should it change in some way or not? It's fairly conservative because, as you can imagine, the American Psychiatric Association would be loath to put themselves in a position of making a change that they then had to go back on, right, so it is a sort of a very conservative overlay on top of the process. And, ultimately, these
14:41
experts sort of say this is what we think the criteria for these disorders ought to be and that ends up, after a lot of other steps, being published. So that that's a system that works well in the kind of medical condition where authority and hierarchy are valued a lot and there's some expedience to that kind of scheme right. People in charge sort of say this is how it works and that's what we're going to go with and that's the truth because I said so. From
15:10
a scientific perspective it's not particularly satisfying right. What we'd like to do is to use data to try to inform the way that symptoms present and how those symptoms covary with one another. And as I said there is a way of thinking about this question from an empirical perspective that is one of the things that happened in 20th century psychology is we've got a reasonable grip on measuring people and trying to figure out how some people are different from
15:36
one another, at least phenotypically. And so you can use quantitative measurement models symbolized by this level CFA diagram here as a way of understanding how the different symptoms people present with covary with one another in populations, right. And that then can identify a new way of thinking about psychology. So I just want to review to their big-picture ways in which this approach departs from the medical model approach, in terms of how we might conceptualize
16:05
psychopathology. The first way has to do with the nature of the variables themselves. I mentioned that there's no evidence, there are ways to test for whether something is binary, that is you have it or you don't have it, versus not being binary. The difference between X X chromosomes and X Y chromosomes is binary genetically even if phenotypically it's a little blurry right there's an example you can test in populations whether there really is a
16:31
difference between men and women, let's say, or between depressed and non depressed, right. When you do that regardless of which kind of psychopathology you're studying, you never find what's called a taxon, you never find categories or bins, you don't find that to say people with autistic diagnoses are qualitatively different from people without autistic diagnoses work, because schizophrenic diagnoses are different, qualitatively, from people
16:54
without schizophrenic diagnosis. There have been a couple of edge cases, interestingly when that happens they get kicked out of psychiatry. So one example is something called fragile X syndrome, which is a chromosomal syndrome where children are born with irregularly large heads, it's fairly severe, and it has to do with a problem on the X chromosome. Once we figured out exactly what the etiology of that was and figured out there is in fact a quality of difference between
17:22
between having this problem and not having this problem, it became a non psychiatric diagnosis, which is kind of interesting. Psychiatry is the home of diagnoses we don't understand. Anyway, when you look at the common mental illnesses, this is the shape of the distribution you have a large segment of the population who are more or less problem-free, or just have a few problems, and then as you get out into the edge, you have more and more people who have that problem, so depression for
17:52
example. There are some people who are so severely depressed, they're thinking about suicide all the time, they can't get out of bed. There are some other people who are so depressed that they go to psychotherapy, because it helps them get along, or take SSRIs because it helps them get along, but they still can parent or whatever and most of us probably live somewhere in this space where we get depressed here and there, but we can basically manage with our lives.
18:13
and and so it goes with other kinds of psychiatric symptoms. So, there's a couple of implications of thinking about psychopathology variables as continuous rather than categorical, one of them is that you lose a lot of statistical power by drawing an arbitrary line somewhere in this distribution. That is, Christian Markham did a meta-analysis of psychopathology studies and found that if you simply take this distribution, say people above a line have it and people
18:41
below a line don't have it, you lose about half of your reliability and about a third of your validity right. So, that alone would help us make predictions and agree with one another; we just stopped thinking about people as either having or not having a disorder. It doesn't solve the arbitrary threshold or cutoff problem, but it changes it. That is, it makes explicit that the cutoff has to be arbitrary, right. there is no principled scientific way to say some people merit
19:12
diagnosis and some people don't merit diagnosis is that if this is indeed the way for the shape of the distribution, right. In other words you could cut right there, let's say, or you could cut right there, and I can't tell you that you're wrong or that you're right by having done so. There are an important implications. Obviously, if you cut right there the prevalence goes down relative to cutting right there, because all these people either are or are not in your
19:38
diagnosed group. So, if you cut higher in the distribution that means that you're reserving resources for people who really need it. It means that you avoid diagnosing people who are relatively well-functioning and don't want a diagnosis in their chart, but it means that you're also missing people that probably could use a little help. So, it becomes in some way an economic or social justice kind of question rather than a scientific or
20:04
psychiatric question where you draw these lines, and then of course will affect prevalence and all sorts of other things. So, that's the first departure is that the variables themselves are not binary rather they're continuous. But, what happens when you put all these continuous dimensions together? Right now in the DSM you have all of these chapters of ostensibly different groupings like anxiety, and mood, and personality, and psychotic, and
20:32
developmental, et cetera, and that model can be tested using empirical techniques, that is you can take all the symptoms in the manual and you can submit them to psychometric evidence and try to see well how do they actually a sort in nature as opposed to how they assort in the minds of the experts who put together the manual. And there are different approaches to doing this, but they converge on a similar kind of idea, which is that there are evidence-based
21:02
dimensions which assort themselves into a kind of hierarchical framework. At the top of this hierarchy is one big chemical category dimension, which reflects the fact that people with problems tend to have lots of them. The more problems you have, the more problems you have. That is, if you're really severely anxious, you're probably gonna have lots of other things going on as well. And so what that looks like empirically is that every symptom of the
21:25
DSM is positively correlated with every other symptom in the DSM, even ones that don't seem like they ought to be, like say anxiety and alcohol use, where it's not obvious at first blush that those would be correlated maybe but in fact they are. And that's why, when you fit a model like this, you're gonna get one huge first factor, if you want to think of it that way. There's going to be a lot of variance and all these different
21:47
types that lumps into one big bin. You can then make a distinction between two broad differences, internalizing and externalizing we'll call them, that is some people tend, when under stress, to internalize, to feel anxious, to go inside of themselves, to feel worthless and helpless and hopeless, and some people tend to act out, to be violent, aggressive, hostile, substance using, etc. So, you can make a kind of broad distinction and then you can break those
22:14
down further, and then you can break those down further, and this picture would continue going down into an increasingly complex hierarchy where the syndromes that are listed in the DSM might live somewhere around this level and the symptoms that comprise those syndromes that maybe live somewhere below than that's here. Now, there's a couple of advantages to thinking about psychopathology it's arranged in this hierarchical kind of way. First of all you can connect the dimensions of this
22:42
system into underlying systems of psychometrics of personality, of neuroscience or whatever. And, in fact, when you look at the way neuroscientists tend to assort their sort of underlined systems of the brain, it looks a lot more like something like this and it's not like the chapters of the DSM. All right so there's converging lines of evidence that we're onto something This, of course, facilitates etiology research and in fact the NIH has issued
23:09
a document saying enough with the DSM diagnosis, we now want you to submit grants based on what they're calling the research domain criteria. These are these cross-cutting transdiagnostic systems of dysfunction that correspond loosely to what you find when you do a psychometric analysis like what I just showed you. This also has implications for the organization of the manual. Rather than having these chapters, which arbitrarily decide mood and anxiety for
23:35
example, are different from one another, you could assort the chapters along the lines of these major evidence-based dimensions. This would then have implications for assessment methods, so clinicians rather than giving assessment methods based on the organization of the manual, would give assessment methods that would be more efficient and reliable because they'd be based on an evidence-based model of how the elements of those assessments covary in nature. And this,
24:00
of course, would impact intervention research and practice. So, rather than treating a depression as distinct from generalizing zygote disorder and ignoring the fact that those two phenotypes correlate point-eight in nature, you could say well why don't we treat internalizing problems, because that's the thing that they have in common as an example. Maybe I should just touch upon that a little bit more. Intervention research has been standing still for 30 or 40
24:29
years so right SSRI's cognitive behavioral therapy, cyclothymic psychotherapy have been established for decades and they haven't really changed a whole lot. What we have learned from that research is: number one, the things that were designed to treat a specific problem tend to treat lots of problems, so SSRIs are used to treat depression but also anxiety and sleep problems and all kinds of other things, CBT works for all kinds of different problems, and that
24:56
there isn't any specificity, that is one treatment doesn't seem to work better than a different treatment that was also designed to work for anything, right. So, when you compare these treatments and RCTs, they virtually always tie, they virtually always be a weightless control, and they almost never beat one another right. And, so, despite the generation of countless acronyms, psychotherapies, and medications, we don't seem to be getting rid of them.
25:23
Perhaps it's because the targets are a little bit murky. That's my soapbox about that. Okay. So the health, and to summarize the first bit, it would help to move from an authority based system for diagnosing psychopathology to an evidence-based system for psychopathology in a variety of ways that may affect both research and treatment as I just described. But let's go back to this person and think about what his or her life is like. Simply saying that this person
25:56
seems distressed or gets distressed, this person seems anxious, this person seems to distort reality doesn't actually get us very far. If we really wanted to know what was going on we would have to say something like, well, when does he get stressed out? What kinds of things make him stressed out? When he get stressed out, like, what does he do? What does he do that's adaptive coping? What does he do that's not adaptive coping?
26:19
We'd have to know something functional is we have to know something about the dynamics of how is or her problems play out in his or her actual life. So, the first 10 years or so of my career focused on what I've just been talking about, getting the structure of psychopathology right. I'm basically satisfied that we've got the broad structured, that worked out, and it's becoming boring and so now I'm starting to be more interested in what I'm gonna call
26:46
dynamics. Just go back to this example to illustrate what I mean. Chemists, I gather, are not actually interested in this, right. What they're interested in is the fact that it's the relationship between protons neutrons and electrons that gives rise to this. It's the relationship between how many electrons you have that gives rise to the formation of molecules, right. That's what we care about. We want to know how the system works without
27:12
having a foundational structure we can't understand. But that's not the end. The end is figuring out how nature actually works and so it goes with psychopathology, we need to have evidence-based dimensions in order to make inferences about how psychopathology actually works, so clinicians for example can say well right now would be a good time to apply this specific intervention for the specific problem, in a way that so far we don't have much. So, how do you
27:41
go about studying dynamics? Well, the first step is organize your the dynamic inferences around evidence-based dimensions. So, we've been using a technique in my lab that I'll just give two brief examples of, and then I'll stop and open up for a discussion, to try to map out how evidence-based dimensions play out in people's world. So, this is the continuous assessment of interpersonal dynamics. It's designed to sample two dimensions
28:09
relevant to interpersonal functioning. The first is dominance versus submission, so how assertive, dominant, powerful am I being; versus how meek and passive and submissive am I being? The second is warmth versus coldness, that is how friendly, kind, agreeable am I being versus how distant, aloof, hostile am I being. The way this technique works is that a coder will watch a video of two people interacting with one another, and while watching the video will manipulate the
28:41
joystick device which samples their behavior every half second. So this coder might watch the woman in this scene and while watching that woman, as the woman gets warmer moves the joystick to the warm side and actually becomes more dominant moves it to the dominant side, etc., and is continuously moving the joystick while observing that person's behavior so that you get this time series, with the data point every half second along the line. So you end up
29:06
with four time series, the dominance and warmth from the first person and the dominance and warmth from the second person. So, she would code this interaction twice, once for each person, and then we'd have multiple coders code for the sake of reliability. And so then what we do is we try to use these time series to test hypotheses or make inferences about the dynamics of dominance warmth. So, we try to go beyond an inference like what he seems to be
29:28
more dominant than she is and she seems to be colder than he is, or whatever, but something about how do their dominance and warmth play out over time in the interaction within themselves and across themselves. So I just give two really brief examples, and then I'll stop talking. So, this doesn't show up all that well. I'll just have to fill in the blanks for you. What is narcissism? You could think of narcissism as a person who's kind of a
30:00
jerk. There's narcissism in the news, for example, particularly lately. In the last year or so. And so that that's kind of what you first think of when you think of narcissism, but even that person that has been in the news all the time is not narcissistic all the time. Even the most narcissistic person you can imagine has tender moments with his daughter, for example, when she was a child or to make it a little less sensational we probably it's possible that you have colleagues
30:30
in your own department who could be described with the term narcissism. But it's not the case that that person is always that way all the time. In fact, from a clinical perspective narcissism does not mean this is a person who's kind of arrogant, grandiose, jerk who thinks really highly of themselves. And, in fact, if that's all it meant, then I'd be concerned about universities because it seems to be full of them. It means more than that. It means that when you poke somebody
30:59
that they are so vulnerable and fragile, they have this immature reaction. That the grandiosity, the arrogance, the hostility comes out because they feel threat. Clinically, that's what narcissism is. So that can't be described by something like this person is kind of a hostile grandiose jerk. Some people are hostile grandiose jerks, but they're immune to being poked. They stay that way no matter what you do. So that would be the difference. So this is a scene in
31:28
which a husband and a wife are being recorded talking about some time where they had a difficulty. And they talked for 300 seconds and I'm sorry that you can't see the wife's time series very well. This is the husband's time series for warmth and this is his time series for dominance. And hers, I'll just kind of outline it because I can see it. It's pretty stable here and here goes like that right there and then goes like that and then sort of stays down here. So
32:05
what's happening in this scene? Well, for the first 30 seconds or so, they're kind of warming up. They're both. She's a little warmer than him, but they're both kind of warm. He's he's a little more in charge than she is which if you saw this scene you'd get the sense that that's kind of a typical way their relationship works. And, then, something happens here where she gets real dominant, just for a moment, for maybe like five seconds and then
32:31
retreats immediately. And while she's doing it she gets just a smidge cold right. So we don't we're gonna forget the content, because it gets too complicated but just think about the process. You're asked to do something you're going to talk about something. She asserts full force, right, and this guy with the assertion he's submissive for a moment, he gets extremely dominant, stays dominant for about a hundred seconds and gets cold and it stays cold. And then they kind of
33:03
reregulate over here, where he's still doing a lot of the talking but now he's a little warm. She's similar and warm. She's submissive but they but the rhythm seems to have been reestablished. So what happens there's 30 seconds of sort of warming up. She threatens him or pokes at him or insults him in some way, he has a narcissistic reaction that lasts about 100 seconds or so, and then they re-regulate one another. And again if you watch this scene you get
33:30
the sense like, this is not the first time that this has happened, and if you notice, you can't see it unfortunately, but if you if you could you would see despite him being having this kind of immature reaction, she stays pretty warm the whole time. And she backs off and she becomes submissive. And you get the sense, if you watch a scene like this, she's probably learned that that's the thing that causes him to re-regulate and they can reestablish some harmonious
33:55
rhythm towards the end of the interaction. So, from a clinical perspective, this is what narcissism looks like. It doesn't mean that the guy's a jerk because, in fact, in the end he's not much of a jerk. At the beginning he wasn't much of a jerk. Well, it's only after this poke that he acts out, and in fact there's a way that she's learned, it looks like, to re-regulate. So, I'm not going to go into much about how we use this to make other kinds of inferences, but this gives you a
34:21
taste of how you can use this technique to get beyond structure and towards a better understanding or deeper understanding of dynamics. I'll give you one more brief example, now in psychotherapy. These are what we call smoke pots in which an entire time series is summarized in one image. I just want to orient you to the image first. Each one of these is one person during an entire interaction. High is dominant, low is submissive. Right is warm, left is cold.
34:50
The density of the smoke pot is where that person spent most of their time during that interaction, and these are about these are these are therapy sessions, but they're brief they're about a half an hour, so most of the time for example this person was dominant and occasionally they were submissive. They were never in the places that are away. Right, so that makes sense? So this is a relatively famous training video in which this woman called Gloria is treated in the 1960s by
35:17
three famous psychotherapists. Is anybody familiar with these tapes? You're familiar with these? Yes, you know what I'm talking about. If you see them before, they're amazing, right? These are on YouTube, by the way, you can watch it. Perls always makes me laugh and Rogers always makes me cry, every time I watch them. So, the idea was that we want to show the world kind of how psychotherapy works from the perspective
35:36
of these three sort of major figures in the field. So, Gloria, this brave woman, goes and agrees to be recorded talking about her genuine problems with Albert Ellis, Fritz Perls, and Carl Rogers. Ellis is the so-called father of CBT, his sort of formulation is if you weren't so stupid, you'd feel better; which is to say, if you think about the world more accurately, then you'll have a more realistic appraisal of what's really going on. You
36:04
won't catastrophize and then you'll feel better all of a sudden. And so his therapy basically amounts to him saying that that's stupid, stop thinking that way think this other way instead and then things will be better, which looks like in our joystick technique a fairly dominant position right. He's doing a lot of like this is what you should do differently. And she more or less complies. She's pretty submissive and that that makes for a smooth interaction. She
36:24
occasionally talked, but not much. It's a fairly fixed dominant to submissive configuration. They're both leaning warm, her a little bit more than him, so it's pretty much the variances and that he's the dominant one and she's the submissive one. Perls' technique is more like, I'm going to screw with you to try to show you that you can tolerate being screwed with and make you a kind of tougher person, more or less. He called psychotherapy a safe emergency. So, he wants to create an
36:51
emergency in the patient so that she can experience the emergency and realize I got my way out of the emergency and I'm just fine. That's more or less the way he thought about therapy. So he tends to be a little colder than Ellis, still dominant but less so than he is. And he has the intended effect. She is all over the place relative to the other two pictures, but she gets pretty cold and dominant. She gets really angry with him. At one point she's beating on the couch and she
37:16
smokes like seven cigarettes during this session. So, his technique is working, for what it's worth. Rogers is like the grandfather you wish you could have had. He's like a sweet kind, genuine, authentic, caring, supportive person so that looks like warm submission in our little joystick technique, and that evokes warmth and some more dominance relative to say Helens right okay the only thing I really want to illustrate with these slides is if you think about
37:48
personality or psychopathology as kind of the way a person is, you would infer the way they are across situations. Now, this is the same woman in the same afternoon right. And, yet, these are three different people. One submissive, listening and meek; one angry, hostile and agitated; and one loving and warm and kind and assertive. Despite the fact that she's one person you've seen three different people because she's in different
38:17
situations. So this is another example of the kind of dynamic you can study using this technique again, just a taste. I did want to leave time for some questions and discussion, so I just want to close by summarizing. The understanding of how psychopathology is structured in nature is changing and, in fact, I would say one step further we have a fairly good idea, at least in a nomothetic population-based way, assuming linearity in all kinds of other things. We
38:43
have a fairly good idea of how psychopathology variables tend to go very open and we know that it's fairly different than the way that those variables are presented in the DSM. That change is happening slowly but surely. There's politics, etc., in the way. It's not in the interest, for example, of Pharma to dramatically change the way that the DSM looks, and so that will take a minute but but the change is inevitable and, when it comes, it will affect research policy and
39:12
practice. I've sort of accepted the change in my own heart and now I'm interested in the dynamics that can be studied with the benefit of more evidence-based structure. So I think I'll stop there and see if there's questions or complaints. *cough*
39:33
Or I can jump on down. *chuckles* Yes? I feel like I've read a few times that, this very broad generalization, that in the nineties everyone was depressed, taking prozac and now everybody's anxious taking xanax, and I wondered if you had encountered any of that in you're own research and if you had any thoughts about whether it's accurate or whatever would be the case. Well, I can start with the accurate piece. I don't believe that
40:00
the prevalence of psychopathology is changing all that much. This has gotten a lot of press with autism and narcissism as well. There's this idea that there's a narcissism epidemic and that there's an autism epidemic. I think a couple of things do happen that cause those kinds of trends. One of them is that we think about constructs differently than across time and, therefore, measure them differently. Two, given that the cut-offs are arbitrary, you can move them around
40:26
fairly easily without consequence, right. So all kinds of social and political factors come into play where, for example, schools may become more sensitive to the possibility that a kid may be having a hard time because of an autism spectrum disorder. At the same time, the way that autism is thought about and assess changes and becomes more sensitive and so all of a sudden a lot more kids get put in that bin. Whereas in 10 years ago they might have
40:56
been described some other way or not noticed at all, now all of a sudden there's a lot more of them and then people say oh my god there's an autism epidemic, like is it because of vaccines or whatever, right? And then now we've got kids not being vaccinated, et cetera, right. So, I think mostly that's the sort of reason that that happens and that there's probably not good reason to think there's a huge change in the actual prevalence of the problem. In other words,
41:21
another way of answering the question, maybe more simply, is I don't think this changes but I think our perception of where this line is changes all the time. And it's kind of arbitrary in the first place, and so nobody's right, nobody's wrong in a sense. So, from your evidence-based perspective of approaching psychopathology from multiple domains, when we're talking of psychology as a field and we have multiple scales that are trying to get out the construct of depression, yeah yeah,
41:51
how would you consolidate all of those varying scales that are measuring depression in adolescence and childhood to adulthood to be able to have this more systematic approach that we're all speaking the same language in a sense? I guess I think of three ways to respond to that. One is it's always better to use more than one measure, if you can, because latent variables are better than manifest variables from a data analytic point of view.
42:19
And you can go wrong easily, narcissism is a good example of an entire field kind of going wrong because they've ended up using a single measure that probably isn't that effective to measure a construct. I can talk about that more, if you want to. The second way to answer your question would be to say I think you need to measure more than just depression if you're interested in depression, because of this picture. In my field, for example, of personality disorders there was this idea that
42:46
borderline personality disorder is like this single discrete type. Turns out borderline personality disorder pretty good proxy for this and so you could think, well, borderline disorder is this disease that's comorbid with almost everything, or you can think, well, borderline personality disorder is like a summary of any kind of mental health problem a person could ever have. And, then, it becomes useful because it comes becomes a proxy for this global factor.
43:08
Now you only understand that if you measure that alongside of everything else. If you only, the people who still argue in favor of keeping borderline personality as a discrete type have tended to study it in isolation; that is have not measured all the other stuff that goes along with it. So, I think the second answer would be measure everything and then try to pull out whatever discriminate valid elements you might think are depression vis-à-vis
43:32
anxiety and other things, right. The third is the most challenging piece, because at the end there you brought up the developmental problem, which is the issue is that psychopathology probably manifests differently across the lifespan. So, the way a five-year-old expresses depression is going to be very different than a fifteen year old, and a 25 year old than an 85 year old. And, you know, the latent construct we presume doesn't actually change. That is,
44:00
depression, whatever that is, it's probably still the same, but because it manifests differently our measurement tools have to, our measurement tools can only connect to the way it expresses itself. And so there's going to be some disagreement across the lifespan in terms of how that expression occurs and I don't know of a good way of solving that problem, frankly. It's a big problem. I mean, child, adolescent, and adult psychologists have tended to not work
44:27
together, historically, that closely either so that might compound the problem. You know, I don't know that there's a great answer to that, I guess it'd be a punt, at the long-run. The solutions probably would be focused on the latent construct by using multiple measures and have people work together in developmentally contextually-sensitive way. That's probably a pathway to a solution but I don't know if the solution is going to look better. I'm intrigued by your ink blots. Oh, yeah? Okay. And
44:56
I wonder if those dimensions, in terms of more dominant-submission, if those have ever been used to look at the effectiveness of psychotherapy. Yeah, yeah. Yeah. I can answer that, Teri Tracey is a counseling psychologist at Arizona State, that's the one in Phoenix, and he's identi- So, the way he's approach this question is he's taken successful and unsuccessful cases, and he calculates a warmth and dominant score across sessions and sees if there's a pattern difference. And what he finds is
45:33
that there's three groups, basically. This is an over-simplification of his work. In essence, there's three groups. In the first group you have what he calls. I have to, excuse me, I have to define a construct first, which is complimentarity. With a construct called complementarity, which means that warmth tends to go with warmth, that is if I'm nice to you you're probably gonna be nice back, and vice versa, and dominance tends to go with submission, that is if I assert myself
46:00
that the simplest smoothest thing for you to do would be to be passive and then make it switch, you could be the dominant and I can be passive. So a complementary interaction is one in which one person is dominant, the other submissive, and both are either warm or cold, yeah, and the idea is that that you know it doesn't have to be that way but an interaction that's like that is going to be more stable and predictable. So, this is a good example of complementarity.
46:25
This is a good example of complementarity. This is not. And if you watch the sessions you're like wow that's intense, relative to these two right. Okay? Does that construct make sense? What Terry finds is that, in successful psychotherapies, there's this three-stage pattern were in the first stage there's complementarity, and the second stage there's not, and the third stage there is again. So, you can imagine Gloria who I think tends to be
46:51
kind of like this in real life. She tends to be a little warm and a little submissive. So, she might present saying you know I don't know doc I have this problem I can't quite figure it out, I really hope you're an expert, I really hope you can help. And a complementary response would be, yeah you know I hear you, it does sound like a real problem. I do think, because of my expertise, I can
47:12
be helpful. It's a warm, dominant response. If she came in saying I know exactly what my problem is, I know exactly what you need to do to help me, you just need to follow my instructions the complimentary responsive would be, sounds like you've got it all figured out tell me what I can do. Right? Okay, so the idea is in the first stage you have to match the person to create a certain smoothness and probably to create some attachment or some alliance with the person. In the second stage, if you
47:37
continue doing that thing that she is evoking, then you're just reinforcing the problem. That is, if her problems relate to always being too warm and submissive and then you take the warm dominant position relative to her, you're just reinforcing that approach to the problem of life. What you need to do is to do something else. Put her in an uncomfortable position so that she can change right so that's the second stage. In the third stage, hopefully, she's a little bit more
48:01
flexible and, so, to return to complementarity but a more flexible variety of complementarity. The two conditions of unsuccessful psychotherapy are number one the person drops out because the first stage is not complimentary. So, dominant patient meets with dominant therapist and is like this person won't listen to me, I'm out of here. Or, passive patient meets with passive autonomy- granting therapist is I don't think this person's an expert, I don't think that
48:28
she can help me. So that's the first kind of unsuccessful and the other kind, which as a supervisor of clinical students was by far the most common that I observed, was stable complementarity throughout the entire therapy and a lack of change. So you keep going because it's nice to go, talk to a nice person who's gonna listen to you and reinforce you. It's going to be smooth and I love going to see my therapist, but the person never actually gets challenged and, therefore,
48:53
never changes. So, I could say more about that, but I've already probably said too much. Well, I was thinking are you measuring the pair then, the therapist and the patient. I see what you're saying. This has been a big problem in psychotherapy research, I think we're a little coy about studying the therapist. We tend to attribute all the variation to the patients, and we don't pay a whole lot of attention to what's going on with therapist. As in like RCTs we say, well, therapist administers
49:25
administers an evidence-based treatment for these people who've been confined to these meds. It's been a mistake. Did you have more to say about that? Yeah. Well, just that clinically it seems like it's the relationship between that counts in terms of people being able to learn new programs, patterns, behaviors. That's consistent with what Teri has been finding. And, indeed, one of the effects I didn't mention before was that the most
49:59
robust predictor of therapy is the treatment alliance. So, if the relationship between the therapist and the patient seems to be the one thing that you can kind of hang your head on. If that's not going well, it's unlikely for the therapy to work. Am I supposed to stop at one? Yeah. There's one more. One more question. No pressure. As I'm doing just kind of fine upon the dynamics, how that would then have implications in diagnosis
50:26
So that is the idea that if you're going to classify or diagnose, you also have to observe in different settings in order to get these different... You would, I mean, the problem there is practical that people just simply don't have time to do a proper assessment, and I don't exactly what to do about that. It's one that pervades medicine. My own practice is in therapeutic assessment, which means that I focus on assessment, and so most of the treatment
50:54
and the assessments kind of blend together. And, so, I have the bene- I also have the benefit of not doing this for a living, and so I can do whatever I want, but in practice it's a big problem. You'd like to use multiple measures along lines of what the response was before. You'd like to assess dynamics, well these lines, and to assess the dynamics of both the therapist and the patient, along the lines of the gentleman's question.
51:13
I think the goal, the challenge is creating a model that's sufficiently efficient to do that. I had the fantasy, five or so years ago, that I could use machine learning to teach robots how to get these joystick ratings, and I've given up on machine learning altogether. Somebody smarter than me is going to figure that out. That's not a very good answer, but it's the best I got. *chuckles* Okay, I think maybe it's time to stop. Thank you. *clapping*
Here to help you learn
Sign in to keep a private conversation with your course tutor.
Ask me anything about the material…
Sign in to chat