Webinar: Facts and Myths About RDoC

Transcript

[music]

[Uma Vaidyanathan] Today's webinar is on facts and myths about RDoC. My name is Uma, and I'm the scientific program manager for the RDoC unit here at NIMH. And in today's webinar, I'll be covering many frequently asked questions and concerns about RDoC. RDoC has been around since 2009, and there have been several articles published about and talks given on it by both folks from within NIMH and outside. However, there still do appear to be a few questions and misconceptions about it. And this webinar is sort of our active effort to address that. So with me today are Bruce Cuthbert and Sarah Morris to provide NIMH's perspective on RDoC, and we also have four distinguished panelists from various research institutions. We'll talk about their experience with RDoC as well. So I'll let everyone introduce themselves starting off with Bruce first, and then we can go around in a circle. So, Bruce?

[Bruce Cuthbert] Okay. Good afternoon, everyone. I'm Bruce Cuthbert. I've been the head of the RDoC unit since its inception last October. As of this November 1st, I will be taking over as the acting director of NIMH following Tom Insel's departure for Google/Alphabet. As a result, I'm pleased to introduce our new acting head of the RDoC unit, Dr. Sarah Morris. Dr. Morris has been with us since 2010 and has been very active in all aspects of the RDoC workshops, authoring many RDoC papers, and participating in all the activities of our working group, so she is very well-suited to take this on, and I appreciate Sarah taking over as acting head of the unit in addition to her regular program duties in our schizophrenia portfolio. So, Sarah, I turn it over to you.

[Sarah Morris] Thank you. I'm Sarah Morris. I just want to say thank you to the panelists for helping out with this effort. And to everyone who's dialing in, we welcome your questions and hope that you find it to be useful. Dr. Gotlib?

[Ian Gotlib] Hi. I'm Ian Gotlib in the Psychology Department at Standford. We have been studying depression for a long time and more recently we've been studying RDoC constructs, particularly in children. So, I'm excited to be part of this. Thanks for inviting me.

[Uma Vaidyanathan] Brandon?

[Brandon Gibb] Hi. I'm Brandon Gibb from Binghamton University at SUNY. We study depression and anxiety in kids, adolescents, and adults, and focus on cognition, genetics, and environment.

[Uma Vaidyanathan] Thank you. Joan?

[Joan Kaufman] I'm Joan Kaufman from the Center of Child and Family Traumatic Stress at Kennedy Krieger Institute also at Johns Hopkins, formerly from Yale as of 42 days ago. And we have two RDoC brands. One is looking at risk and resilience in maltreated children which combines measures of gene, environment, as well as imaging. And we also have a [K-series?] Bridge Project which is designed to create a tool that will ultimately bridge DSM-5 and RDoC diagnostic perspectives.

[Uma Vaidyanathan] Vilma?

[Vilma Gabbay] Vilma Gabbay from Mt. Sinai and Nathan Klein Institute in New York. And we have an RDoC project where we're looking at the reward processes in a [inaudible] sense across all psychiatric conditions and we've linked through-- links to the [inaudible] of the immune system both in the periphery and in the brain.

[Uma Vaidyanathan] All right. Thank you everyone. To our viewing audience, again, we apologize for the delay in the start. We welcome your questions as well during the webinar like Sarah said. If you move your mouse over the zoom window, you'll see a button that says questions and answers. And if you click on that it'll open a chat window where you can type in your question and answer to us and we'll do our best to address them during the webinar today. If not, we'll try to get a hold of you later after the webinar. All right. So let's go ahead and get started. So we'll start off with the elephant in the room question which is that RDoC has been criticized for being reductionistic and focused only on neural circuits at the expense of psychological phenomena. Sarah and Bruce, would you like to address that first?

[Bruce Cuthbert] Sarah, go ahead.

[Sarah Morris] Sure. So while it's true that RDoC does have its foundations in a modern understanding of brain-behavior relationships, we're not only interested in neuroscience or circuit-based measures, but all methodologies are units of analysis as we refer to them, including self-report questionnaires, data from EEG, genetics, behavior are equally important, and they inform and constrain each other in an integrative way. So in a sense, it's a pluralistic system in which different types of measuring and getting at constructs are meant to be integrated.

[Uma Vaidyanathan] What has been our panelists' experience with their grants? Do you feel similar or do you feel [inaudible] different?

[Ian Gotlib] I mean, speaking for myself, we're certainly studying neural circuitry but in addition. And I think it's a real strength of RDoC. We're are also studying a range of other, as Sarah mentioned, units of analysis where we're examining telemere length, we're examining cortisol secretion, we're looking at self-reported questionnaires. So I don't think we would have done nearly as well or gotten the information that we needed if we had restricted our study to neural circuits.

[Sarah Morris] Going to add that I don't think we can about neurocircuitry without thinking about psychology and experience as we're learning more and more about neuroplasticity. And in our studies we're finding dimensional measures of trauma, predict individual differences in connectivity within the brain, and that the effect of trauma can be moderated by social supports. So I think needing to embrace kind of the dynamic interaction of all these things, none of these systems work in isolation, and I think the RDoC's goal is to really think about how to integrate across units of analysis and recognize they're in dynamic interaction.

[Uma Vaidyanathan] Good point. Brandon, Vilma, anything specific to that?

[Brandon Gibb] 06:15.280 In our study, we're not looking at imaging at all, and I think what helped us get our initial grant was probably the really explicit focus on looking across units of analysis. So looking at EEG and eye tracking for example or looking at physiology and self-report. And I think the bigger key is looking across units of analysis and trying to integrate across those instead of any one being foundational or all-encompassing.

[Uma Vaidyanathan] Okay, good. Okay. So referring to what Joan kind of mentioned in her question. I think the next sort myth, if you will, about RDoC is that RDoC ignores developmental and environmental factors and that it focuses only on neurocircuits. And since Joan [already?] hadn't answered that, I'll kind of let her go with that first if that's okay.

[Joan Kaufman] Yeah, so I think as we're learning more and more about how the brain functions and develops, the brain is [not?] a static organ in that it's very dynamic, influenced by development as well as by experience. So incorporating development, experience into studies that are looking at kind of more basic neuroscience outcomes is absolutely key both from a translational perspective as well as in clinical studies.

[Ian Gotlib] Yeah, I'd want to echo that. In fact, our study is explicitly developmental longitudinal through puberty and it's an RDoC study. I just think it's not the case that RDoC is ignoring developmental issues.

[Uma Vaidyanathan] So Sarah, how would you answer the question then that there is nothing explicitly about environment development in the RDoC matrix.

[Sarah Morris] Well, part of it, I think that misconception comes from a practical fact which is that it's very hard to depict four-dimensional space in a slide or in a journal article, so we end up showing our matrix with its rows and columns. But it's hard to depict what we really imagine to be, in a way, a separate dimension within every cell in the matrix, which is the relationship of environment and of development, in a sense, in every single cell in the matrix. But RDoC is focused on measurements that are done at the individual level, and so, I guess, those factors of environment and development are measured in the way that they impact individuals and can be detected in individuals.

[Uma Vaidyanathan] Okay. Does anyone else have anything to contribute to that? Vilma or Brandon?

[Vilma Gabbay] Brandon? from our study, also, we focused on adolescents. Many of the psychiatric conditions start during this vulnerable period of time. There are multiple symptoms which we don't know what they mean. Are they going to become chronic, or is the kid is going to be just fine and normal with no psychiatric condition in adult? So looking, for example, at adolescents, would allow us to come up with preventive and early on biomarkers, so that's as developmental as can be in an RDoC study.

[Brandon Gibb] Yep.

[Uma Vaidyanathan] True. All right. So to reiterate then, it sounds like several of you do have grants and obviously focus on the developmental aspects of RDoC and environmental aspects as well. And would you say it's sort of been freeing to not have it specified in the matrix because it allows you to study it in the way that you'd like, right? If I'm correct based on your replies?

[Brandon Gibb] Yeah. I mean, I thought that if there was a-- even in the original writings, there was a huge section about development and environment. So it's always kind of surprised me that people either miss it or didn't see the emphasis from the beginning. So I think it's always been there, as it should for all of these disorders.

[Ian Gotlib] It's interesting to think about anything being freeing when you're writing a grant [laughter]. So it's tough to put these constructs, especially early on in RDoC. Bruce and I have had lots of conversations about what's an outcome. And it's really difficult to try to understand exactly what we're doing. But I'm just assuming that this is RDoC 1.0 and that we're going to be having changes to RDoC, giving findings from these studies on their own. But I think in the sense of going beyond the DSM diagnosis, that was just a nice thing to be able to do in the grant. Changed recruitment, for example, tremendously.

[Uma Vaidyanathan] So you also bring up an important point, Ian. In your grant, how did you define outcome variables? That's a question we get often. How do you define outcome variables in RDoC applications?

[Ian Gotlib] Yeah, I think it's a really difficult thing to do without relying-- I mean, you could say anxiety symptoms or regressive symptoms. We're actually much more interested in our work on the impact of early stress on neural trajectories of boys and girls through puberty. So we sidestep the issue of outcome explicitly until we start putting in follow-up applications for this. But I think it is really difficult. You want to say, "Does it lead to depression? Does it lead to anxiety?" But without being constraining by the diagnostic criteria.

[Uma Vaidyanathan] Okay. What about any of you, Brandon, Joan, or Vilma?

[Vilma Gabbay] What we are doing in our study, we're looking at the different level of activation of the immune system under stress, under psychological stress, under biological stress, and then our outcome is how our brain may respond to different level of rewards as well as clinical phenomenology of symptoms. So getting it out of the categorical diagnosis. And the truth is, especially in adolescence, the clinical categories of diagnosis are really a mess these days. I mean, many of them change, you don't really know what you treat. So our outcome are really labored on the neurocircuitry on chemicals in the brain and on the clinical symptoms.

[Uma Vaidyanathan] Interesting.

[Joan Kaufman] And, I think, for us it's, likewise, looking at kind of multiple neural measures as well as dimensional measures of a range of different clinical symptoms recognizing a comorbidity as the rule. So we're not just interested in one like depression, but we also know that kids with a history of abuse are an increased risk for substance use. So really looking across multiple types of outcomes dimensionally and ideally working to link it to various brain systems. So.

[Uma Vaidyanathan] Sarah, any prospectives in the program side on how people should define outcomes?

[Sarah Morris] I think we've heard some really great examples here. And it's just capturing that basic idea that we will get more traction by not whetting ourselves to complex heterogeneous diagnostic categories, whether we're talking about understanding existing illness, or predicting and preventing illness.

[Uma Vaidyanathan] Which leads us right nicely into our next question, which is that is RDoC intended to replace the current DSM-5 diagnostic system? Who wants to tackle that one?

[Bruce Cuthbert] So this is Bruce, and since, unfortunately, I have to run off to another meeting in a minute, I'll put in my two cents here. I'll pick up on what Ian said about this being RDoC 1.0, and that's exactly right. RDoC is a framework that is intended to grow and change dynamically and continually as new research data come out. And so the whole idea is to create a burgeoning literature that it can inform how we might create new revisions of the DSM and the ICD in the future, but there is no anticipation that we will release any kind of competing document in that sense. In fact, we'll never really release RDoC, which implies some cohesive, finished product. Rather, the result is to stay continual, expansion, and we hope, increasing sophistication of the research literature that could tell us in the future how we ought to make a diagnostic system that works in terms of being informed by neurodevelopment, brain circuits, behavior, and indeed people's lived experience of what it's like either to have these disorders or even to be in a risk state for these disorders, which, of course, can lead us to prevention.

[Uma Vaidyanathan] All right. I guess the next question comes directly from what you just said, Bruce. If RDoC is meant to evolve, people have criticized it by saying, "Well, you're putting a lot of these constructs and specified them in the matrix. So, essentially, you've just replaced one set of categories in the DSM in the new set of constructs. What if I want to study a new construct?" In fact, we've had questions coming in right now that it's like, "What if I want to study a construct that's not there in the matrix? So I can't use RDoC. How would I go about doing that?"

[Bruce Cuthbert] That's actually one of the real misperceptions. Again, it's the idea that while this is what's out there, so that's all you're allowed to do. In fact, our current constructs have been vetted by work groups as having a lot of promise and validity for study, but explicitly we also say these construct represent exemplars of how you might go about creating constructs that cut across the units of analysis and probably cut across many of our current diagnostic groupings. So we explicitly encourage the study of new constructs. The investigator simply needs to provide information in the grant application indicating the literature to date that would suggest that this would be a viable new construct and propose how they would go about studying it and instantiating it. And the only way that the RDoC matrix can, in fact, grow as intended is if, in fact, people do submit applications to study new constructs and propose them. So far from not having that available, that is very much explicitly encouraged. And it's always a good idea to talk to program staff here at NIMH about how to put together an application like that and to lay the scientific basis for proposing a new construct in an application. Sarah, do you have anything to add from a programmatic point of view?

[Sarah Morris] Yeah. I think maybe perhaps one source of the confusion has been the difference between our RDoC RFAs that we have published for-- to invite R01 applications. And in those RFAs, we did explicitly say that the proposal needed to involve one of the defined RDoC constructs. And the reason behind that was simply that we felt it was important to have an initial critical mass of applications looking at-- of projects that looked at those constructs that we thought were likely to be useful. But outside of those specific RFAs, investigators are welcome to submit investigator-initiated applications that propose other constructs.

[Bruce Cuthbert] I think it would be interesting to hear from our panelists about whether you feel that you introduced a new construct or sort of a new take on a current construct in your work, or whether you've thought about that. It would--

[Ian Gotlib] If I can do this quickly. I think a perfect-- I have a post-doc who just got an NRSA to look at irritability as a construct in this sample, and Ellen Leibenluft is a consultant on that project. But it's basically integrating irritability into the study of threat, acute response, how to respond to acute stress, to the prolonged stress. So I think that there are different kinds of mechanisms available to start to introduce constructs that aren't in RDoC to start with, to see what they do with the other constructs.

[Uma Vaidyanathan] Brandon, Joan, or Vilma, how about your grants? Have you tried using constructs that are not there in the matrix, or have you stuck with them mostly?

[Vilma Gabbay] We also looking at irritability which is another co-feature in adolescence which may induce multiple related symptoms, including threat, fear, and [affect reward?]. So it's part of our matrix even though it's not-- it's part of the grant but it is included in on our investigation. And we did also look at relationship to white matter compared to other symptom dimensions, and it does seem like it's associated with a distinct neurocircuitry, which makes sense, and it's probably different neurobiology.

[Joan Kaufman] And I think the complexity of these issues, we're not even at the beginning of understanding because we know irritability, when it's chronic, seems to have a different biology than when it's episodic. And irritability is totally nondiagnostic specific. You can have it with anxiety. You can have it with depression. You can have it with bipolar, so how we assess even these dimensions kind of over time because there's probably something very meaningful in these clinical dimensions from a longitudinal perspective, which I think we're all in the infancy. And how do we relate that to what we know and understand about neurocircuitry and risk and all those sorts of things? So--

[Brandon Gibb] Hold on. [inaudible] had something. We submitted our grant at the first call, the first RFA. And what we did that I think might've been helpful was instead of proposing a new construct, we said, "This construct, this thing that we're looking at seems to occur across a couple different constructs. And so we're going to try to compare and contrast between these constructs to get a little bit more clarity." And that's another thing that I think is probably interesting within RDoC is comparing across constructs or across systems.

[Ian Gotlib] And across units.

[Uma Vaidyanathan] Yeah. So it sounds like what's more important to derive from the matrix, or the ideas behind RDoC, rather than the constructs themselves. Which is that these are dimensional constructs at a measured range and that these constructs are based on both biology and behavioral. Would you say that's fair?

[Brandon Gibb] Yes.

[Joan Kaufman] And highly interrelated. So they may be under different columns, but the idea that they're totally distinct and separate, I think, is not what RDoC is trying to say.

[Uma Vaidyanathan] All right. A related question to this which has come up often is that RDoC does not define impairment. There's no way of knowing what is normal or abnormal on a construct. Whereas the DSM at least you have a cutoff for a certain disorder. Does anybody want to tackle that question?

[Vilma Gabbay] Yeah, I mean I can answer it, and I'm sure others would answer it too. I think that the binary inclusion of our DSM is really artificial because we really now don't know when impairment starts and when it ends. Are NOS syndrome are not significant clinically? Of course they are. So I think what, at least in our study, what we are trying to measure is when impairment starts. When reward processes, the reward circuitry, or fear gets to the point where it's impairment. The goal would be, I hope, that with technologies like with Google and Apple, we'll be able to come with specific items which make us abnormal and then we need to seek treatment. And I think that's where that's the goal. So I think that would be the future to do come with levels which are within normal range and without normal range.

[Joan Kaufman] And I think, just to add to that, I mean, I think that's part of the goal in creating this database. And I think, just like with hypertension where you have a subclinical range of functioning where you're not really sick, but it would say, "Aha. Some sort of change in lifestyle, change in behavior, or medicine," might be indicated so that you don't get into the pathological range. And I think ultimately the goal is to be able to collect this sort of data to understand where the cutoffs for absolute treatment are or where the cutoffs for prevention might be, and what things in recognizing-- it may not just be medicine. It may be change in lifestyle, change in environment, so.

[Brandon Gibb] The other piece of that is also explicitly testing for non-linear effects. So seeing is there a threshold effect at which point something becomes impairing, which I think RDoC is very well set up for.

[Ian Gotlib] It does raise the importance of those outcome measures, right? And impairment is one, just general impairment, in a GAF sense.

[Uma Vaidyanathan] Sarah, anything else to add to that?

[Sarah Morris] 24:16.629 Well, just to elaborate, I think Brandon's point about looking for tipping points or natural discontinuities is a very important one because if we conduct research with this false dichotomization of individuals into patient group and control group, we would [miss?] all that. So it's not that RDoC, that one can't do between group analysis in an RDoC project, but that this group should be defined on the basis of your empirical findings.

[Uma Vaidyanathan] All right. So now I think I'm going to transition a little bit more into questions related more specific to sort of grant applications. [inaudible] talk a little bit more broader conceptual [inaudible]. So this is one that we get often which is that in order to get NIMH funding, your proposal must include RDoC in it. Is that true, Sarah?

[Sarah Morris] That is not true. At this time, most of the clinical research grants that NIMH funds are actually not incorporating RDoC principles. Of course, we hope that that will change and that as researchers and reviewers become more familiar and persuaded by the logic behind RDoC that they will adopt some of those principles of dimensionality and multiple units of analysis. But at this point, there's no policy or change to the review criteria from NIMH's perspective that would have the effect of only funding RDoC grants.

[Uma Vaidyanathan] Okay. And I assume our panelists can vouch for that, if they have non-RDoC grants that are--

[Ian Gotlib] We have not RDoC grants.

[Uma Vaidyanathan] Okay, good to hear.

[Vilma Gabbay] We too.

[Uma Vaidyanathan] Related to that, the next sort of issue that's come up is that a lot of people ask us, "Do you have to have a neuroimaging measure to get an RDoC grant? Or how many units need to be included?" which is another version of it. Units of analysis. So Sarah, why don't you lead off on that and then we'll let the other people--

[Sarah Morris] Yeah, sure. So although RDoC constructs are based on brain circuits and an understanding of brain-behavior relationships, it's not necessary for every RDoC study to include neuroimaging measures. The general principles are that we get better validity in our measurement by using converging different methods and that we should be open to indicators of dysfunction that aren't just symptom-based. So for example, if we wait for a symptom to be available to self-report, we may have missed an opportunity for intervention earlier on. So again, that idea that multiple measures lend better validity and that other indicators of dysfunction other than self-report-- but they don't have to be brain measures, might be more predictive or objective.

[Uma Vaidyanathan] For our panelists, are most of your grants more biased toward brain measures, you think, or an equal mix of brain, self-report, and other measures?

[Brandon Gibb] I think I might be the only panelist that has no imaging in the grants at all. And so--

[Joan Kaufman] Well, on that myth, shattered that one.

[Brandon Gibb] I mean, I think that what we did do, though, is recognize that the constructs were chosen because they have an assumed underlying neural circuit. We based our argument around that and said, "We're going to work at each of the other units of analysis. And so we're still going to try to converge and work toward this kind of integrated view of the construct. But we weren't criticized at all in our grants for not having imaging. No one brought it up.

[Ian Gotlib] But, to be fair, you are doing EEG.

[Brandon Gibb] Yes.

[Ian Gotlib] Right. So, I mean, it is a measure of brain function. Right. Just a different measure. And so I think it is a reasonable question to ask whether you need some measure of brain function. I know that there are RDoC grants without any. They're probably not the rule, and Sarah would have a much better sense of this than we would from this side of it.

[Sarah Morris] I would guess it's on the integration of ideas, and ultimately, the research question but that it's not one measure that says, "This is fundable. This is not fundable." It's what's your pilot data? What's innovative that you're bringing? How can this move the field forward? So I think that's really where people need to be, kind of evaluating and looking at that.

[Uma Vaidyanathan] That's a very good way to phrase it. Yep. Related to this, we just got a question that asks, "Do you have any recommendations for looking at some of the more biological dimensions when you may not have the resources to analyze these? Is the best way to collaborate with those that do, or are there other ways?" Have any of you had this experience where you wanted to recruit more or less and you didn't have the resources to do so in terms of RDoC studies?

[Brandon Gibb] Collaborate. How we've done it is collaborate. The team's large, and everybody has their expertise. And so there's a genetics expert. There's a physiology expert. There's an ERP expert. I think that's the only real way to do this.

[Ian Gotlib] Yeah, I don't know that this was intended as a consequence of RDoC, but I think the multiple units of analysis on these grants has just created collaborations for most of these grants that probably wouldn't have been created had it not been that it needed to be an RDoC application.

[Vilma Gabbay] Yeah. I agree. I remember when I put the application I looked at the key personnel and I said, "Oh my God. They have so many key personnel." But that's the consequence of [MOT?] analysis RDoC application. And I think that's the only way to really move the field forward where the brain is so complicated, our body's so complicated and putting it all together, we have to get applications from cancer, from genetics, from imaging, from chemistry. And that's the future.

[Uma Vaidyanathan] Any other insights to add to that from anyone else? No? All right. We'll move on to our next question. One of our viewers asks, "All of this information pertains to what NIMH would like to see, but it is reviewers that decide whether something should be funded. How confident are you that reviewers take the same positions as your panel?"

[Sarah Morris] Can I answer that [laughter]?

[Uma Vaidyanathan] Definitely.

[Sarah Morris] It's true that reviewers play an important role in the process of getting grants funded, but they don't make funding decisions. But I can say that we are actively reaching out to reviewers who serve on panels at CSR to educate them about RDoC and NIMH's interest in projects that adopt RDoC principles so that they know, perhaps, how better to think about them when they come their way in standing review sections. So we get a kind of equal number of complaints that RDoC applications get bad scores at CSR and then versus the idea that you can't get funded if you don't do an RDoC grant. And so we're looking at that really carefully. And we have been doing some analyses about differential scoring. And we're not finding strong patterns of differential scoring and funding based on whether a grant invokes RDoC or not. So we are keeping an eye out for that and doing everything we can to make sure that reviewers are educated about NIMH's priorities and interests in these applications.

[Uma Vaidyanathan] All right. The next question we've gotten is, are RDoC grants for human clinical perspectives only? Can you speak to other RDoC dimensions, perspectives, and grants can be applied to non-human basic science perspective. And this, again, one of Sarah's. Not to shut our panelists out, but.

[Joan Kaufman] No, absolutely. I mean, there's a really nice paper written by Arie Kaffman and John Krystal, talking about taking the RDoC perspective and bringing it into animal models. We've been trying to model psychiatric disorders based on these complex heterogeneous things that don't exist in animals. And that taking the RDoC perspective and a brain circuitry perspective into the animal work is key. And it can be extremely synergistic. There's things you can do in animals that you can't do in kids. So if I see a methylation change, and I really want to understand what's the mechanism leading to that epigenetic change, I can't do the molecular studies in kids. But I can collaborate with someone who works in mice and can do that. And so, absolutely, RDoC translates perfectly. And it's Kaffman, K-A-F-F-M-A-N, Arie Kaffman. It's a great paper, thinking about how to bring RDoC perspectives into animal models of mental illness.

[Ian Gotlib] I would be even stronger. I was on the negative valence workgroup. And there were 40 of us, half of whom were animal researchers. And my perspective, which I know is wrong, is that almost all of the negative valence system con domain was created by the animal researchers, and I think it's perfect for them. I mean you just look at those constructs of sustained threat and acute fear. And it's just made for-- it's just made for animal researchers. So I think it's sometimes tougher for people doing human research to fit these constructs in a valid, creative way. Sorry, that was too strong.

[Uma Vaidyanathan] No, we like [crosstalk].

[Ian Gotlib] But I think it's important. I think that it's not can animal researchers play a role. I think that they play a tremendous role in creating the matrix.

[Joan Kaufman] And refining it.

[Ian Gotlib] And refining it.

[Uma Vaidyanathan] That too, yeah. All right. Another question we've gotten is can you comment on whether using an RDoC perspective would be useful in pursuing why the R21, R33 grants? I think Sarah, want to take that one again?

[Sarah Morris] Yeah, sure. So that's one of our new clinical trials RFAs so to determine the efficacy of novel treatments. And I believe that there is language in that RFA about RDoC, and we certainly encourage applicants to take to adopt RDoC principles as they're thinking about those treatments. And that would take the form of not conceptualizing the work in terms of developing a treatment for a heterogeneous categorical disorder but perhaps focusing on a more RDoC, either a defined RDoC construct or a similar disorder independent construct.

[Uma Vaidyanathan] A related question we've gotten to that, Sarah, and I think the others may help answer this one too is that when your grant funding budget is essentially limited. And RDoC requires that you recruit patients from every diagnostic group, how do you sort of go about doing this? Do you take all comers? How do you deal with the recruitment criteria?

[Sarah Morris] Right. So I can get started and then our panelists have lived experience in this. So just to be clear, the basic principle is that you would not recruit or frame your hypotheses on the basis a DSM category. Your recruitment would be guided by your hypotheses and the goal of ending up with a participant group that has the full dimension of the construct that you're interested in studying. And so that might mean that you target your recruitment to one sort of group of DSM disorders. Or you could achieve dimensionality by looking across disorders or by looking at subsyndromal symptoms and achieve that dimensionality from the clinical population into the general population. So you really have to be guided by the hypotheses. But the hypotheses, in essence, should not be about treating or testing the validity of a DSM disorder.

[Ian Gotlib] Yeah. If I may, Lea Williams, who's a colleague of mine here at Stanford, has an RDoC grant in which she is taking all comers to a treatment clinic, for example. So that's one way of getting a range of scores or functioning along whatever dimension you happen to be looking at. Because we're looking at stress in kids, our best friend is Craigslist. And we're taking all comers, but with ads eliciting kids who have experienced stress, I mean through their parents, of course, but trying to get a higher range of stressful experiences. But it's been, as we were saying earlier, freeing not to have to fill cells of MDD or bipolar to get the numbers just right.

[Brandon Gibb] Ours is also all comers, so we're taking all seven to eleven-year-olds. You can complete the protocol.

[Vilma Gabbay] Yeah. In our case, we're looking at toddler stage, which is three, four and five, puberty, but of course to catch a condition as long as they can participate in the study with the aim to get the full range of the specific symptoms. For example, full-range of anhedonia from zero to full-scale. Same with pleasure, same with hopelessness, so our outcome [inaudible] inclusionaries for the whole group to have a full range of the dimension that we're looking into.

[Joan Kaufman] We're looking at stress-related mechanisms associated with psychopathology and so our inclusion criteria are really looking at kids at the extreme, so kids who've had a recent out of home placement due to severe abuse and looking at ranges of psychopathology they experience and looking at comparison children who have not had that interfamilial type of trauma. So [that's?] one of the ways in which I think Sarah's point initially is the hypotheses should drive the inclusion criteria.

[Uma Vaidyanathan] But in general, it sounds like it's actually been easier for folks to recruit rather than the other way around, and there's concern that if you had to recruit too broad a spectrum, whether like you said, based on your hypothesis, actually been easier to take all-comers rather than restrict them by diagnoses, whatever.

[Vilma Gabbay] Yeah, 100%.

[Uma Vaidyanathan] Okay. Good. We have a couple of sort of repeat questions, but I figured since people are logging in late, I just want to go over them quickly if you want to address them again. Someone asked for the Kaffman citation again. Can you go over that quickly, Joan?

[Joan Kaufman] So it's Arie Kaffman, K-A-F-F-M-A-N, together with John Krystal, I think it's 2014. I don't have the-- I can't pull it up on my computer, but it's Arie Kaffman that will come up on PubMed, and it's specifically about RDoC perspectives in animal research. So a very well-written paper.

[Uma Vaidyanathan] Another question asks, "I understand that a brain measure is not necessarily required to get an RDoC grant but is a biomarker of some sort required however though?" I think we covered this earlier but can you just sort of go over it quickly?

[Sarah Morris] I guess my short answer would be that you would want to think very carefully about a biomarker for what. We have not had much luck identifying biomarkers for DSM disorders, and that's part of the motivation behind RDoC. And also, again, to think in terms of the principles of RDoC. It is possible to do an RDoC informed project using only self-report or behavioral measures. Again, incorporating those ideas about dimensionality and not simply looking at one extreme of a disorder.

[Uma Vaidyanathan] All right. And a related question we got Sarah, "Is there a place in RDoC for research and parenting or family?" This kind of goes back to development/environment. [crosstalk] later, I just want to make sure we cover them twice.

[Sarah Morris] Yeah, I'll give my short answer and then our panelists, again, have better detailed experience. There is certainly an appreciation of the importance of all kinds of different environment and developmental trajectories in RDoC. Just because they're not articulated in the matrix doesn't mean that we don't know that they're important. So we do encourage investigators to test hypotheses that involve those kind of factors and to measure them at the level of the individual.

[Joan Kaufman] Just to sort of reiterate, I think, without a doubt, we know that risk for psychopathology brain development is enormously influenced by parenting and very experienced, whether you're looking at animal studies, or you're looking at human studies, children born or raised in institutions. So without a doubt, these questions can be examined from an RDoC perspective, and it's an important-to-have experience as part of the measures accounting for variability and outcome.

[Ian Gotlib] And to make it concrete if it helps. Sarah [Ordaz?] in my lab has just got a K Award to add parenting to the RDoC project that we have already. So I think that-- I mean, Sarah is absolutely right and they're putting their-- they give you your money where your mouth is [laughter].

[Uma Vaidyanathan] All right. So we have about another five minutes left. So I'm going to start getting onto the end of our list of questions here. So one that we've gotten is, RDoC is essentially asking the field to wait another 10 years for neuroscience to inform new treatments. How do you deal with that? That's a long time to wait before you do anything with it.

[Vilma Gabbay] How long we've been waiting since Freud [laughter]?

[Uma Vaidyanathan] So we can wait a little bit longer? Yep.

[Vilma Gabbay] What is 10 years? I mean, for every treatment to come and when you look at clinical trials, it takes years, and years, and years. Here we have nothing. I think this is our chance to get into specific pathway and biomarkers that can really inform our new treatments. I think 10 years it's very brief period of time.

[Brandon Gibb] I don't know who's waiting. I don't see these as conflicting at all. I think that this compliments the new outcome trial initiative really well. The only difference between how it was is focus on a key mechanism, and show that you're moving that mechanism. But you're still treating impaired people, and you're still studying how to best do that. So I don't know of anybody waiting for 10 years to start doing treatment trials again.

[Ian Gotlib] It also seems to imply that we've learned nothing at all from DSM, and I don't think that's the case. I mean, I think we've learned a lot, and there are obviously problems with it, but we can take what works, what we know, and build on it, which I think is what RDoC is trying to do.

[Sarah Morris] Yeah. That's how I think of it, too. Not that everything we've learned about DSM disorders is wrong, but that if we think of those disorders, those diagnoses, as serving as a proxy variable for something, and we need to do a better job of figuring out what exactly do those diagnoses reflect, and how might they be improved and refined, and become more informative by testing outside of those boundaries.

[Ian Gotlib] Yeah.

[Brandon Gibb] Which is what the transdiagnostic treatments have been starting to do anyway.

[Sarah Morris] Yes.

[Vilma Gabbay] And we are using the same medication across psychiatric conditions, so be able to really specify what we're really treating, which we're doing, but without any real biomarker these days.

[Uma Vaidyanathan] Okay. And again, we sort of have a little bit of repeat question again, like I said, since we started late. One person asked, "How is RDoC being transmitted through NIMH program directors? Some we have talked with seem focused on behavior without knowledge of RDoC principles." Sarah, want to address that again?

[Sarah Morris] I'm sorry, did you say through NIMH directors or through--?

[Uma Vaidyanathan] Program directors is how the question's phrased.

[Sarah Morris] I still didn't hear it, sorry.

[Uma Vaidyanathan] How is RDoC being transmitted throughout NIMH program directors? Some have talked with seem focused on behavior without knowledge of RDoC principles.

[Sarah Morris] Well, I can only say that we certainly have talked about it a lot here on the program and at NIMH, and certainly having Bruce serve as division director and now as acting institute director, the message does get out. But I do think that there is more work to do in making sure that individual program officers have a good shared understanding about RDoC so that we can be consistent in the guidance that we're giving to applicants. But I hope a webinar like this will help clarify.

[Uma Vaidyanathan] I would think it would. You guys have all given very frank answers. So we're almost at the end of our time slot here. So before we end though, I would like to ask our panelists about this. You name one thing that following RDoC principles has been very useful. So one unexpected advantage you had in your studies and one, perhaps, disadvantage. It would be interesting to hear your perspectives. Do you want to start off on that, Ian?

[Ian Gotlib] Sure. It's a tough question. I mean, I'm trying to think of-- it's been a big change from the way that we had done work in our lab, to go from diagnosed clinical samples to a recruited sample of say, kids experiencing stress. And I think we're actually learning a lot more about the stress process with this broader group, with the broader perspective, than we have been learning from a much more circumscribed diagnosed sample. So I think it's, again, we're not learning as much about depression as a disorder, but we're learning a lot more about a much broader construct for us.

[Uma Vaidyanathan] Okay. All right.

[Joan Kaufman] For me, it--

[Uma Vaidyanathan] I'm sorry, Joan. You were saying?

[Joan Kaufman] Yeah. I think, for me, it means embracing the complexity and the ambiguity. And it's that kind of fine balance of you have specific hypotheses, but you also are open to surprises. And I think one thing that many people have written about in writing the RDoC is that people are agnostic in what kind of diagnoses are going to look like in 2020 or 2030, and I think designing your experiments to have some specific hypotheses but also be open to surprises because the relationships among these different units of analysis are really complex, and the things we're studying and interested in - improving the life trajectory of people with these illnesses - is complex, and so it's embracing that complexity.

[Brandon Gibb] For us, it's probably been two big things. One, it's encouraging us, even more, to include as many units of analysis as possible in every single study so that we can see where things either line up or break down, instead of just assuming. And then also like I was mentioning before, really testing whether the relations you're looking at are linear or nonlinear and really looking for these break or change points.

[Vilma Gabbay] I think, for us, when we started doing some more RDoC, we actually started within the categorical diagnosis of our depressed adolescents, and we started commuting these anhedonia scores, the irritability scores, both to core symptoms. But what we found that both of them has a full range, so you can be very severely depressed, but with no anhedonia, which we know that clinically because adolescents are highly reactive. And the same with no irritability, and that both of these have full range, as if some of them completely overlap without the control. It also allows us to examine which core symptoms may be associated with more severe outcome, either in a specific clinical group or a cross-psychiatric condition, with suicidality, with-- be more prodromal symptoms. And I just want to repeat what Joan said. Many of what we find, sometimes we do have hypotheses, but we have no clue if we're really going to have our hypotheses supported. One of my colleagues told me it sounds like a fishing exhibition. It's not [inaudible]. It is in the sense that we don't know what we're really going to find because it's so wide.

[Uma Vaidyanathan] All right. Okay, well, thank you for your answer. I think we're slightly past our time now, so thank you, everyone, again for a fantastic discussion today. Thank you to our viewing audience as well for logging in. If we weren't able to answer your question during the webinar, we'll do our best to reply to you via email in the coming days. Feel free to email as well at rdocadmin@mail.nimh.gov if any particular, burning questions you want answered. This webinar has been recorded, and we'll be posting it on YouTube in case you wish to revisit some of the discussion today. So thank you again for your time, everyone.

[Ian Gotlib] Thank you, Uma.

[Joan Kaufman] Thank you.

[Brandon Gibb] Thank you.

[Vilma Gabbay] Thank you.

[Uma Vaidyanathan] Bye, everyone. 

[music]