New Director, New Directions: An Interview with NIMH Director Tom Insel

Numbers tell the story. And the latest numbers from the National Institutes of Health fairly scream that the National Institute of Mental Health is the NIH home for psychology. In 2001, NIMH spent almost $360 million on social and behavioral science, and a PhD in psychology is the degree most represented among all NIMH grantees – behavioral or not. Psychological research is represented widely throughout the various divisions and branches of the Institute, including the divisions of neuroscience and basic behavioral science, services and intervention research, mental disorders, behavioral research and AIDS, and the Division of Intramural Research Programs.

With this as context, and against a backdrop of a doubling of the NIH budget over the last five years and prospects for a tightening NIH budget over the next five, APS Executive Director Alan Kraut sat down recently with still-new NIMH Director Tom Insel for a perspective on what his plans are for NIMH.

Although Insel has been NIMH director for only one year, he is not a newcomer to the Institute. Insel first joined NIMH in 1979 as a clinical associate in the Clinical Neuropharmacology Branch, and went on to hold several administrative and leadership posts. During his 15 years at NIMH before heading to Emory in 1994, Insel conducted research in obsessive-compulsive disorder, initiating some of the first treatment trials for OCD using serotonin reuptake inhibitors. Five years later, he launched a research program in social neuroscience, focusing on complex social behaviors in animals. Insel presented his research at the 2003 APS convention, where he also chaired an NIMH-sponsored symposium on social neuroscience.

NIMH in the Post-Doubling Era
KRAUT: Tom, thank you for agreeing to sit down with me. You’ve been in the job close to a year now. What are some of the things you’ve put in place, or are thinking about putting in place at the Institute. Are you taking the institute in any new directions?

INSEL: From my perspective, we are at a point of spectacular and rapid scientific growth and opportunity. For example, a philanthropic group in Seattle just announced publicly that it is going to put $100 million into mapping all the genes in the mouse brain. And they’re going to do it in two years. Lots of things are happening very quickly to move the science forward, using genomics, using neuroscience, using cognitive and behavioral science – all of those together in a way that says, “we can make progress.” That’s spectacularly exciting. But at the same time, we’re looking at a moderation of the growth in NIMH funding.

It’s important for your members to be clear that the change in funding can be seen in one of three ways – either the glass is half-full, it’s half-empty, or we’ve got too big of a glass. One scenario says we’re going to be dropping our payline, and that is undoubtedly going to be true. That’s the half-empty part. We’ve been trying to maintain about a 20th percentile for the payline, and that’s likely to fall significantly, possibly down to the 16th percentile. We are carefully considering ways to avoid this in a number of ways, including using the utmost care in making new investments.

The glass is also half-full because we have more money than we’ve ever had. That’s important to remember. There has been a tremendous increase in the amount of grant applications received, and NIMH is funding more grants than we’ve ever funded. In this next year, we will fund probably as many as five or even 10 percent more research project grants beyond where we were this past year.

KRAUT: This is a result of the five-year doubling of NIH funding that has just been completed?

INSEL: Yes, there’s a doubling of the overall NIH budget; though for NIMH, it’s about an 80 percent increase since 1998. And in these five years the number of applications has skyrocketed, too. Last year alone, there was a 40 percent increase over five years ago. CSR [the Center for Scientific Review, the gateway through which all NIH applications come] last week projected 66,000 NIH applications for the current year, about double where they were three years ago.

So even with more money, the increase in research costs and the increase in applications mean that if we get to the 16th percentile or better, it’ll be only because we’re being very frugal in how we’re paying out the grants we are able to fund.

We have to make more tough decisions. We’ve talked about everything from whether we should look at our continuations, to whether we need to make more drastic cuts in other grants we’re funding, to the possibility that we need to simply find ways to concentrate on certain areas of mental health and research needs and not spend across the board the same on every area, setting priorities for funding in areas where there is “traction,” where we know research will have an impact.

We clearly are not going to be able to continue doing what we’ve done, or we will become spread very thin and never get very deep. We need to have more depth in certain areas. This is an institute that has covered the waterfront. Relative to most NIH institutes, NIMH has a very broad portfolio, much broader than that of several other institutes with similar budgets.

On Budgetary ‘Whiplash’
KRAUT: Do you think the same kind of post-doubling conversation is going on at most of the institutes at NIH?

INSEL: Not to the same degree, although everybody is planning for a likely change in funding. It’s what we call preparing for a soft landing, or whiplash, depending on who you’re talking to. It depends on the institute. For instance, I’m aware of one institute that is still up to the 26th percentile on its payline. Others have had far larger increases

It’s going to be tough to produce a lot of new programs, unless we can find a way to cut back on some of the things we’re doing currently.

KRAUT: What are some of the options you’re looking at?

INSEL: There are lots of questions, but one is: what are we doing that could be done as well by other institutes? That question has been raised in particular with basic science, anything from basic molecular and cellular science to basic behavioral and social science. Are there parts of the portfolio that should be done by the National Science Foundation, by the Neurology Institute, or the National Institute of General Medical Sciences? There are parts of our portfolio that will not be covered by anybody else – and I want to make it clear that we know there are spectacular new opportunities that are obviously ripe for behavioral research. In addition, of course, the more applied work, the more disease-oriented research, will end up here at NIMH no matter what. We need to make sure we have the funding in those areas.

Another question is, even if the science is still very good, is it relevant? We’re concerned about what we call the ‘fossil’ regions of science, the areas where science has moved on, but the field hasn’t quite shifted – a kind of modern phrenology. Are we doing phrenology in a new form? We’re looking very carefully at where do we have traction now, where can we make the most progress, and where are we doing the same thing over and over again just because that’s the way we’ve always done it. That covers the waterfront from clinical trials to psychopathology to basic science portfolios.

This is not a black and white issue, but on the basic science end, often, the science itself appears quite elegant, and then the question is, is it going to have an impact because of its effect on mental disorders? That’s a very difficult call to make. Study sections aren’t asked to make that judgment; they’re scoring “best in breed.” It’s up to our council and to the institute itself to decide whether best in breed is also “best in show.” Having said that, I want to reiterate that we view this time as one of tremendous opportunity for behavioral science as it becomes increasingly clear that the expertise and knowledge it can bring to bear on understanding and treating mental disorders.

KRAUT: Back to your earlier point about wondering, because of scarce resources, whether NIMH can cover the waterfront in terms of basic science, one of the risks, it seems to me, is having even less connection between areas, say, basic cognitive science and schizophrenia. How do you protect against that?

INSEL: It is a risk. That’s where the judgment call comes in. You’ve got to pick the area where there will be discoveries that are relevant, and there are bridges to be built. Understanding basic aspects of cognition, social behavior, fear conditioning, those are areas where we will be able to build the bridges, and all three are areas where we have some traction. So those would be areas that will remain very important to the Institute.

Where the rubber hits the road is on individual proposals. That’s where it sometimes becomes a very difficult call. To fulfill our public health mission we need to relate our grant portfolio to an ultimate goal of understanding or treating mental disorders. You have somebody who may be an outstanding cognitive scientist, let’s use that example, but there’s no way to link her or his work to mental disorder, something that will actually impact the burden of disease. Yes, it’s important to cultivate researchers who are developing innovative behavioral and cognitive science, but at the end of the day we need to show that this work will address the burden of disease.

KRAUT: And, to the extent that you let those basic proposals go elsewhere, you’ve got the risk of well-meaning clinical researchers putting in place clinical interventions that could have been much better if they were grounded in basic cognitive science.

On Linking to Biology
INSEL: One place where the judgment call comes in, as I look at the portfolio so far, is basic behavioral, social, cognitive science. The part of these areas that includes biological underpinnings is where we are progressing at the fastest rate. When there’s a link to neuroscience, for example. You can already see that happening. Whether it’s [psychologist] Richie Davidson and his group at Wisconsin, or [psychologist] Dan Shachter and his group at Harvard doing imaging of memory, a number of people have already made that transition.

Does everyone need to make that transition? There is the danger of someone who becomes very invested in neuroimaging or genetics, or some other kind of more biological aspect of cognition or behavior, potentially becoming weaker in more general aspects of cognition and behavior; we don’t want to dilute our very best work by making that happen. On the other hand, you want to go where the action is, and right now, from my perspective, the action in basic behavioral and cognitive science is in behavioral and cognitive neuroscience. Certainly, neuroscience is giving us tremendous insights into behavior, as well as into the brain. And that is an important opportunity that we want to be able to grasp. You can see this in the study of face processing, an area of cognitive science that has been rejuvenated by fMRI, not only in terms of identifying brain circuits but in re-conceptualizing the encoding of faces.

The harder question is where do we get traction in basic behavioral or cognitive science that doesn’t have the biology? What do we want to develop or hang on to, and how should that be supported? A lot of it should be supported by us by becoming linked to our disorders, but we have to find a better way to do that.

KRAUT: There is such a sophisticated, almost mathematical elegance to much of modern cognitive science that’s not related to biology or neurobiology. You could argue that there’s a greater theoretical and methodological sophistication on the non-biological side than on the biological side. We started this conversation with a discussion of phrenology, and I remember, it might have been your predecessor who would call much of the early imaging “color phrenology,” or “high-tech” phrenology, because he had these wonderfully descriptive color pictures of the brain, but not much in a theoretical sense of what’s going on behind those images. Doesn’t that mean that we still need non-biological behavioral science to guide us?

INSEL: We get a lot more grants that talk about testing theories than about making discoveries. The Institute’s mission, where we make progress, is on the discovery end, more than on the testing theory end. You do need to have the theories, but we also need to be looking at mechanisms, and mechanisms often lead to absolutely new insights and new discoveries.

I think of our science as being at a point similar to the way it must have been in 1803 with the Lewis and Clark voyage of discovery. In the case of genomics, for example, we know a number of genes that are out there. But of the 30,000 or so genes that we’ve got, we’ve got expression information on, at most, two to three percent of them. It’s quite likely that these next three or four years will be like the Lewis and Clark expedition. You’ll find out where Seattle and San Francisco and all these great places are going to be, these really spectacular opportunities that are out there, remaining to be discovered. I’m not sure we’re even ready to make theories; we’re really in the discovery phase of the science. The theories of genomics and how genomics ties to neuroscience and behavior may be 10 years away.

It’s all going to happen very quickly. The tools are all in place. The information is largely in the public domain. We can go out and explore that territory, and anybody can stake a claim, right now in any part of it. The best estimates are that there’s something like 6,000 genes that are only expressed in the brain, not in any other part of the body. We don’t know very many of them at all, we’re just beginning to put numbers and names on these things. What they do is a complete mystery.

Against that you have to ask, where are we going to put our energy? I agree that pure cognitive science is generating some elegant mathematical models and theories. But I suspect that in the next few years, some of the most critical discoveries will not come from testing theories but from mapping uncharted frontiers.

On Translational Research
KRAUT: One of the buzzwords you hear around NIH a lot is “translation.” Even when, as you say, the basic science is at the top of its form, how do you translate it to a clinical area? In psychology, for example, you may know the basics of cognition or emotion or development or language, and the question is, since schizophrenia is at least partly about disrupted thinking and memory, about unregulated emotion, and about disordered language, how do you take basic psychological science and apply it to those areas? [articles from the Observer “Translational Research” series can be accessed online.]

INSEL: That is the Holy Grail. But 10 years ago, that was mostly a pipedream in mental disorders. In fact, I largely left the field of clinical research to do basic behavioral science because I felt we didn’t have the tools. We just weren’t there. Now, though, when people ask me what’s the priority, what do I really want to see happen here, it’s translation.

Sometimes that means translating the best clinical science to service, and making sure people are using evidence-based practices. But in a research environment, it means translating the best of basic science and the discoveries there, new treatments, new drugs, new uses of imaging, into the clinical realm.

The pendulum is now swinging back. One area where there’s real opportunity for translational research is extinction, really a subset of studies of learning and memory. We’re looking at the process by which learning decays, and how it is that someone who develops a fear memory can begin to overcome that fear. We have just spectacular animal models of that phenomenon. We’re beginning to understand it at the molecular, cellular, and systems levels, both in rodents and non-human primates. Now we have people who have taken some of those discoveries, and are moving into the clinical realm.

There’s some recent work by Michael Davis, a former colleague of mine at Emory, who has been very interested in fear conditioning. He has been able to look at this extinction process and has found a particular compound that facilitates extinction in a rat model. He now has begun applying what he has learned to human subjects with PTSD. It’s a remarkable example of how you go from really the most basic science, in this case emotional regulation, and ramp it up to a novel treatment for a disorder that has a really significant public health impact.

KRAUT: I should note that Mike Davis is a psychologist.

INSEL: He certainly is. Mike originally trained in the Wagner Lab at Yale.

KRAUT: We know that so much cognition is affected in a disorder like schizophrenia, but how do you attract help from someone who’s spent his/her career being very productive in an area – cognition, for example? What kind of incentives do you have that would convince those NIMH-supported National Academy of Science members who are basic scientists in cognition to start thinking in more applied ways about how to use their work, or their students’ work?

INSEL: The area of cognition and schizophrenia is ripe for picking. In fact we have a whole initiative in that area, bringing very good cognitive science and neuroscience together to focus on the cognitive deficits in schizophrenia. The field needs this, And it is where it is going. For example, we know the cognitive deficits in schizophrenia lead to disability, much more than hallucinations and delusions. But all of our medications are currently focused on hallucinations and delusions, so they’re not reducing the disability. For that to happen we’ve got to get people in who really understand cognition.

KRAUT: It also sounds like you’re describing another buzzword often heard at NIH, and that’s “multidisciplinary,” that there have to be multiple perspectives involved in going from cellular to behavioral. Of course, not everybody doing every experiment has to be knowledgeable in every discipline – it can’t work that way – but what do you do about promoting training that’s more multidisciplinary? How do you, for example, give the young geneticist more of a sense of what’s important in behavior, and give behavioral scientists in training the requisite knowledge of genetics?

INSEL: That happens on several levels. On the one hand you’ve got to cultivate a generation of people who understand that. I believe that “multidisciplinary” is a buzzword, and it’s been overused in a lot of ways. There will always be a role for people who are very focused and get very deep into a very narrow problem. But increasingly, science is going much more in terms of teams, because nobody can know all the techniques and all the implications of the techniques and know how to manage this from the molecular, cellular, systems, and whole animal levels. So excitement is coming from working across those domains, and that’s where we’re really seeing some of the most interesting insights.

For training in that realm, there are two possibilities. There are multidisciplinary training efforts underway, and there are many NIMH training awards that do that. You may know that NIMH has about double the amount of money in training, percentage-wise, of any other NIH Institute, except for NIGMS [National Institute of General Medical Sciences]. So training is certainly a very important part of our portfolio, and will continue to be.

But there is another piece of it. We usually think of training as being for someone either in graduate school or just out of graduate school. But increasingly, people who have followed a particular path, becoming molecular biologists, for instance, are finally asking the more molar questions – what does this mean, in terms of phenotype? – and then realizing that if to define a phenotype of a mouse or of a person, you’ve got to know a lot more about behavior and cognition. There’s been a lot of interest from that molecular community in getting re-trained, or at least getting exposed in a more adept way than most people have been in the past, to what it takes to reliably define behavioral or cognitive changes. There are workshops, summer programs, a variety of things that we and others have funded to make sure that happens.

Similarly, behavioral scientists increasingly realize they’ve got to learn about the most modern techniques in biology to understand the foundations of behavior. This includes genetic manipulations in model organisms, neuroimaging to trace functional pathways and sophisticated environmental manipulations. And, again, there are opportunities to do that. But what I tell young investigators all the time, that the tools are changing so rapidly and developing so quickly that what you see frequently is that people are having to re-tool at various stages of their careers. Training doesn’t stop in graduate school or even in your post-doc.

On Science-Based Clinical Training
KRAUT: You mentioned earlier that one of the gaps you need to address is the gap between what we know that works in a clinical sense and how to translate it to a service community.

INSEL: That’s a tough one. The Institute of Medicine says there’s something like 15 to 17 years between discovery and translation into action. That’s certainly true in the realm of mental disorders. You might think, innocently, if you came up with a new treatment for a mental disorder, or a physical disorder, that you would begin quickly to wipe it out. But in the case of mental disorders, when you come up with a new treatment, in a sense you increase the prevalence of the disorder rather than decreasing it, because everyone starts coming in for the treatment. It’s an odd aspect of this field and we’re still young enough, we haven’t gotten a good grip on any one of these disorders to really reduce its prevalence in a serious way with the treatments that we have. Too many treatments are only modestly effective.

The problem is true even with the good treatments that we have. This is probably most true with behavioral therapies, where the evidence is fairly compelling that cognitive behavior therapy [CBT] is either equal to or perhaps superior to medication, depending on the outcome measure. But if you ask how many people are qualified to use CBT for obsessive-compulsive disorder, or IPT [Interpersonal Therapy] for depression, it’s a only a relative handful of people at academic institutions. Medications are used for the vast majority of patients, by primary care docs who don’t have the time to really follow the patients, don’t spend the time to follow-up on whether the medication is being taken, maybe giving an inadequate amount of medication for an inadequate amount of time. We’re nowhere near where we need to be in terms of getting evidence-based treatments into the hands of the people who are treating most of the patients with mental disorders.

KRAUT: Do you see that as an NIMH role?

INSEL: We have to be a large part of it, along with our sister agency SAMHSA [Substance Abuse and Mental Health Services Administration]. We have the science, our role is establishing the evidence base, and SAMHSA’s role is the dissemination of evidence based treatments at the community level.

KRAUT: Do you worry that there is still training being given in therapies that have been shown not to be effective?

INSEL: Oh, absolutely. I joke that when I was in training in a very psychoanalytic department of psychiatry in the late 1970s, just 25 years ago, much of what I was trained to do would now be malpractice. I usually tell that to young trainees, to convey to them that much of what they’re being trained to do will also probably ultimately be considered out of touch. But what’s not so funny is that what I was trained to do is still too represented in training in the United States.

We need to make sure that training is based on the science. We’re talking about a tremendous public health problem. This is why I am looking forward to working with you and the academy [of Clinical Psychological Science] that you introduced me to. It seems to me that, in psychology, these Academy programs are the ones most likely to be in the business of science-based training.


APS regularly opens certain online articles for discussion on our website. Effective February 2021, you must be a logged-in APS member to post comments. By posting a comment, you agree to our Community Guidelines and the display of your profile information, including your name and affiliation. Any opinions, findings, conclusions, or recommendations present in article comments are those of the writers and do not necessarily reflect the views of APS or the article’s author. For more information, please see our Community Guidelines.

Please login with your APS account to comment.