How to Develop a Career in Translational Research - Interview with Dr. Greg Siegle
Dr. Greg Siegle, Ph.D., is Associate Professor in the Department of Psychiatry at University of Pittsburgh School of Medicine, and Senior Advisor to the Chair of our Neurocognitive Therapies/Translational Research SIG. I (Jonathan Stange, NT/TR SIG Student Representative) interviewed Dr. Siegle to ask his advice for graduate students who are interested in developing a career in translational research.
Your graduate school background was in cognitive aspects of depression. What led you to your interest in translational research?
I was actually interested in the translational work right from the beginning - specifically, applying the science of clinical psychology to intervention. When I got started, we were so far in the initial stages of specifying mechanisms that I wanted to do that for a while before applying these mechanisms to the translational/interventional research. I figured if I was going to intervene on mechanisms, we had to know what was wrong, and we were still debating about things when I first got started like, “do depressed people have attention biases?” So we did basic work for a while, and got a basic sense of the mechanisms. I got a sense that depressed people didn’t have these attention biases that were early; rather, they just elaborated on things. So then I went to do some physiological work with Eric Granholm in graduate school, and we saw that there was sustained physiological activity in depressed people. So we did some computer modeling, got a sense for what structures might be involved, and did some fMRI when I did my postdoc, so that we could say, “here are the structures that made that go.” At that point we saw that there was an amygdala that turned on, a prefrontal cortex that wouldn’t turn on to dampen down the amygdala, and then finally I knew what we could intervene on. And I went to do translational work to intervene on the prefrontal cortex. But it was a linear path. I really keep doing the same experiment over and over, iteratively refining mechanisms, to where we can intervene.
It seems that conducting translational research requires a solid background in understanding both mechanisms of and treatment for disorders. During your career development did you feel torn between these two directions when you were deciding which experiences to pursue? How should one balance clinical experience with research opportunities when positioning oneself for a career in translational research?
I was always interested in doing both, but this is a primary obstacle to one person doing translational research. That is, it’s too much for one person to do excellent basic research and excellent clinical trials and put them together, especially as a graduate student. That said, I think that just means that it’s an excellent argument for doing your work collaboratively. Getting people who are good at each part and being a member and a leader of a team - say, a group doing science, and a group doing trials, and I’m going to be integrating them. This is not an argument against specializing – I really do believe you need to specialize somewhere so that you have something unique to offer (there are a lot of good managers out there). Once you have something you are really good at, collaborating with people who know more than you do about the other pieces is essential. So when I went to do my first clinical trials, I had good clinical trials people on board – Michael Thase was my mentor in clinical trials from the time I got to Western Psych, and he really schooled me in how to do these things even though I didn’t have the ability to do them myself.
What are some opportunities that you considered/applied for in building your career? Are there current themes in translational research that you have noticed have become priorities of funding institutions?
In terms of opportunities, let me give you three areas that are useful to reach out to and be exposed to as a graduate student.
1) One area is modeling. As a translational researcher, if you model the systems you are interested in, you’ll be able to see whether your assumptions about what you might change actually lead to mechanistic change even in your conceptual framework, and often they don’t. Working with good modelers can give you an insight into a part that would be hard to get otherwise. By modeling, I mean computational modeling at the level of systems. I do a lot of neural network modeling, other people do semantic network modeling, some people do systems modeling, there are lots of ways of doing it, but the big idea is translating your assumptions into something a computer can interpret, and then saying, “if I could change it in some way in terms of mechanism, do I get changes in the system’s behavior that is reminiscent of recovery?” As a graduate student, it would mean finding people who do this. Certain kinds of modeling, such as neural networks, there are good books for out there.
2) Understanding brain circuits is a priority right now at, and the primary way we’re investigating brain circuits is through neuroimaging. So, if you happen to be able to collaborate with someone who’s doing neuroimaging or if you can get some training in it, it makes you really marketable.
3) As a trainee, getting onto a clinical trial with somebody who’s good at it is a good idea, because you’ll see day to day what it takes to make one of these things go.
What experiences would you recommend for graduate students who are interested in developing programs of translational research but whose faculty mentors’ research is primarily non-translational (e.g., risk mechanisms or pure treatment research without bridging the gap)?
I would say the big advice I have is to collaborate more broadly: have multiple advisors, work with people who are not your advisor. In this day in the age of translational, transdisciplinary, integrative researchers, the idea that somebody could be mentored in a career path by one person is almost antiquated, and a good mentor should understand that you may need input from other people to go beyond where your mentor is. So having multiple advisors, seeking external opportunities, is not weird. I have people visiting my lab all the time from around the country to get training in the stuff I do. Specifically, if you happen to be interested in psychophysiology or neuroimaging, the Society for Psychophysiological Research has scholarships for external mentoring for student members. So student members can identify a mentor anywhere in the country who happens to be an SPR member, and they can end up getting mentored there. Another really good mechanism is the F31, the NIH predoctoral research grant, which will give you some stipend money so you can do things like spend time outside of your mentor’s lab.
Graduate students may develop an idea for an ideal research study which involves evaluating mechanisms of a clinical disorder, but may find it difficult or infeasible to recruit and gather adequate data from the desired population. Thus, students often resort to studying mechanisms in non-clinical analog samples (e.g., studying symptoms rather than disorder with a convenient student sample). To what extent would you advise graduate students to collect well-powered studies using less-desirable samples, versus conducting less-sophisticated and potentially underpowered studies with the desired clinical sample? Is it better to wait until post-doc or internship (when one may have better access to a desired clinical sample) to conduct a more ideal study with the desired sample?
I’ll give you two contradictory answers. There is some cache to being able to say you have done work with diagnosable populations. If you are going to be applying for a K award in your first year out of graduate school, having a publication with a diagnosed population is good.
So let me shift to the other answer. Right now at NIMH, the Research Domain Criteria (RDoC) is huge. The RDoC basically say that clinical disorders are terribly heterogeneous, and there is no reason we should have really ever gotten so far down the path that we reified them as entities. Instead, we should be breaking that apart, and saying that if what we want to do as translational researchers is intervene on mechanisms, we should be recruiting people who have the mechanism we’re interested in. So this is different than analog populations – it’s different than saying I’m going to recruit subclinically-depressed people because I believe in “depression-lite.” It’s saying, I’m interested in reward-processing deficits, for example, so I will recruit people with anhedonia, even if that is not depression, or schizophrenia, or some diagnosed category. So I do see that as a way out if what you’re interested in is a mechanism as a translational researcher, go right for the mechanism, but hit the mechanism hard. And I’m not saying not to go for sample of convenience – often mechanistic samples are terribly inconvenient to get. When I wanted to put in a grant looking at reward processing, just taking people with low reward processing wasn’t enough – the RDoC wants to see you go the whole range, so I had to find who is high in reward processing – for example, gamblers, and sexual risk-takers, which are not really my area, but if I wanted to pursue research in this area then I had to find collaborators to do that. So we did put in a grant looking at people everywhere from anhedonia at the low end to sexual risk at the high end.
It seems that clinical psychology research is increasingly utilizing physiological techniques (e.g., attempting to identify “biomarkers,” evaluating respiratory sinus arrhythmia, skin conductance, using EEG, fMRI, or other approaches based in cognitive neuroscience). What experiences would you suggest for graduate students without much experience with these newer (and often more expensive) methodologies?
My advice is to learn what you can. Collaborate, collaborate, collaborate, even if it’s in a basic cognitive lab that’s outside of your area of expertise, if you feel like you need this technique, go help somebody else out who is using it and apprentice so you can get this. I apprenticed in a schizophrenia lab – I was a depression person – but that’s where they were doing physiological research that I wanted to get experience with in graduate school so that’s where I went to get it. There is a historical bias against peripheral psychophysiology. Don’t have it – it’s coming back into vogue.
The last point I’d like to make here is that it really doesn’t have to be expensive. This technology is getting cheaper and cheaper and look for affordable things coming around the corner. I’ll give you two examples. One is EEG. The EEG system in my lab is $80,000. It’s a professional system and it was industry standard a few years ago. Recently, I’ve been exploring a $500 EEG system from gamers which is giving us beautiful signal for the research we want to do. There is another technology if you want to do peripheral psychophysiology called Phidgets, which is $75 for a little amplifier board where you can put any transducer in it (whereas the amplifiers for physiological signals in my lab are $20,000). So these things are coming around the corner. One thing I would suggest is as stuff comes around the corner, it’s not always the most user-friendly. While you’re a graduate student and you are young and still have energy, don’t be dissuaded by that. Learn to program, learn to script, learn to use whatever technology is available on its terms so you don’t have to wait for somebody to package it in some expensive way.
Would you advise those at the early stage of their career to gain in-depth training in one methodological approach within translational research (e.g., fMRI, cognitive enhancement strategies, psychophysiological indicators) rather than venture into broader training in applying multiple methods? In other words, is it important to be focused in your methods at the early stage of your career?
If you think of a continuum between very low level and big picture, a good scientist always has the big picture in mind. But in my experience, good scientists are also good at something. If you have nothing to contribute yourself, then you’re just a manager. There are a lot of good managers out there. So I would say, have a big picture vision, but get good at something, so that when you get into your own lab, you can set up there and do something without somebody there to help you. The other thing I would say is have a path for the things you need to be good at – e.g., in five years I will have done these things – and hold yourself to that so that you keep getting training. And there are training mechanisms. So at the level of the doctoral student there’s the F31 mechanism. At the level of a post-doc there’s the T32s (which are institutional) and F32s (which are your own grants). And at the level of junior faculty there are K awards, which are mentored training awards, and they are five years to get good at something. So I’ve had my list of the things I want to know in 10 years and I’ve just been working through them.
Do you have any recommendations about ways to get translational neuroscience training outside of one’s graduate program? For instance, if one’s program doesn't have specific resources for learning about that area, are there summer institutes or other training programs that are good for young investigators?
If you want physiology specifically – things like heart rate, pupil dilation, eye tracking, EEG, any of that stuff – become an SPR member, go to their conferences, get their student training awards. As the NT/TR SIG is an ABCT institution, tell ABCT we need these institutes at our conference. We always have preconference workshops in things like statistical modeling. Ask them to have preconference workshops in things like psychophysiology. Chris Deveney and I have offered, and they’re not sure that there is really a call for it. So, make the leadership aware that there is a call for it and they’ll do that. As faculty advisor to the SIG, if you’re interested in an area and you can’t find a faculty advisor in an area who is doing it, talk to me – I’m happy to try to point you to somebody. The best way, though, if you’re interested in an area, is to Google it, see who’s doing it, and see if there are training institutes for it. For fMRI, there are lots of training institutes all around (e.g., there are institutes in Boston, Woods Hole), all over the country. For computational modeling there are also institutes.
What is the best way to obtain a post-doctoral fellowship that will help you gain experience with translational research?
Three answers here.
1) Write to people in your area who are doing your ideal work, do not be shy, and say “I love what you’re doing, can I work with you?” I think that’s the #1 best way to get a post-doc, is to find someone who is doing what you want and contact them.
2) Look for translational T32s (institutional training grants), and see if you can affiliate with them. We have one here at Western Psych, and there are other postdoctoral training grants around the country that support people.
3) Finally, having your own grant really helps to get you in the door as a postdoc. A dissertation grant helps because it shows that you can do this work; having more publications is always better; either of those is terrific. And having a grant that pays for a postdoc is writing your own ticket, because a lot of people will take you into their lab if they know that all they have to do is provide a desk, rather than a desk and a stipend. How easy it is to get these grants depends what you are doing. There are niche areas where they come easier than others. The Templeton foundation might give grants if you’re doing positive affect. Writing an F32 is always a good way to go – you identify a prospective mentor and write an F32 with them so that you can go to their lab and work with them.
I would also add that if you are an undergrad, definitely don’t be afraid to take a few years as an RA between undergrad and graduate school. If you are in grad school and get an opportunity after your degree to be a study coordinator for a good study that is in a translational area where you could write a postdoc grant, that’s one option you could consider just to get experience in some area or to be writing F32 in area with someone. While you are on internship you could be writing an F32. Also, try to do your internship somewhere where there is someone with whom you can do your postdoc. At Western Psych especially we take people with often the expectation that we will try to do what we can to let them stay as a post-doc, so it’s kind of a combined internship/post-doc for some people.
How does the job market look for researchers in translational research, what types of positions are available (e.g., for recent PhDs) and what can graduate students do to make themselves appealing candidates for these positions?
If what you mean by translational research is combining, for example, neuroimaging with clinical trials, that’s very hot, people want that, there are a lot of clinical departments retooling to do neuroimaging. If you can get a neuroimaging or EEG background as a graduate student, I would do it. If you can cultivate strong math and technical skills as a graduate student, I would do it, just because that is in demand. There are a lot of people who do good clinical work and really want to add the translational piece and you’ll be in position to do that. On the other hand, there are a lot of people who are specializing more in clinical trials who are having trouble getting out because NIH isn’t funding big trials especially at the level of post-docs, so that’s sometimes harder.
What suggestions do you have for networking and “getting your name out there” among translational researchers?
Three things to say here.
1) Be brave, introduce yourself, find someone to introduce yourself. Use LinkedIn, use Facebook, use your advisor’s network to get introductions.
2) Present at conferences. My advisor, Rick Ingram, used to say “surround yourself with greatness.” As a graduate student, have a symposium with the people you want to work with, who you idolize, invite them to be part of a symposium. They’ll get to know your name, they’ll get to know you as an authority figure basically, and they’ll be looking up to you while you look up to them if you’ve invited them to be part of a symposium.
3) Publish, publish, publish. Our old department chair, David Kupfer, used to say, “there are few ills that cannot be cured by more publications.” If you get good publications out, people will seek you.
Do you have any other advice to share with the SIG’s student members?
1) One is to consider the SPR training award, you can find out more at their website: http://www.sprweb.org/grants/index.cfm
2) The F31 mechanism – we wrote an article for the Behavior Therapist with advice about how to obtain one (see p. 12): http://www.abct.org/docs/PastIssue/33n8.pdf
3) This coming year we will have 2 faculty advisors for the NT/TR SIG: Thilo Deckersbach and myself. Ask either of us questions about where to get training if you would like advice.
4) This translational research stuff is hard - don’t expect it to be easy. What that means is that if you’re not passionate about it then don’t bother, but if you are passionate about it, it’s easy to speak about passionately and it’s very persuasive. So, if you have the internal fire for it, most of the people I’ve seen with that end up succeeding.