Skip to main content

Transforming the understanding
and treatment of mental illnesses.

Celebrating 75 Years! Learn More >>

 Archived Content

The National Institute of Mental Health archives materials that are over 4 years old and no longer being updated. The content on this page is provided for historical reference purposes only and may not reflect current knowledge or information.

IRP Podcast: Dr. Robert Savoy: fMRI Master Educator


NIMH IRP podcast with Dr. Robert Savoy: fMRI Master Educator


Dr. Peter Bandettini interviews Dr. Robert Savoy from Harvard Medical School about his experiences training thousands worldwide in the use of Functional Brain Imaging.


Peter Bandettini: Welcome to the Brain Experts podcast, where I meet neuroscience experts and talk about their work and their field of expertise. My name is Peter Bandettini and I am a scientist at the National Institute of Mental Health. I aim, through this podcast, to engage, educate, and inspire. The views expressed by the guests don't reflect NIMH policy. In this, the sixth episode, I have a discussion with Dr. Robert Savoy, a close friend and colleague, who has had a unique and eclectic career in fMRI education. 

Dr. Savoy received his bachelors and masters in applied mathematics at MIT and, in 1980, received his Ph.D. in experimental psychology at Harvard University. In 1981, he joined the newly formed Rowland Institute for Science, under the direction of the late Edwin Land. In 1991 he first learned of fMRI and in 1993 he joined the Massachusetts General Hospital fMRI group, becoming the Director of Functional MRI Education in 1994. Since then, he has conducted fMRI training workshops in fMRI, multimodality imaging, and functional connectivity regularly at the MGH NMR Center several times per year, attracting thousands of researchers from around the world. In addition, he has run similar programs at conferences and at other institutions in the United States, Europe, Asia and Australia.

Dr. Savoy recently won the prestigious education award given annually at the Organization for Human Brain Mapping Meeting as his work in education has been fundamentally influential to early careers of thousands of researchers around the world. Let’s get started. 

Peter Bandettini: So today, uh, with the Brain Experts podcast we have, uh, a good friend of mine I've known him since about '94, uh, and he's been a pioneer in education of fMRI. He's educated thousands of people over the last, you know, 30 years or so. And, uh, uh, he's made quite an impact in the field through his work with education. You know, his, his path that he's taken is, is not the standard path of a scientist. It's, uh, more towards education, but he still works a lot with scientists. So welcome to this podcast, uh, Robert. so why don't you just start us out by, by telling us what your situation is now and generally how you got there, what, what motivated you to get there, and, uh, uh, maybe we'll go from there.

Robert Savoy: Thank you, Peter. Thank you for inviting me, um, to this interview. I don't have a standard academic path in one sense or standard, uh, research lab path in another sense. But in fact, in many ways, I have a very standard path. I, I majored in applied mathematics, uh, at MIT as an undergraduate. But I-- while I was there, I took a lot of physics classes and a lot of psychology classes. So if you think about those three things, you know, applied mathematics with a lot of computer programming on the side, um, physics, and experimental psychology, it's almost a made-to-order background from the 1970s and early 1980s to suddenly find myself hearing electron functional MRI at the end of 1991 by Bruce Rosen, uh, um, MGH, the Martinos Center and saying, "Wow. This is a, a new technology that might allow me to use all these things in a good way." And, um, shortly after becoming involved with that lab, uh, Bruce asked me to think about doing a class. This was in around 1994. And that led to something I thought would go on for somewhere between two and four years and has gone on for more than 25 years and more than 100, um, versions of it both at MGH and in other places.

Robert Savoy: During that time - I mean, prior to my involvement with fMRI - I was kind of a traditional vision research scientist, but I also a novel human computer interface device, which I actually patented, though it didn't go anywhere a-as a real product, um, and published papers in visual perception, something called the McCollough Effect, which is a, a long-lasting visual aftereffect. But functional MRI as a research tool and a-- and ability to look at the brain in a novel way, that was very exciting. And then it turns out getting involved with teaching and lecturing was real natural fit for me. I think I've always enjoyed explaining things to people. As one of my, uh, exceptional students said, uh, "Yeah. It's always fun to look like you're the smart kid in the class and answering questions." And, you know, he sort of nailed me with that. It is fun. And it's fun to know something about a field and be able to answer people's questions. But, also, in the early years in 1994 when this class first started and especially at the first-- one of the first, uh, meetings of the organization for human brain mapping-- it wasn't officially that name yet, but that was in 1996. And I organized a, a, a educational workshop there. Then that kind of put my name on the map for teaching, and it just sort of never went away even though many people have done, um, much more thorough and systematic kinds of teaching. there's a lot of other people doing, um, very good, good teaching. Uh, so that-- other than the fact that I've probably been a compulsive teacher since I was five years old. Uh, the, the family story is I showed my mother how to use her mix master, which she couldn't figure out how to turn on. I'd seen somebody do it when I was five and showed her.

Peter Bandettini: You, you had a novel computer interface device or, or, or hu-human computer device?

Robert Savoy: Yeah. I was working at a place called the Rowland In-Institute for Science, started by Edwin Land who started Polaroid Corporation. It later and is now the Rowland Institute at Harvard. But while I was working there, for a number of reasons, I, um, thought about using a tongue to control, um, systems. So imagine a surgeon doing a surgery but suddenly wants the light moved. Wouldn't it be cool if you had a sensor or she had a senor in her mouth and could just move her tongue on it, and the light would move? I mean, that was the idea behind it. And I actually developed some hardware and software for doing that. It turns out, there were one or two other companies-- I got to go out to Apple and meet with one of their, uh, their people and designers there. And, and they were reasonably impressed with it, um, but, you know, they have lots of people working on, uh, computer human interface devices, and, uh, uh-- so that was fine. And it was-- it was a fun, little episode. And part of the reason for it, I thought it would be useful for, um, quadriplegics or for anybody else who didn't have use of their limbs or hands or usual devices for, for controlling things. And there's a lot of other people who developed other sensors, um, in a variety of context. I'm sure you're, you're well aware.

Peter Bandettini: Yeah.

Robert Savoy: Um, but that was, uh-- that's one of my only two patents. The other-- the other patent was unrelated to high tech. It had to with, uh, a company my father and his relatives were involved in for storing music systems like CDs and tapes and things like that.

Peter Bandettini: What do you think differentiates you from, like, the other, you know, s-- teachers of fMRI?

Robert Savoy: Part of it-- part of it is historical. I was in the right place at the right time, and there were a handful of people involved and in fact specifically, visual perception, who when they heard about fMRI just sort of stopped their normal careers and jumped in with both feet. But at that time, I knew very little MRI physics. I knew some physics and a lot of math and, and computer stuff and certainly experimental psychology. That was my degree. But I really didn't understand MRI physics very well, and I wouldn't do any teaching without having a real MRI physicist to help out. And when my first offer to do an off-site workshop came up, that was in 1997. As you no doubt recall, you and I went to Perth, Australia, [laughter]

Peter Bandettini: Yep.

Robert Savoy: --uh, I think the thing that was special about me was, relative to other people at the time maybe, is in, in addition to the cognitive fit, it's a real ego syntonic fit in the sense that I love teaching. And I think I kept doing it partially for the theater, partially because I like, you know, being on stage and talking to people about these. It runs deeply in my family. All three of my children are basically performers. And even my wife who's a clinical psychologist doesn't spend a lot of time doing clinic-- traditional clinical things. She mostly teaching parents and educators of young children. And in my case, it's trying to treat people in a way that they feel like they're being treated as a colleague or somebody who will soon be a colleague. It makes you wanna live up to the way you're being treated.

Peter Bandettini: I think so much of your teaching is not only during your lectures but, uh, outside your lectures. And when people approach you, you're very, very approachable. you respect them. And they-- and they want to fulfill that sense of respect.

Robert Savoy:When I was approached to teach a class, I went away for two weeks and started thinking about what would I do if I-- you know, wh-what would make it special if I were to do this? Because I didn't wanna do something sort of more routine. And I realized the key was designing experiments. And that hallmark of all the classes I've taught, there's a-- there's a part of this-- the week where people in small groups design their own experiment. And that forces them-- sort of no matter what they're designing, no matter what the experiment is, it forces them to run into all the fundamental problems of the field. How do you design experiment to ask a question? How do you analyze the data that successfully addresses that question? And I think when you were talking about interactions outside of lectures, that's really the main time it happens is working in small groups

Robert Savoy: In the early days, I could do this with one other teacher. But over the years as more people came to the class I realized I couldn't meet individually with people or even the groups. And over the years, there's been just a remarkable collection of scientists who also really enjoy teaching, who really enjoyed interacting with people and helping them design their experiments. And, thankfully, that's continued throughout the whole period

Peter Bandettini: You know, I learned a tremendous amount when I'm teaching, but also, not only as I'm preparing but also from the feedback, from the questions. They sort of uncover gaps in how I communicate.

Robert Savoy: So things like that were very gratifying, but a lot of it had to do with being in the right place at the right time and also feeling strongly about making the classes a little bit special. Forcing everybody to design an experiment was a good idea back then. It's a little different now. I don't have people, you know, design from the beginning because the field has been going on so long. I usually have people pick out some kind of paper or set of papers in the literature and think about extending that design rather than starting from scratch because that doesn't happen so often in real life now. But that's always been the most fun part of the class, I think, for the participants as well as for me.

Peter Bandettini: Yeah, yeah, yeah.

Robert Savoy: While the fMRI course is still the one that I do the most teaching for, over the years, we added a, a two-week multimodal program, which talked about other technologies. And now, by far, the most popular course is one on connectivity where we talk about functional connectivity, structural connectivity, diffusion imaging, um, like matter tracking, etc. That class, which we only do once a year in that form, by far gets the largest turnout of, of any of the classes. And then in addition to that, that, I've been affiliated with, with people who develop the CONN software. Um, and that class, which is just the software training class - I, I mostly play an administrative role there - is another very, very popular, uh, program. And, and indeed, one of the questions you had about sort of where the biggest challenges are. It's largely in data analysis. That's where you need to be the most careful. There are plenty of places you can make a mistake when you're doing fMRI or other brain imaging experiments, but I think the most-- the single most, um, tricky part is analyzing the data. And, you know, programs that teach how to use that software, no matter what the program isthey're always oversubscribed. People are just dying to get good training in those software packages. And I think that for good reason it's hard to-- it's hard to do that well, and I think it's probably at a practically-- practical level, the single most important thing to learn to do well.

Peter Bandettini: I think one thing that you instill in everyone is, to realize that the really interesting research goes on with a certain amount of creativity it's not like a formula that you follow. Uh, it's sort of like a set of tools that we're still figuring out how to use as best as possible.

Robert Savoy: One of my more favorite moments each time I teach a class is I make the following assertion - and I've checked this with every software developer, and they all agree - and that is, you can collect data for two hours in a single subject with one fMRI experiment and spend the rest of your life analyzing that data. The amount of information there potentially and the number of different kinds of analytic approaches that you could use with the data is so large that there's essentially no end to it. And so it becomes-- if you actually wanna be productive and get things published, you actually have to design your analysis almost as carefully as your experiment. And it's-- so that's a general statement. I-I've, I've had contact with people in the pharmaceutical industry more, uh, in the last few years. One of my colleagues gave me a sort of lesson about, uh, the way you have to do data analysis if you're trying to develop drugs for the FDA is very, very, very carefully looking over your shoulder. In fact, it's not looking over your shoulder. It removes the entire process from you. You fund it. You design it. You specify how the data is gonna be analyzed, and that's the last time you touch it until somebody else gets the data and analyzes it. Because it's the only way you can reasonably protect the integrity of the experiment. And that has all kinds of interesting consequences. But one of the things is-- and that context, it absolutely forces you to say exactly the way you're going to design and analyze-- exactly the way you're gonna analyze your data.

Robert Savoy: And you and I both know over the last 20 years, there have been a number of controversies associated with fMRI data analysis. And one could argue that at root, most of them stem from seeing something, thinking it's interesting, and maybe analyzing the data a slightly different way in the light of that information. And that's precisely what you're not allowed to do in drug development. You can get great ideas for your next drug study, but the study that you're doing a phase three trial or whatever that you're spending $100 million on, that you have no say beyond the design as to how the data is gonna be analyzed. And that's a really interesting form of discipline. Uh, and I tried to teach that a little more or emphasize that a little more than I have in the past.

Peter Bandettini: When I give courses, uh, when I give talks, it's like I always like to emphasize that, "It's not just a black box." I mean, it's not like, you know, the fMRI signal goes in, and you pull it out. It's sort of like, you know, it has a latency. It has a shape. It varies box to box, and we're still-- we're still trying to figure this out. Um, so what do you think, uh, in your teachingWhat do you think is the most common deficit that people generally have coming in, or has that changed over the years, and, and is it fixable?

Robert Savoy: Has it changed over the years? absolutely. Um, you know, in the earliest days once the, the word of fMRI got out, everybody, every clinician thinks about how they're gonna use it for their-- especially neurological or mental disorder. So people jumped on the bandwagon, some of whom had no background, no experience in psychology or experimental design. I mean, they were radiologists. They were physicists. They were anything but experimental psychologists. And that showed up in some of the flaws in the designs that people did including in published literature. You know, it came out of the fact that that was not part of their training. That's much less true now. People because-- you know, these things are taught in college courses. Nobody comes to the classes anymore. Uh, w-what-what's the story? When we first did the class in October of '94, none of the 14 people had ever been near an MRI machine before, okay? So they were seeing it for the first time in that class. Whereas now with 30 people, virtually all of them have been involved with being a subject, running experiments. That typical person is somebody who's been around an fMRI-based lab for a while and wants to get a sort of more thorough, you know, soup to nuts basics in the foundation.

Peter Bandettini: So where do you see this going? you're still teaching your classes and, and, uh, uh--

Robert Savoy: Sadly, not, not this year. This is the first time that I haven't because of COVID, of course. Um, and I chose, for whatever reason, not to work hard to develop online versions. lot of the, um, the strength of it is the interaction in the classroom is interacting with the people directly. And I hope to do that again on October, but if not October, then after that. So, um, yes. I continue to do that. I have done four or five courses a year..

Robert Savoy: I would like to say a little bit about, um, psychology, multiple personality disorder, and psychoanalysis because they actually played a role, not only in my personal history, but specifically in my getting involved with fMRI in the first place.

Peter Bandettini: Yeah, yeah.

Robert Savoy: So, um, you know, I am a physicist by training, and I almost got a degree in physics except they required a dissertation as an undergraduate and applied math didn't. So even though I took all the courses for both physics and applied mathematics, I didn't wanna write the, the physics dissertation, so I just got the math degree. But sitting on top of those very heavily math, physics traditional science background, I've really always been in-interested in psychology. I remember Hans-Lukas Teuber who famously started the psychology or neuropsychology department at MIT. Um, when I was a freshman, the week before classes started, people asked him questions about his intro psych class, which was an incredible class. I still have the notes from it back in 1967. And one of them said, um, "What about behaviorism?" And he drew a sketch of the brain on the board, and, you know, Harvard which is about up two miles from MIT, uh, was a bastion of behaviorism historically. And he just said, "Uh, this in an outline of the brain. Unlike our colleagues up the street, we think it has something to do with behavior." [laughter] And then somebody asked him, "In your introduction, will there be any Freudian psychology?" And he said, "Freudian psychology? Isn't that a contradiction in terms?" [laughter] And what struck me was that he had-- I mean, brilliant teacher, great guy, great scientist, many things positive to be said, almost endless. But he dismissed two really interesting, important pieces of the history of psychology in the 20th century with jokes. And that didn't make sense to me at the time, and I think it was just wrong. And, you know, I was a graduate student. And as an undergraduate, I learned a lot about behaviorism. But when I was a student, psychoanalysis was a no-no. It wasn't science. It was all religion, blah, blah, blah, blah, blah. And for a variety of reasons, I knew that that was shortsighted. It wasn't that it was science. It wasn't that Freud was a great scientist. But Freud, like people who developed fMRI, found an interesting window unto the brain and behavior in the tool of psychoanalysis, just in the process, not unlike Zen meditation of saying, lie on the couch, freely say whatever comes into your mind. That turns out to be a really interesting way to get at psychology. That's all general background.

Robert Savoy: Um, when I first visited the MGH fMRI lab, I saw the room, bay three, as you no doubt recall, which had an MRI machine. But I was told that bay two, which was right next door, would be getting its own fMRI machine. And I suddenly realized, "Oh, my goodness. You could do an experiment where a patient or analyst and-- was in one magnet, and a psychoanalyst was in another magnet. And you could watch the brains of both of them at the same time, and wouldn't that be interesting?"

Robert Savoy: And I really thought potentially that fMRI would allow, um, the study of both people doing that, you know, on both sides. And in theory, it can. In practice, it has never come close to succeeding in doing that. But that was a, a, a piece of the background. It was that interest. And the closest I ever came to it was when somebody asked me to look into patients who had dissociative identity disorder. Because I think the mechanisms-- this is not a clinical statement. It's my sort of intuition, scientific intuition, not clinical. There's clearly a splitting off of part of the brain's function from another part and-- whether motivation or other stuff. We're not gonna have that conversation. But just on a phenomenological level, that's an interesting piece of psychology. So that was part of the reason I was interested in all these things, why I got involved in that research, and, and why the idea that there would someday be a second magnet. But nobody's doing psychoanalysis in the magnet yet.

Peter Bandettini: Yeah. That would be-- I mean, having like a, you know, hyperscan where the psychoanalyst is in one scanner. And--

Robert Savoy: Right. That's the term. Hyperscan is what people-- um, I forget the name of the gentleman in, um, in Texas who sort of got funding for having five magnets for doing social experiments, and he called it hyperscanning.

Peter Bandettini: Yeah.

Robert Savoy: But he didn't succeed in getting adequate funding for doing things like that as far as I understand.

Peter Bandettini: Yeah, yeah. But I mean -- there's a really strong sense that there's still so much to do, uh, with fMRI to look at neural correlates of, you know, uh, uh, either social interaction or engagement to therapy or even biofeedback. And so it's still wide open, and it still lends itself to people who are creative and can make connections between other modalities. I mean, you even talk-- you even teach-- I mean, it's sort of like they shift a little bit but you-- a big part of what your course is, you sort of branched off into not only basic fMRI, but also, you have a multimodal, uh, course as well. Do you wanna talk a little bit about that?

Robert Savoy: Sure. That, that course actually we don't do anymore. It was-- there was an NIH, um, program that funded, I think, four institutions for-- and it ended up being 10 years, two 5-year rounds, for doing something like multimodal imaging. And I think three of the four places chose to focus on, uh, fMRI and EEG and MEG. But because the Martinos Center also had optical imaging and PET and other thingsI heard a lot of people talking about different aspects of multimodal imaging, and I proposed that that's what we do. That we take advantage of all the different modalities and the expertise of the Martinos Center, and we did that for about 10 years. It was a very select group of people. You had to really apply to get into it. Um, and it, it was fun. And-- but it went from 2007 through 2016 for 10 years. Um, and, and then we tried doing versions of it, but it, it-- very hard to get people to make that kind of commitment, both the students and the faculty, um, for, for doing a class like that. So I haven't managed to figure out a way to make that still go. It was a lot of fun, very interesting. But I think the fundamental problem with it was when people wanna learn about fMRI, they wanna learn about fMRI. And if they wanna learn about doing PET studies, they wanna learn everything about doing PET studies. And if people wanna use TMS, transcranial magnetic stimulation, they wanna take a class just on TMS. In those classes, there are optical imaging classes., that's the way people want their training. There are a small handful of people who would love to get a little bit ex-- of exposure to all of them, but there weren't enough to justify the faculty commitment and the amount of time that students would use it.

Robert Savoy: So that's what happened to that program. It's, it's one of my favorite memories because it was two weeks long not a week long. The students, they just bonded much more. Very funny, enjoyable things happened.

Robert Savoy: The main new direction is connectivity, structural and functional. And that's where I incorporated in the-- um, uh, in a regular fMRI class a little bit, but it's mostly its own class, and I should add. That came out of the multimodal course. In the middle of those period of doing multimodality, it seemed to me a lot of people were interested in connectivity, so I organized a one-afternoon workshop. The class had 20 people. 200 people from the Boston area showed up for that workshop.

Peter Bandettini: Wow.

Robert Savoy: And that's when I said, "Okay. We need a course on connectivity," and that's, that's what happened back in 2011 or '12.

Peter Bandettini: The whole idea of connectivity is sort of, like, you know, you have structural connectivity with, you know, doing anatomic imaging of white matter tracts and functional connectivity which is this co-covariance of fluctuations, and, and it's tricky I always try to caution people that, you know, it's interesting to try to pull out the information but not to over interpret, you know, what those coactivations mean or, you know.

Robert Savoy Absolutely, absolutely.

Peter Bandettini: Yeah.

Robert Savoy: One of the fun things is there are some anatomists who really look at what white matter is actually doing in the brain, and they give, you know, sort of a lecture at the end to say, "All right. This is good for much of the brain, but when you actually get to the ending of these axons, they're making right angle turns, and they're doing all these things that MRI can't pick up at this point." And, and so you have to be very cautious in the interpretation as you say.

Peter Bandettini: Yeah, yeah. So I have a few more questions that I kind of really wanted to, uh, sort of get your perspective on. so when you're teaching-- I mean, obviously, different students become successful for different reasons or not successful for different reasons. But is there anything that you feel that you can teach and things you can't teach. I mean, are there things you just can't-- that can't be taught?

Robert Savoy: Um, that's a wonderful question. My general sense was there are extremes. There are people who no matter you did, they were going to be good scientists, like, from day one. And there were a very small number of people who nothing was gonna help. There was one particular person who was so clearly clueless about what it meant to be a scientist that I-- it, it-- I was not involved because I had to pull my hair outUm, but those were rare. For the vast majority of people, you cangive people enough information and enough of the ideas that if they are motivated, they now know how to go further. And they have to do it on their own. They can get help. They can get information. But for most people, getting that kind of information is gonna be useful, even if it-- what it does is it leaves them to say, you know, "This stuff's too complicated. It requires too many people and too much money. I've got better ways to spend my career." And, you know, a number of people did go that way. some people said, "No, no. I can do this. And it's really cool." And then they go on to build their career around it. I can't take, you know, pretty much any credit for those people and their success, but I can say that I did something to give them enough information to make intelligent decision about how much time and effort they wanted to put into it. Um, that's the best I can give you for that question. It's a good question, but I don't think there's a, a suc--

Peter Bandettini: Yeah. Nobody knows really. I mean, right. Yeah just to shift a little bit. What does it really mean to, to communicate clearly?

Robert Savoy: Let me say two things about that, one of them very general. Uh, for a long time, I wanted to learn to draw. But you don't learn to draw. You learn how to draw a nose, and you learn how to draw an eye. You learn how to draw a duck, and you learn how to draw a house. And if you do enough of that, things start to come together. Similarly, you don't learn carpentry. You learn-- you learn how to build a house. You learn how to hang a door in a doorjamb and how do you make it level so that it opens and closes just right. You learn how to use a chisel so that the thing actually catches in the latch because it didn't get put there right. And that's different from what you need to know to put in the electrical system or the plumbing. And it's this large collection of things that turn you into a real carpenter or home builder or whatever. In a similar way, there are a collection of things that contribute to good explaining and good teaching. Some of it has to do with the mechanics of how you make slides. I remember having a colleague who I always felt put too much information on his slides. [laughter] And, um, and I try with-- occasionally with my teachers. I will take them aside and teach them how to make better slides, how to focus the story a little bit. Now, I tend to wonder too much. I tend to put too much in

Robert Savoy: I think that's the general thing that there's a lot of tricks, mm, th-th-that, that generally, it's a lot of little pieces, making good slides, not trying to do too much. But if I had to give one overarching thing, I use the electrical engineering term of impedance matching. And what I mean by that is trying to get a sense of who your audience is and tailoring the way you explain something to the person you're talking to.

Peter Bandettini: Yeah.

Robert Savoy: And then there are-- there are occasional cases of people who teach to show off how smart they are, and I would argue, intentionally make it difficult for the student That's rare. More common is people who really, really are motivated to teach and really try and have just have not had the right training or don't have the right intuition for it. So they don't do the impedance matching. They don't prepare slides in a way that, um, are effective in conveying the idea they want and keeping the audiences' attention.

Peter Bandettini: Yeah, yeah. Okay. So one of my last questions at least is, uh, is there anything in your whole career that you'd have done differently? [laughter]

Robert Savoy: So there are many things that I think I could have and should have done differently or at least would have opened other paths. You know, starting from becoming an MD. I mean, there's-- there are things you learn by be-- getting a medical degree as opposed to, you know, a more purely academic degree. But there are many things like that. Um, and also, as you well know, I, for historical reasons, for my employment reasons at the Rowland Institute, I was not under the usual publish or perish pressure. And that made it easier for me not to, you know, tradi-- build the traditional lab. There have been prices and consequences for that, not all of them good. And, you know, certainly I would rethink some of those. I don't know if I'd make a different decision, but I wish I'd had more perspective when I was younger about that. But with regard to the teaching as such, um, I wish I had written more. I wish I had taken the time and put the effort into creating things that were more polished in terms of writing for people. In terms of the classes themselves, aside from individual mistakes-- I've certainly made personal mistakes occasionally. People have said I was too pushy about something, but lots of people have also said, you know, "This is, like, the best class I've ever taken." And, you know, it's hard to please everybody. But separate from the fact that it's hard to please everybody, as you know, I have a somewhat stronger dynamic personality, and sometimes that can get in the way. And I always start off my, my lecture series now talking about the effects of coffee.

Robert Savoy: I, I, I think you've seen that slide. They're my favorite pair of slides where I talk about the way computers can be thought of and understood at different levels from, you know, quantum mechanics, up through the internet. And I point out that you can understand one level by just knowing a little bit about them below. You don't have to understand how the internet works in order to study how to build a better transistor and vice versa. And then I give the counterexample of looking at a brain from neuroanatomy, up through social interactions. And then I point out that if I have a cup of coffee before a lecture, I start talking too fast, and I get very impatient with questions and loud noises bother me, and I even say things I'm not supposed to say. And what I'm pointing out there is that a small biochemical change not only affects my biochemistry, it affects it all the way up to my social interactions. And similarly, if I were to say to you in a convincing way something really horrible about a relative being in an automobile accident or something and you believed me, that would be a purely social communication, but it would measurably affect the chemicals in your blood stream right away.

Peter Bandettini: That's--

Robert Savoy: So the, the, the levels of explanation and the interconnectedness in neuroscience is very different from many other areas of research, and that-that's a big challenge. Um, so my point about that was it took me a while to learn not to drink coffee. And sometimes, my-- I, I probably did or said things I would rather have not done. Um, and, and, and it, it depended on the student. Some students-- you know, there's a wide range of people in the world.

Robert Savoy: when I'm forceful in one context with somebody who's strong in their own right, it can be enjoyed and may be educational. And if I do the same thing with somebody else, that's not good. And it can be intimidating and, and be very counterproductive. So it's a little bit like impedance matching. You wanna be sensitive to who you're talking to. That's sometimes hard when you're talking to a large group. Uh, it's a little easier and more important when you're talking individually with people.

Peter Bandettini: Yeah.

Robert Savoy: I hope that a-addresses your question, but it's my attempt.

Peter Bandettini: Yeah. That's perfect. Um, okay. So I think if people are listening to this they get a good perspective of, the things that have interested you and your methods of teaching a little bit and, and sort of the whole landscape of, of education in fMRI which is-- which is sort of something that, that people-- uh, and also brain imaging in general that people-- it's sort of a blind spot, I think, in the field. People talk about it. There's all kinds of communication. There's all kinds of tweets and posts and publications and things like that and meetings, they're getting educated but, but having people like you to sort of, uh, uh, gather and direct, uh, the information and share the info--

Robert Savoy: I, I try to give them a running start. I tell everyone, a week of this kind of information will give you information to decide if you wanna go further. But if you do want to go further, you're gonna have to spend six months to two years in somebody's lab and be surrounded by people, analyzing data all the time, surrounded by people who are doing the same kind of thing just as in almost any other serious research endeavor. You'd, you'd need that, that much for this stuff. But I can supply kind of a running start. And that's always the way I've thought about it.

Peter Bandettini: Yeah. And I think your course is also not only give a running start in terms of information, but you're suddenly introduced this very supportive network, uh, of fellow students.

Robert Savoy: I think that's fair. It's both within the class and every once in a while, I, I-- usually on the first day, I comment that, "Look, we're gonna cover a lot of topics. But if you have a question or topic typically something like animal research that we don't talk about, I'll try to arrange for you to have lunch with somebody at MGH who studies animals, or we're not gonna look at pain research as such in detail, but I know somebody at McLean who could come in and have lunch with you and talk about pain research." And, and that typically happens once or twice in each class. Somebody takes advantage of that, and I try to connect them. Uh, that-that's more of the impedance matching. It's trying to be specific to the individual needs if you can.

Peter Bandettini: Yeah, Well, , we're getting close to our time. And so I just like to thank you again, uh, for, for taking the time during this, this crazy COVID time and, uh, uh, to, to give this interview.

Robert Savoy: You know it's been a pleasure. and thank you for inviting me.

Peter Bandettini: Well, thank you. Thank you.