00:00I can still remember the day that Aaron and I were sitting there and you know he had run functional connectivity with some of these lesion locations and we saw boom this anti-correlation to the extra striated visual cortex the exact spot of the brain involved in visual imagery just popped out moments where you know I kind of couldn't believe that it was working right yeah you know like when Ryan Darby first showed that lesions that caused delusions of familiarity right are connected to the familiarity detector and they're all connected to our reality monitor and like all of a sudden was like oh for for a century people have hypothesized this is a dual hit syndrome there's been a lot of very very exciting moments as you know we saw a different lesion that would map the results and kind of you know put our hands behind and said wow that's really cool welcome to stimulating brain you Mike Fox is the director of the Center for Brain Circuit Therapeutics where he is also the Raymond 01:21D. Adams Distinguished Chair in Neurology at Brigham and Women's Hospital in Boston. A few years back when Mike moved from the Beth Israel Deaconess Medical Center to the Brigham the reason was to create exactly this Center for Brain Circuit Therapeutics where the idea was to have colleagues from neurosurgery, neurology, psychiatry and neuroradiology all under one roof to work together and solve problems of neuromodulation in all types of brain disorders. I should probably mention that Mike recruited me to the center last year so he is my boss and dear friend and dear colleague and also one of my most important members of the Center for Brain Circuit Therapeutics. I'm very happy to have him as my mentor. I owe a lot to him and have learned so much from him. So that's why I'm super 02:03excited to share this conversation I had with him about his work in resting state connectivity in the earlier phases but more and more so on brain mapping using a connectome. In 2016 when I was a postdoc in Mike's lab he once phrased that you know we have the connectome now so we have the map but where do we go with it? And I think Mike is one of the few people that really has solved this question especially with the aim of clinical research. So I think it's really important that we have this conversation and that we have this conversation. So I think it's really important that we have this conversation and that we have this translation using the connectome to map functions and symptoms to brain networks with the aim of translation. He uses causal sources of inference for example brain lesions, TMS stimulation sites or DBS electrode placements to look at which networks map to which symptoms or functions in the brain. The aim here is to use these causal sources of information to create novel neuromodulation targets both for invasive and also non-invasive neuromodulation. Our conversation was enriched by guest questions of former and present colleagues that have worked 03:04with Mike such as Shan Sidiki, Aaron Bose, Michael Ferguson, Dan Korb and Fred Schaper and I'm super excited to be part of this community of brilliant minds that trained under Mike and many of which have now begun to start their own labs worldwide in different places such as Utah or Finland or Germany. This is a growing community of people that are using these network mapping techniques to find out which networks map to which symptoms and symptoms. So I think it's really important that we understand which neuromodulation targets are suitable to treat which symptom in the brain. So for me Mike is really one of these very stimulating brains and I'm pretty sure you're going to enjoy this conversation I had with him. Thanks for tuning in stimulating brains. Okay Mike thanks so much for doing this. As you know to break the ice before we get into science I always ask about hobbies. So what do you do when not involved in science? 04:01When not involved in science. Well I like recruiting very talented faculty members to my center. I don't know if that counts as a scientific hobby or not. It still works. As you know I live on a lake. I absolutely love the outdoors so paddle boarding, kayaking, playing with my kids either on the lake or in the pool. But my big pastime right now is being a dad and then if I can be a dad outdoors even better. Great. You once mentioned that as a kid you would go outdoors and find fish in the lakes, put them into your aquarium. You were always an outdoor person right? Yeah well not just fish it's whatever I could catch. Crawdads, turtles, frogs. I grew up in the woods and the more I can get back to being in the woods. Now that I have a seven and a nine year old I have an excuse for catching craudads and fish and turtles. Sounds great. Sounds great. So going into life. Your career, science and medicine. 05:00Who were key mentors and turning points? Yeah so certainly the dominant was my Ph.D. mentor Mark Riegel. He was my first scientific father and just learned a massive amount from him. Obviously I learned a lot about brain imaging and just how to do science and write papers. But more he continues to be my role model for how you run a lab, how you run a center. Yeah. Just politically how you solve problems amongst different people in the lab. And I still remember going to a conference and anytime Mark Riegel goes to a conference it's, it is, he's like I guess now the grandfather. Yeah. Walking through these imaging conferences where it's not just what he did with his own career it's what all of his students did. And you know I'd say that's my model for where I want to go. Yeah. And you train the best people and that allows you to one let go of things and pivot and focus on something new. 06:04Yeah. So that's a real important lesson that he taught me. And then two it's, you know you see how science unfolds and how your efforts become multiplicative as people go off and do their own thing. And for me that was so salient you know watching him walk through this conference that that stuck with me ever since. And it continues to be my career goal. So essentially creating mini mics that would then pivot off and do their own thing. No, no, no that's the whole point is that if it's a bunch of mini mics then you failed. Okay. Right it can't be mini mics. It's got to be you know people that think differently than you. They learn something from you. That's right. You had an influence on them but they think differently. They do different things. They do things that you never would have thought to do. Okay. Right. Yeah. And so they're not mini mics. They're better than mini mics in important ways. Yeah. And that's how you actually have influence on a field. So Mark Riegel taught me all that. And then obviously Alvaro Pascaleon my post-doctoral mentor really got me laser focused on translation. 07:05And always thinking about cool science is cool science but what do you do with it? How do you take that science and translate that into in the case of Alvaro a neuromodulation intervention. Yeah. And so that obviously had a huge effect on how I think about things and how you run a center and how you build a center focused on that goal. Cool. And do you continue contact with both of them? I do. Yeah. It's very important to me. I think a good mentor is a mentor forever and they continue to have things that they can teach you. Yeah. Sounds great. So I will start with the first guest question by Sean Siddiqui about that first part of your professional life. Hey. Sean Siddiqui here. So I've got two questions for this interview. First of all, where do you think you'd be right now? If you've never gone to medical school or grad school and you decided to continue being an engineer, what direction do you think your career would have gone? 08:00Yeah. It's a good question. So at the time that I decided to give up electrical engineering and go to medical school, I did have a nice job offer on the table where I had worked for Procter and Gamble for a couple of summers. And I did not fit in with the corporate culture at P&G. And so their offer to me was, hey, we're launching this new initiative where we're going to take all the crazy people and stick them together in a building and hope they come up with creative things. And your job as the electrical engineer will be figuring out can you manufacture any of these crazy things that people come up with. So every week you're going to be trying to build a new assembly line with a bunch of robotic arms doing different things. And so that's probably the job that I would have taken if I had not gone to medical school. Now, would I have stuck with that? Would I have been a good student? Would I keep doing it? Would I have gotten bored of making diaper machines at some point? I don't know. But that was the job I probably would have taken if I said no to medical school. Okay. 09:00Sounds great. And I think you did then go to an MD-PhD program and wash you. And you once said something like it's a good program for people that can't decide yet. Yeah, that's fair. No, I was very attracted to medical school. I was also very attracted to graduate school. I also liked being an engineer. But I think the MD-PhD, it gave me, at least me, I'm sure there are people out there that know what they want to do, but it gave me a low-risk way to explore both options and figuring that I'll push the decision down the line to some point and decide if I actually want to be a doctor or be a scientist. Sounds great. Sounds great. So your earlier work with Mike Reichel that you just mentioned focused on, I think, a lot of things, but also anti-correlated networks. So you were a pioneer in using and applying resting state functional connectivity, also developing it in the brain. Maybe we should start by briefly mentioning what resting state fMRI is as opposed to task-based fMRI. 10:03Yeah, sure. So, you know, task-based fMRI is, you know, what a lot of people, I guess, used to think of as standard fMRI. I guess it's become pretty popular. But you can think of the standard preoperative mapping pair. Yeah. So, you know, the standard preoperative mapping paradigm where you put somebody in a scanner, you have them tap their fingers, and you see what activates in response to the finger tapping. Yeah. And that identifies brain regions that activate during a certain task. And people use this task-based imaging for the early PET studies, for the early fMRI studies, and it was mapping out, you know, what goes up when you do a particular task function. And then resting state was the idea that you could put someone in an fMRI scanner, not have them do any specific task, right, just rest quietly in the scanner and try not to fall asleep, although it turns out that you can still see spontaneous activity when people are asleep or even under anesthesia. But you just look at these spontaneous fluctuations in the fMRI signal, and those are temporally correlated within connected brain regions and brain networks. 11:05And so it gave us a new way to map out not what makes this brain region go up or what activates this brain region, but what is the functional relationship between different brain regions. And that was heavily influenced, or at least it was informative to have PET, which Mike Reichel worked before that, right? I think he used, would you say so? So it was motivated by a couple different things. One was just, you know, observation by Bharat Biswal way back in 1995, who just noticed, hey, gee, you know, there's a lot of spontaneous activity in this fMRI signal. Oh, look, it seems to be correlated between the left and the right. Yeah. And the left and right motor cortex. Yeah. Right? And it was just something he saw, and it was an observation that sat out there for people to capitalize on. And eventually they did, and did big time. But I think for Mark, he came at it from a slightly different perspective. So, you know, Bharat Biswal's finding had been sitting around for 10 years before, you know, I or Mark Reichel ever got involved in the field. 12:06And I think what piqued Mark's interest was two things. One is Mike Gracious published a very nice resting state fMRI study showing the default mode network. And the reason that that really raised our eyebrows is, one, we were interested in this network, but then, two, it was in multiple different vascular distributions. So there was no way that that could have emerged through, you know, vasomotion or respiration or cardiac pulsations or a lot of the artifact that people were worried about following the Biswal paper. Yeah. So the Gracious paper got us really interested. And then Mark had always come at the brain from an energetics and pet-motivated perspective, which is what we see is that the brain is extremely active at rest. Yeah. And that when you actually do a task, you get this itty-bitty increase above and beyond this massive amount of ongoing resting activity. Yeah. And for Mark, his first insight into that was this default mode network, you know, areas that go down when you do a task, right? 13:04Sure. Well, okay, if there's areas going down when they do a task, then they must be doing something at rest. Yeah. Right? And so then when this resting state fMRI came around, Yeah. you could see that this is another reason to get interested in it is this could be a window into the brain's enormous ongoing resting energy expenditure that you could see with these PET scans and you could see based on regions that went down when you did a task. So you're saying with PET, you wouldn't be able to see that the default mode goes up during, could you do that? Like if you, in theory, you could probably. You could, yeah. But it's easier with fMRI. Right. Yeah. So you have the, you know, the resting FDG scan. So just what brain regions are consuming the greatest amount of glucose at rest. Yeah. Right? So the FDG PET, you know, was a resting state scan. Yeah. Right? Before they really had task activation mapping. Makes sense. Super cool. So, and then your like landmark paper in PNAS from 2005, it was cited, I think, over 8,000 times by now. 14:08What did you find there and why was it so important? Yeah. So first off, at least a third of those citations are saying that the paper's wrong. So if you do a controversial paper, like, you know, you can accumulate citations. But, you know, I think most of the citations are just, that was the first paper where we kind of, if you will, put together the processing recipe for cleaning up resting state fMRI signals. So I mentioned, you know, Brad Biswell had a beautiful paper. Mike Gracious had a beautiful paper. And then when we tried to do it at Wash U, we had a lot of trouble. Okay. In fact, I wasn't the first grad student that they assigned to the project. They had tried multiple times before I tried. And the problem was is that the resting state signal does have neural activity in it, but it has those cardiac pulsations, respiratory artifact, movement artifact. And how you get rid of that wasn't trivial. 15:01And so for us, we had some resting state data. And it was trial and error over the course of about a year and a half. Where, you know, sitting down every night. Sitting down every night with Avi Snyder, trying different things, trying different algorithms, trying different combinations of regressing out different signals. And then finally we kind of developed a recipe that worked. And all of a sudden these maps started popping out that just looked way cleaner than anything we had seen before. And so I think a lot of the citations of that paper are, you know, that process of, you know, regressing out the motion parameters, the temporal filtering, you know, the global signal regression. Yeah. But regressing out these covariates kind of led to a way to clean up the fMRI signal. Yeah. So that's part of it. People just cite it as a generic resting state fMRI paper. And then the other observation was this idea of the push-pull networks, these anti-correlations. And, you know, yes I am. I do realize anti-correlation is not a word. 16:02But we had multiple versions of the paper. We had the test positive network and then the test negative network. And we actually had a version where it was the test positive network, test negative network, and negative correlations. And it was just unreadable. I see. Right? Because the positive negative was meaning multiple different things in different places. And so anti-correlation was stuck in there just to make the paper more readable. More readable. Okay. But for us I think it was exciting because it suggests that these resting state fMRI maps told us something more than just what is anatomically connected. Right. It was telling us something about the polysynaptic functional relationships between regions. And over and over again with TASC you saw that a certain set of regions went up, a different set of regions went down. And the idea that that functional relationship that we saw over and over again in different TASC paradigms was there represented intrinsically in the pattern of the ongoing spontaneous activity. It really made us think differently about what these maps were showing us. 17:02Yeah. And what the maps are capable of telling us. Yeah. And I would still say that, you know, sometimes there's criticism that fMRI hasn't taught us much about the brain yet. But I think that is one key thing that the whole organizational architecture that these massive networks ramp up and down and seem to stop each other. Right? When one's active it might have an intrinsic function of tuning down the other one. I think that's one possible interpretation of the brain. Okay. I think that's one possible interpretation of the data. But that's not proven yet. You know, there's not, you know, direct inhibitory connections between, you know, these two circuits. And so I think all the fMRI signal really tells us is that when this one goes up this one tends to go down. Yeah. But that doesn't tell us the mechanism. It doesn't tell us how that happens. It doesn't tell us why that happens. Yeah. 100%. But it's, I do think it is helpful for mapping out the topography. It tells us where it happens. Yeah. Sounds great. 18:00Do you want to briefly tap into the global signal regression debate or shall we skip that? Well, the short answer is no. But I think it's for the listeners. I think it is useful as an example where everyone has to figure out what their balance is going to be between digging into the methods and just taking a method and running with it. Yeah. And actually before we ever published that paper we had the follow-up paper on the data. We had the follow-up paper on global signal regression fully written. Right? We knew that it was an issue. We knew that it was a confound and it was actually part of the original paper. The problem was is it was going into a problem or a question that nobody was even asking. Yeah. Sure. And it's very hard to answer a question that no one's asked yet. Yeah. And so Mark's advice and it was I think accurate was you know let's just publish the paper. Someone will then ask the question. In fact we would oftentimes talk about this issue in talks. And so then people started asking the question. Then we were ready with the follow-up paper to say hey here's our best set of analyses 19:03that go after is this a real artifact or is it something meaningful biologically. Yeah. And then the field kind of did its own thing where lots of different very smart people went after this. Is there a better replacement for global signal regression? Is it a confound? Is it a not confound? And I'm really glad that people dug into it but it became such a debated methodological step that at one point people were getting their papers rejected solely because they either did this processing step or didn't do the processing step. Yeah. And that was a very frustrating period of time. And actually Kevin Murphy and I got together and wrote this you know opinion piece on just trying to get a consensus. Yeah. So hey actually we pretty much agree on everything here. Yeah. These are just different ways to look at the data that are useful in different circumstances. And you know the analogy for neurologists I always can give is nobody fights over how you montage the EEG. Yeah. Right. There are different ways to look at that EEG study. Yeah. And so I think that's a really good point. Yeah. Yeah. There are different ways to look at that EEG signal. The right way to look at it is the way that shows you the seizure. 20:01Sure. And it's the same thing with the rest of this data from Ryan. Yeah. But I do think that it's great that there are people out there that really dig hardcore into the methods and that's how we make the methods better and refine them better over time. That being said if the whole field did that then we would never be able to move forward and use any method for something useful. Yeah. So I think there's a balance there with any methodological debate. Sounds great. So but one thing I think that will be important for the next guest questions is that these anti-correlations or negative associations, negative functional connectivity values might tend to increase a bit if you do this global signal regression step. So it was a bit up for debate. Are these real? Are these physiologic and so on? So I'm going to play two questions because they are quite similar and you can maybe answer them together. So this is by Aaron Bowes from Utah. What do you think is the significance of anti-correlated networks in understanding human brain function? 21:04And then the second one is from Michael Ferguson. Hey Mike, it's Michael Ferguson. My question for you is what finally convinced you that anti-correlated activity was a real physiological principle and how did you feel? All right. So both great questions. You know, there's lots of ways I can answer Michael Ferguson's question. And there's a development. We did a lot of analyses and, you know, we've published on a lot of them of different ways to look at the data and simulate the data and cut up the brain and, you know, trying to convince ourselves that it's not all artifact, right? Do anti-correlations get a little stronger with global? Yeah, they definitely do. And so it's a question of degree. But to link the two questions very nicely, I'd say one of the best pieces of evidence that convinced me that the anti-correlations were real. And I think the most real is when Aaron Bowes developed lesion network mapping. And so I still remember the day where Aaron and I were sitting there and Aaron was trying 22:04to understand lesions that cause visual hallucinations. And he had, you know, put together lots of cases of patients with peduncular hallucinosis. So a stroke in the brain stem erythalamus that ended up having visual hallucinations. And it was very hard to understand how a stroke in these non-visual parts of the brain, you know, the brain stem, the brain stem, the brain stem, the brain stem, the brain stem, the brain stem, the areas of the brain that are involved in visual areas. Could result in hallucinations. And for a century, there had been this long standing theory of calling them release hallucinations. The idea that somehow the lesion is releasing activity in areas of the brain that are involved in visual imagery. And that leads to hallucinations. And I can still remember the day that Aaron and I were sitting there. And, you know, he had run functional connectivity with some of these lesion locations. And we saw boom. This anti-correlation to the extra striated visual cortex. No. No. No. No. No. No. No. No. cortex, the exact spot of the brain involved in visual imagery just popped out. 23:01It made so much sense with so many long-standing theories of where these hallucinations must be coming from that that really hit me. That without these anti-correlations, there was no way to link up these lesion locations with the part of the brain that we were pretty sure was involved in visual imagery. And then there's been lots of other pieces since that time. But I think we had to switch to causal sources of information. So there's been lesion data that convinced me that there might be something there. There's been TMS and DBS data that's convinced me that there might be something there. But in all these cases, there's not a good way to link a causal site, a site of brain stimulation or a lesion site with its effects on symptoms. So, without taking into account the anti-correlations. Sounds great. So, we will of course dive into lesion network mapping. I wanted to briefly maybe before we go into that, ask you. 24:01So after your MD-PhD at WashU, you completed your residency in neurology at MGH, so Mass General Hospital. And I remember you recollecting a typical Boston residency anecdote once where on a snowy day, one would see medical residents on their skis coming down Beacon Hill to be punctual for rounds at MGH in the morning. So I just thought to pick your brain because I rarely ask you that. Any anecdotes from your residency time at MGH that you wanted to share? Now we can skip to. Oh no, so many. But I got to pick a politically correct one. Sure. So one of the ones that I'll share comes from when I was on the wards. My attending was Jeremy Schmommen at MGH. Brilliant. Brilliant. Brilliant clinician and researcher. And I still remember we had seen a stroke in the emergency room. It was somebody with a thalamic lesion as I recall. 25:00And as a neurologist, you go there, you do all your testing, it takes you 20 minutes or 40 minutes depending on how comprehensive you want to be to figure out do they have a stroke, where might the stroke be as you're waiting for the imaging to be collected. And I still remember Jeremy Schmommen comes down, pulls back the curtain, walks in, introduces himself. The patient reaches out, shakes his hand, introduces himself, and he turns around and leaves the room. And we're all kind of looking like, well, where are you going? He's like, oh, he's got a lesion in his left alabum. This is a clumsy hand dysarthria syndrome. And literally within five or 10 seconds, he had diagnosed the syndrome, the exact location of the lesion, the patient's entire problem. And it just kind of blew me away of, wow, if I could ever be a neurologist like that. Right? So that's just one anecdote that I remember. That's great. It reminds me of the video challenges they sometimes have at the Movement Disorders Congress where sometimes you see these cases where patients with movement disorders were not 26:03diagnosed for 10 years, but it takes one guy looking at the video for 20 seconds and they know what it is. Right? So that is sometimes super impressive for clinical practice. I totally agree. Great. After residency, you joined the BIDMC as an instructor. Yeah. And there, of course, Alvaro Pascual Leone was your postdoc mentor. And then one key method I think you developed in that time is called lesion network mapping. And you already tapped into it just now, but could you summarize a bit more generally what it is, what the idea is about? Yeah. And I guess I just first want to mention, you know, before lesion network mapping actually was this idea of TMS network mapping. Okay. And then after that, you know, we went over to Beth Israel to work with and learn from Alvaro with the idea that, you know, with functional connectivity coming out of WashU, we had this tool for mapping out how different brain areas were related to one another. 27:00And, you know, I was definitely a tool builder in search of an app. Yeah. And so I spent my entire neurology residency thinking through, okay, we've got this great imaging technique that can tell us how different brain regions are related to one another. Where's that going to be useful? Yeah. And at the time, one of the big clinical problems was that we didn't have a single brain. Right. And one of the big clinical problems was where do you hold a TMS coil to help something like depression? Yeah. And there was evidence that TMS was working. Alvaro had done great work, you know, along with Mark George and many others, showing that TMS could help depression. But it was administered using scalp measurements. And we said, all right, well, we know that TMS is having an effect not just at the stim site, but everything connected to that stim site. And it was kind of this perfect combination of, hey, I know a little bit about resting state functional connectivity. You know a lot about TMS. Yeah. Let's get together and see if this, you know, brain connectivity information can be useful for improving TMS therapy. So that was actually the reason I went to Beth Israel. Makes sense. Yeah. Absolutely. 28:00And the reason I highlight that is because the lesion network mapping emerged fully from, you know, patient motivation and some degree of serendipity, right? Where, you know, Aaron Boes also came over to Beth Israel as a fellow to work with Alvaro and work with me. So, again, for TMS network mapping, right? How do we combine brain connectivity to improve TMS therapies for... Maybe let's briefly talk about TMS network mapping before we go into the lesion. Yeah. Go into the lesion. So, was... So, the idea there was that you want to hold the coil over a spot that's anticorrect, again, to the subgenual. And I think that's your 2012 biopsych paper. Was that motivated by the DBS trials of subgenual or...? 100%. Yeah. Okay. So, you know, we had a clinical problem of where do we hold our TMS coil to try and improve depression. But then we needed another angle on it, another hook. And so, Helen Mayberg's work showing how critical this limbic region, you know, the subcolosal 29:03cingulate was to depression was very influential. And at the time, you know, she had recently published her papers, you know, motivating implantation of DBS electrodes into that site. And I think you're hitting on something that is an important trend or focus of mine, which is how do you relate these different neuromodulation interventions to one another? I said, all right, it can't be a coincidence that we have a DBS target for depression and we have a TMS target for depression. And you can administer neuromodulation to either of these targets and make depression better. They've got to be related to one another. Yeah. And so, it was Helen's work that motivated us to say, all right, we think the subcolosal cingulate is important for depression. We know that somewhere in the left prefrontal cortex is important for depression. How do they line up? Yeah. And so, the 2012 paper was just a test of that hypothesis of can we look at connectivity with this critical limbic region and does that help identify topography in the frontal cortex that lines up with different TMS sites that seem like they're working or not working. 30:05Super cool. And it seemed to work. And then I think one other landmark paper was your 2014 paper in PNAS where you were showing, again, more conceptually, but across, I think, 10 or 12 diseases that the DBS sites and TMS sites would also, again, line up in the same way, right? So that is another really cool paper. Do you want to briefly? Yeah. It was just the same idea. Yeah. And again, it was, I had the privilege of, you know, talking to Alvaro Pascaleon every day and then he introduced me to Andres Lozano, one of the world's experts in DBS. And I had, you know, training in movement disorders and did DBS every day for Parkinson's disease. And so it was just this idea that, you know, we have DBS therapies that are effective for certain symptoms. We have TMS therapies that are effective for certain symptoms. And we've got an imaging tool that can link up different regions into networks. And can we test the hypothesis that our DBS sites that are working and our TMS sites that 31:01are working are part of the same network? And if that's true, then maybe we need to be thinking about our targets for neuromodulation, as brain network or brain circuit targets that are independent of whether you're targeting a subcortical node with DBS or a cortical node with TMS. In the end, it's a circuit target. So it was just, it was more of an idea paper. In fact, we originally tried to publish that paper as just a perspective, right? As a hypothesis or an idea. And it got rejected everywhere. And so then we tried to turn it into a quote unquote research paper. But again, it got accepted and it got attention, not because the science in that paper is particularly It's not good. It's not good. It's not good. It's not good. It's not good. It's not good. It's not good. It's not good. It's not good or rigorous. It's a concept that I think has proven to be useful as time's gone on. Going into lesions now, finally. We will get there eventually. Aaron Bowes came along. What did he find? He had a patient, right? Yeah. So it all started with a patient. 32:00So he literally saw a patient when he was a fellow over at Mass General in his pediatric neurology clinic. And this was a 16-year-old girl that came in with a peduncular hallucinosis syndrome. So she was driving her car, all of a sudden was looking out her window and boom, the whole countryside looked like it had been drawn in by crayon, right? And as Aaron tells the story, and I think he actually wrote up her case report, she pulled her car over to the side of the road, looked at the passenger seat sitting next to her, and all of a sudden a bunch of flowers sprouted up from the passenger seat. And then she reached over to try and touch the flowers and they all fell over, right? Wow. Wow. Wow. Wow. Wow. Wow. Wow. Wow. Wow. Wow. Wow. Wow. Wow. Wow. Wow. Wow. Wow. Wow. Wow. Wow. Wow. Wow. Wow. Wow. Wow. Adam Adam Adam Adam Adam Adam Adam Adam Adam Adam Adam Adam Adam Adam Adam Adam Adam Adam Adam Adam Adam Adam Adam Adam Adam Adam Adam Adam Adam Adam Adam Adam Adam Adam Adam Adam Adam Adam Adam Adam Adam Adam Adam Adam Adam Adam Adam Adam Adam Adam Adam Adam Adam Adam Adam Adam Adam Adam Adam Adam Adam Adam Adam Adam Adam Adam Adam Adam Adam Adam Adam Adam Adam Adam Adam Adam Adam Adam Adam Adam Adam Adam Adam Adam Adam 33:03Adam Adam Adam Adam Adam Adam Adam Adam Adam Adam Adam Adam Adam Adam Adam Adam Adam Adam Adam Adam Adam Adam Adam Adam Adam Adam Adam Adam Adam Adam Adam Adam I understand that we see this and we've seen this for 100 years, but I still don't understand why a lesion in these subcortical locations would be causing visual hallucinations. And so, you know, he dug at it, dug at it, dug at it, and just wasn't satisfied with the answer. And he had a very nice paper ready to go, just, you know, kind of a comprehensive analysis of all the lesions that he could collate in the literature that had ever caused this syndrome. And then he just happened to knock on my door and say, hey, Mike, I hear you do some kind of brain connectivity stuff. Do you think that would be useful for understanding this patient's lesion? And that critical question, right, that Aaron asked ended up launching this entire technique because, you know, Aaron was onto something. He had the right idea. And it, I think, has proven useful for understanding peduncular hallucinosis, but for linking up lesions that cause a wide variety of different 34:01neurological and psychiatric symptoms. Super. And we forgot to summarize what the general principle is. So you have lesions causing the same symptom, and they don't overlap always, right? They can be always, like, everywhere in the brain. Yeah. And then, yeah. Yeah. So I'd say there's two problems that Aaron ran into, right, that lesion network mapping was able to help solve. So one problem is the problem of heterogeneity in lesion location, where it's very, very common that lesions that cause a similar syndrome don't intersect the same brain region. Yeah. So in peduncular hallucinosis, all the lesions are in the brainstem and the thalamus, but they can be in lots of different locations in the brainstem and thalamus. And so when Aaron looked at all these different lesion cases, there was a lot of heterogeneity. They didn't all intersect one spot. So that was problem number one, was heterogeneity in lesion location. But problem number two was even if all the lesions did intersect one spot, let's say in the medial thalamus, that still doesn't tell you how the 35:00lesions causing visual hallucinations. Not connected to a visual cortex. Exactly. And that gets to this diascesis phenomenon, right, that a lesion in any location has a functional effect on any brain region that's connected to that lesion location. So even if the lesions weren't heterogeneous, there would still be an extra step that could be needed to understand where the symptoms are coming from. And so lesion network mapping was the idea that, hey, we know that there's heterogeneity in lesion location. We know that there are diascesis effects of the lesion. And so we know that there's heterogeneity in lesion location. And so to get to these these these these these these these these these these these these these these these these these these these these these these these these these these these these these these these these these these these these these these these these these these these these these these these these these these these these these these these these these these these these these these these these these these these these these these these these these these these these these these these these these these these these these these these these these these these these these these these these these these these these these these these these these these these these these these these these these these these these these these these these these these these these these these these these these these these these these these these these these these these these these these these these these these these these these these these these these these these these these these these these these these these these these these these these these these these these these then turn to a normative wiring diagram of the human brain. And it's important to emphasize that. We never scanned any of these patients with functional connectivity to see what the lesion was connected to. And in fact, if we had scanned the patient, it wouldn't have worked. 36:02Because the tissue at the lesion location is dead. It's not connected to anything anymore. So you go to a normative wiring diagram of the human brain collected from 1,000 healthy individuals. And it's an atlas that tells you what is this lesion location connected to. And you do that for each lesion location. And that allows you to say, hey, do these lesions in different spots all line up? Are they all connected to anything in common? And then two, what is that? And it might be a brain region that's actually not lesioned by any of those lesions because the symptom might emerge from a diascesis-like effect. Super cool. And that shows they all, so you find the common denominator. And then in that first paper already, you mentioned that the antigen. Anticorrelation seemed to be important because they would connect to the spot in the visual cortex. So all the lesions were anticorrelated to that spot. Is that correct? Correct. Yeah. Very nice. So what were maybe general challenges developing this method further? 37:07Do you have any, or also wins from the method? So what followed after that? Yeah. So I think it. So the wins were easy. So it just, it sort of became a, you know, so Aaron, after he got his win with peduncular hallucinosis, I think we submitted that paper to a couple of different places and it got rejected. Okay. And you know, people weren't interested in just peduncular hallucinosis, right? Neurologists are interested in that, especially really nerdy neurologists like us, right? But, but, but they were interested in, in could this be a, a, a technique that holds true more generally? Sure. So even in that first paper to get it accepted, we had to go through and show, Hey, this isn't just a unique phenomenon that's only relevant to the peduncular hallucinosis that it might also be useful for subcortical aphasia, um, for pain, for auditory hallucinations, right? 38:02And, and Aaron put all those together and said, Hey, look, this, this isn't just a one off. This could be a general principle for how we understand strokes and focal effects of lesions and network effects of lesions. Um, and, and Aaron was right. Um, and so after he did that initial paper, it was almost off to the races and then it just became any fellow or resident that knocked on my door. You know, I was like, all right, what are you interested in? Yeah. You know, I'm interested in hemichorea. All right, go for it. Yeah. Map out lesions. And so it, and then I started to get, um, fellows knocking on my door with totally crazy ideas and, you know, Ryan Darby knocking on my door, be like, you know, I'm interested in Capgras syndrome. Uh, you know, can it sell the Capgras syndrome? I was like, well, probably. Probably not, but give it a try. And, um, and then it just, um, it was just shocking where it just kept working. Um, and it wasn't like, you know, it worked half the time, you know, we don't have a whole bunch of skeletons in our closet where it's, oh, you know, we tried it all these different syndromes and we got absolutely nothing. 39:03Um, it, it, it just kept working. And so I think it, it, Aaron just stumbled upon a fundamental principle of how the brain works and how symptoms localize. And they just, they don't localize to regions. They just localize really well to networks. And this resting state connectivity, because it's polysynaptic, because it, it reveals functional relationships ended up being a good way to, to map these things out. And then I think the challenges, um, that you asked about, you know, anytime there's a new method, it gets to this debate of the, the, you know, the processing algorithm for resting state data. Sure. You know, there are many different ways you can do this. Um, you know, do you, do you take the. Take sites of lesion intersection and run that as a seed, or do you run each lesion location as a seed? Um, and, and actually it was a debate Aaron Bose and I had for that first paper and it remains an ongoing debate. Okay. Um, and some of his recent papers, he's done things that way. Um, there's a debate of, you know, is functional connectivity the way to do it or is anatomical connectivity a more useful way to map lesions onto, onto networks and onto circuits? 40:08And, um, you know, are there different symptoms where anatomical connectivity might be better than functional connectivity? So I, I think. There's a lot of ongoing good methodological debate and development to figure out how we optimize the technique. But I think the general principle that lesions causing the same symptom map onto a network or a circuit, um, is, is holding true. Seems to work. Yeah, definitely. So how, how did it feel once you realize this is probably took a few years, right? Get the first papers out, but at some point you realized, Hey, this really seems to work. What was that feeling like? Probably was it wasn't it? Yeah. There was never a moment, right? It's, it's cumulative. And, and, and I wouldn't say, you know, even now I'd say, Oh, it, it, you know, I'm now confident it works. Right. I keep looking for that example where it fails. Right. And, and, um, you know, I, I would say that there were definitely moments where, you know, I, I kind of couldn't believe that it was working. 41:08Right. You know, like when Ryan Darby first showed that lesions that cause delusions of familiarity. Right. Are connected to the familiarity detector and they're all connected to our reality monitor. And then like all of a sudden it was like, Oh, for, for a century, people have hypothesized this is a dual hit syndrome. Like the lesion must be doing two things and disrupting two different processes to cause this very, you know, amazing delusion of familiarity. And it was like, okay, sure enough, the lesions connected to two different things and two different networks. And, and. It's too good to be true. Yeah, yeah, exactly. It's, it's one of those things where it. It just makes, I, I don't know. I, I do believe in science that things that are true feel simple in retrospect. Yeah. And, and again, it just was like, it made so much sense. Sure. Yeah. Yeah. And, and so I, I do think that each time we study a new syndrome and a, and a complex syndrome and things all of a sudden make sense, it, it, you know, it just. 42:06Yeah. Continues to amaze me. Right. That. Absolutely. That, you know, we, we, we have a tool that can even begin to study some of these more complex things. Yeah. You know, we have papers on free will and criminality. Yeah. Yeah. Just things that, religion. Yeah. Right. That, that you just, you can barely even make these topics accessible to neuroscience in general with any technique. Yeah. And the idea that I, you know, happen to have a tool that allows me to address these fascinating questions that go beyond, you know, neuroscience or direct patient care is super fun. Super cool. Yeah. Very, very nice. I, I remember loosely that you were at the time in this beginning phase collaborating strongly with, you know, methods, people like Heshing Liu, Randy Buckner, Tom Yeo. Was that, was that important to get the, the setup and the, you know, the connectome running or would you, would you think in general your network mattered a lot at that point to, to, to get this, this running? 43:01Oh, 100%. Yeah. No. I, I, you know, one of my general principles is to try to always find something that works. Yeah. And always find someone smarter than me. And then partner with them and collaborate with them. Yeah. And, and certainly at that period in time that, you know, group that you mentioned was so helpful in so many ways where, you know, I had just come in from WashU. I was spending most of my time being a resident. That's right. And so I only had time for science on the nights and weekends. And I certainly didn't have time to, you know, rewrite all the WashU code to work with, you know, the system in Boston. Yeah. Right. And so just the, even the, the format of the imaging files and, and the, you know, the resting state data sets and, you know, all that was different. Sure. M and I space was not a thing at WashU. Oh, really? Right. MIFTI files were not a thing at WashU. I knew. Yeah. And luckily Randy Buckner had moved from WashU to Boston just about five years or so prior to me. And it already solved all these problems. Right. So having Randy as a mentor, I, I mean, I, I would have been, he, he basically allowed me to accept the fact that I was a student. 44:00Yeah. I mean, I, I would have been, he, he basically allowed me to accelerate my entire research portfolio, you know, by five to 10 years because he had already solved so many of these problems. And then, you know, in, in his lab and collaborating with him were Thomas Yeo and Hsien-Liu and they were very generous with their time and their expertise to, to help get some of these things working. And so it was really the right group of people at the right time that, that helped me tremendously. So you mapped since then quite a bunch of data. Got to get to these standards to get to these standards to get to these standards to get to these standards to get to these standards to get to these standards to get to these standards to get to these standards to get to these standards to get to these standards to get to these standards to get to these standards to get to these standards to get to these standards to get to these standards to get to these standards to get to these standards to get to these standards to get to these standards to get to these standards to get to these standards to get to these standards to get to these standards to get to these standards to get to these standards to get to these standards to get to these standards to get to these standards to get to these standards to get to these standards to get to these standards to get to these standards to get to these standards to get to these standards to get to these standards to get to these standards to get to these standards to get to these standards to get to these standards to get to these standards to get to these standards to get to these standards to get to these standards to get to these standards 45:00Autophagy, autoscopic, oh, my German accent is too terrible, blindside, bodiless awareness disorders, bodily self-confidence, and so on. Coma, criminality, central hypoventilation syndrome. Alphabetical. Exactly. I wanted to make the point. I just made it to the third letter, and I even skipped a few now. So lots of symptoms have been mapped. Why do you think is the technique so powerful and also now popular? So I'd like to believe that it works so well because we stumbled upon a fundamental principle of how lesions cause symptoms, and we just happened to have a tool at our disposal that is very, very useful for understanding lesions and lesion effects and relating lesions to one another. So I want to believe that that is why it has worked across so many different neurological and psychiatric symptoms. Right? So I would say there's a skeptical side, which in a concern that I will continue 46:04to harbor because you've always got to be very skeptical of your own stuff. And the skeptical side is to say, oh, well, it has worked and it's gotten popular because it's super easy with a very low bar, right, where all you have to do is search for case reports of lesions causing fill in the blank, right, copy those lesion locations onto an atlas, take a publicly available connectome, right, which anybody can download. And then you just take the lesions and run them as seeds, right? So you can run a lesion network mapping study in, what, three months, six months? Yeah. And so it's a very low bar to go in and study an interesting symptom. Well, that creates a danger, right? Sure. And so it's going to gain popularity because it's easy to use and apply. Yeah. But it also comes with a lot of risk that, you know, because it's easy to use and easy to apply, maybe people have done 20 times these number of conditions. And then through publication bias, we just don't even know about the times where it doesn't work, right? 47:03You know, there are probably instances where people are applying the technique to things that it shouldn't be applied to. You know, and it's funny, reading that off that table actually reminds me of a review article I did with Mike Gracious early in the days of resting state functional connectivity, where we did this, you know, ridiculous table of all the disorders where resting state functionally, functional connectivity had been applied. Oh, interesting. And in that review article, we actually made the point that it's being applied to too many things, right? And that it can't be possible that all of these are, you know, disorders of resting state connectivity in the default mode network. And so, you know, I think it's great when a new technique is developed. I think it's great when it's widely applied. But when it does get applied that broadly, you do wonder, should it be applied that broadly? And is this just a fundamentally valuable way for mapping out different symptoms? 48:01Or is it being over-applied because it's just easy to do with a low barrier to entry? And I'd say time would tell. You know, I think, you know, we haven't yet done the prognostic prospective study of, do any of these lesion networks actually predict which lesion patients are going to develop these symptoms? And, you know, those studies are ongoing. Yeah. But I think that the prognostic value of lesion network mapping will be a real, important test of does this hold water? And then there's a lot of therapeutic targets or possible therapeutic targets that have come out of these lesion network mapping studies. And those are all testable hypotheses. Yeah. That's the perfect segue into the next guest question, again, by Aaron Bowes. Let's hear that one. Of all the lesion network mapping studies that you've worked on, which do you think has the most promise for a novel therapeutic target? Yeah, great question. So, I'd say, you know, that one's, 49:01you know, the front runner in my mind is definitely addiction. And, you know, we've studied mostly lesions that cause a symptom, right? And you can map it onto a network. Yeah. But there's now a, you know, a conceptual leap to say, all right, that network for lesions that cause a symptom, is that the network you need to target to make the symptom better? Yeah. Because it might, it might be, but it might not be, right? It might be you have to target a different network that helps you compensate for that symptom. Yeah. But then a very small number of symptoms, we, we study lesions that get rid of a symptom. So, lesions that get rid of tremor map onto a brain circuit that exactly highlights the VIM nucleus of the thalamus, which is our DBS target for tremor. And then the next one that Yuhou Yutsu did was, you know, lesions that get rid of addiction. And because it's lesions that improve a symptom, you don't have that conceptual leap, right? Sure. Right? So, it's, this is the connectivity profile of a lesion that gets rid of addiction. 50:01Yeah. So, where are the spots in the brain that best match that connectivity profile? In theory, that's where you should put a lesion to get rid of addiction. Yeah. And so, you know, we have a spot near the anterior cingulate that comes out, which is a classic lesion target for addiction. Yeah. We have a spot in the anterior insula in the frontal operculum that comes out, that's a classic, well, the insula, I guess, is a lesion-based target for addiction. Nobody's actually lesioned the insula that I'm aware of. Yeah. But we do have a couple spots where I'd say if those spots fail, if you lesion those spots and they fail to help addiction, that would be a pretty direct challenge to the therapeutic value of lesion network mapping. Makes sense. Okay. And then, if you would want to go one step in between before lesioning, you could also apply non-invasive stimulation, for example, TMS and TDCS. Would you say the same approach? Would you say the same applies there, that the lesions that, you know, 51:01cause the loss of the symptom or improvement of the symptom would be better for that? Or would you say there is different because... Yeah. So, it's the challenge, and that's why I said if you put a lesion there, that's the best test for the hypothesis, right? Yeah. Sure. And the challenge with different non-invasive technologies is that they're not really a lesion, right? So, people talk about TMS as a virtual lesion, but it's really not. Yeah. Right? And even DBS, right? We think of DBS sort of like a lesion if you administer it at high frequency, but really that's just because, through trial and error, it kind of seemed like it acts like a lesion in certain movement disorders. Yeah. And so, it's very possible that it is the right therapeutic target, that a lesion could work really well at that location, but that TMS to that spot or focus ultrasound to that spot or DBS to that spot might not work. Yeah. Now, that might be the right next step, right? Sure. Certainly for, you know, lesion therapies, there needs to be a very high bar before you go in and start lesioning patients. 52:00Yeah. But that is scientifically the best test of the hypothesis. Yeah, yeah, yeah. Makes sense. That's very…makes a lot of sense. Great points. So, I think all of these symptoms or many of these symptoms led to an invited review that you wrote for the New England Journal of Medicine in 2018, which also shows, I think, the big impact and interest in this technique. From all these symptoms, and functions that you've studied, which…do you have a favorite one? Just, you know, personal favorite. Yeah, I'd have to think about that one if I had to choose a favorite. There's been a lot of very, very exciting moments as, you know, we saw different lesion network mapping results and kind of, you know, put our hands behind our heads and said, wow, that's really cool. Yeah, I think, you know, going back to that story I told earlier, like the day Ryan Darby, quote, unquote, solved the Capgras delusion, right? 53:00That one to me was…that was really cool. Super cool story. It's hard to know if it's right, but it makes a lot of sense. Let's go with that. So, one key point that you recently also highlighted together with Sean Siddiqui in a review paper is that, and you mentioned it before, that the lesions introduce a causal modification to the brain, right? You have a stroke and then something changes. It's a causal intervention. There are other causal interventions and namely, I would say, brain simulation, as you point out. So, deep brain stimulation and TMS and other forms of neuromodulation are other forms of causal intervention, and your lab has combined all three of these in some papers, sometimes two of these, in sources of information. Can you talk a little bit about that? What's the importance of these causal… Yeah, yeah, and so it's…I was…got to give a very nice line. I was at a lecture at the Organization for Human Brain Mapping a few years ago, and it was…I think it was, what is the message that I want to give to all the neuro-imagers out there? 54:02And it was very clear, which is, you know, we need to rethink neuroimaging. We've been doing neuroimaging for three or four decades, looking at, you know, imaging correlates of these symptoms we want to treat. But our track record of translating those imaging correlates of symptoms into newer effective treatments is not good. Yeah. And why is that? It might be, oh, we're just not there yet, and it has to evolve over time, and that's the standard answer. Or it might be that fundamentally there's something wrong with this approach. And I'm not the only one to worry about this, right? It's kind of emerged as this idea of the causality gap, where you can have an imaging correlate, but that doesn't mean that that's an effective therapeutic target. And so if we believe that that's a problem, then how do we close that causality gap? And one way to close it is, you know, through what I call the causality opportunity. It's to focus on causal sources of information for mapping out symptoms. And brain lesions is our original one. 55:00And in fact, brain lesions and these incidental observations of lesions doing different things led to many of our effective neuromodulation targets. And so lesion-based localization actually has a very good track record for turning into effective treatments compared to neuroimaging correlates. So it's all right, let's double down on the path that's working. And maybe, you know, deviate a little bit from the path that's popular. And so we focused all in on lesions as the dominant causal source of information, you know, followed by DBS electrode locations or TMS stimulation sites, with the idea that you can map things out based on one causal source of information, and that should be valuable. But if you can actually do convergent causal mapping, if you can map out the circuit based on lesions, map out the circuit based on DBS effects, and map out the circuit based on TMS effects, and these three causal sources of information all converge on the same brain circuit, 56:05well, now you've got a really strong case that, you know, this circuit is causally linked to this symptom in humans, and modulating this circuit causally leads to a change in that symptom. You know, that's as strong as you can get as your candidate therapy. So, you know, we've got a very strong case that this is actually a therapeutic target. And that's what Sean Siddiqui did recently for depression. Just wanted to say that's probably the canonical study so far, right? Can you summarize what he did in that study? Yeah. So he took, you know, five lesion datasets where they did depression scores after the stroke. And then he took, you know, roughly five DBS datasets and five TMS datasets. Where in the DBS datasets, they measured depression scores before and after DBS. TMS, same thing. And in each one of these cases, he mapped out what the lesion, the DBS site, or the TMS site was connected to. And then asked, okay, which connections co-vary with depression score or change in depression score? 57:02And so you can map out a lesion-based depression circuit, a DBS-based depression circuit, and a TMS-based depression circuit. And just ask, do these three things line up? And they did. They lined up really well. That's another one of those moments where I remember Sean popping those three maps up on the screen and just looking at me like, no, this can't be real. Wow. That's great. But yeah, no, I think it was the best and the most rigorous test yet of this idea that lesions and brain stimulation sites or TMS sites and DBS sites are all going to converge at the circuit level. And that resting state connectivity might be a useful tool for testing whether these things converge at the circuit level. And I mean, we can even say, I think in that study, that the DBS sites were not, all in depression, right? They were somewhere in epilepsy, somewhere in Parkinson's disease. So you even have some cross-disease symptom mapping in that paper as well. So I think it's a really landmark paper, really great work there. 58:05I think the next guest question, again by Aaron, will, you know, I don't know if you have a spontaneous answer to that. Maybe you have to think about it a bit, but we'll see. What do you envision lesion studies will look like 50 years from now? Yeah, it's a great question. You know, I think everything's moving towards big data and big lesion collections. And, you know, Aaron Bowes being at Iowa is just such a beautiful, beautiful thing because Iowa's got this rich history of really collecting lesion cases and studying and phenotyping lesion cases. And Aaron has kind of revitalized that at Iowa. And, you know, I think it's a really, really important thing. You know, you build big ends. You know, the discovery lesion cohort being run by Natalia Roast over at MGH, you know, the goal is to collect 8,000 lesions, 6,000 of which will be focal strokes, 59:01and, you know, and phenotyping all these patients. And so I think it'll move towards, you know, a lot of our lesion papers now are 30 lesions that cause this thing. Well, you know, 6,000 lesions, right, is a different level of power. Now, they won't be as clean phenotypically. One of the things that's been nice about lesion network mapping is these, these case reports of this really clean phenotype that somebody was motivated to write up have been just worth their weight in gold. So I'll answer it two ways. I think one is moving towards big data, huge lesion data sets where we've got depression scores, memory scores, you name it. And I think those efforts are happening. But I think the other effort that has to happen, and I'm not sure how to put some motivation behind it, is doubling down on these incidental cases. That just have these very clean, striking phenotypes. You know, where you see the case and it's like, wow, right? The lesion did that. And right now, I don't know that we have a great way to capture those very important and interesting lesion cases 01:00:05that are not, you know, subtle changes in depression score, but just that striking phenotype where the lesion either caused a severe depression in somebody with no history of psychiatric illness or got rid of depression in somebody with a very, very strong history of depression. And somebody with a big history of psychiatric illness. And right now, those cases are happening. I'm convinced that they're happening. And if you could find a way to collate the world's, you know, database of stroke, you know, all the stroke cases in China and India and all the different institutions in the U.S., I think if you could capitalize on that in some way, you would be leveraging serendipity towards therapeutic targets at scale. Yeah. And so I'd say that's the other thing that I'd like to see, happen in the next 50 years of lesion studies. Sounds great. I very much share that vision. Very cool. I think the podcast, as you know, is about deep brain stimulation, mainly, or brain stimulation. So we've talked a lot about lesions, and they have a great value for informing both invasive and noninvasive brain stimulation targets. 01:01:05But you also clinically and scientifically are very much involved in deep brain stimulation. So are there maybe, in parallel to what Aaron asked, any key highlights you foresee for the future in that field? Just in DBS or TMS? Both of them, but yeah, maybe DBS. Yeah, it's a great question. So it's an area of rapid development. Your podcast has covered those developments across brilliant speed. By the way, I'll take a moment to say I love this podcast. And I love all the speakers you've had. And it is an absolute privilege to be interviewed, given the set of people you've interviewed so far. So you've well documented a lot of different visions of where DBS is going. And I'm trying to think if I have anything actually useful to add or complement what all the other speakers have said about DBS. 01:02:01My follow-up question would be, what do you think about closed-loop deep brain stimulation? You go on and talk about it. And I'm excited about it. There's an electrical engineer. It's super-duper cool, right? I think I'll mention a couple of things to complement what other people on this podcast have said previously. So one is there's a lot of development in the DBS area that I think is driven by very cool engineering and the ability to do cool things with our engineering, and maybe a little bit less by patient need. I believe fundamentally in the clinician-scientist model. And doing DBS clinically really tells me what the problems are. And the problems are balance. The problems are freezing. The problems are speech side effects. And some of those problems can be addressed with closed-loop or adaptive design, 01:03:02but some of them maybe not so much. And so I think for me, I spend a lot of time thinking about, you know, stimulation site and symptom-specific circuits and side-effect circuits. And so I think one is can we change the location of stimulation slightly to better avoid a side effect like cognitive decline and STM, DBS, or Parkinson's disease, or maybe even have some benefit on gait or balance or freezing. Now, not with our traditional target that improves tremor or improves rigidity. It's going to be a different target, but it doesn't mean that there's not some circuit that we could reach with, you know, flipping on a different contact that could have a symptomatic benefit on some of these things that right now we feel like can't respond to DBS. Or even the psychiatric side effects. You know, it's very common that we'll have a patient with psychiatric issues, you know, depression, anxiety, and they want to know, 01:04:00can you do anything about this with my DBS electrode? And right now we say no, but that doesn't mean we can't. That just means we don't know how to do it yet. So I do think it's moving towards this idea of symptom-specific circuits, of reprogramming trials, of flipping on other contacts in different ways that might help different symptoms. And then the other one that I'll mention, just because you already know what I think about closed-loop, but I like being controversial and telling people that one future of DBS is extremely cool engineering and fancy devices and closed-loop designs with sensing all over the brain and stimulating in different areas of the brain. And the other future of DBS is just burning on fire. And I think that's the key to the whole and going back to lesions. And that again just came from my patients, where as focused ultrasound became an option, I would see patients over and over again where they would, you know, see somebody for a focused ultrasound intervention, they'd see me for DBS. Both of us would tell them that DBS is better and it can give you better tremor benefit with fewer side effects. 01:05:01And time and time again, the patients were opting for the lesion. And it just kind of blew my mind. And then it suddenly dawned on me that some of these patients don't really want a fancy piece of engineering in their head if they can avoid it. And the idea that they could do one and done might be preferable for certain patients, certain symptoms, certain conditions, especially as we start moving into, you know, psychiatric symptoms. So, you know, my controversial take on it is that, you know, DBS is just a stepping stone to figure out exactly where those symptom-specific circuits are and the exact spot in the brain that we need to modulate without side effects. But once you know that exact spot, you know, burn in a small hole there. If you can be confident that it has the effect, it might actually be the future. And I guess it's two things, right? It's knowing exactly where it's a lesion but also being able to ask precisely as possible the lesion there, right? And I love the idea just, you know, the idea that the MR-guided focused ultrasound 01:06:00happens in the MRI because even if it's not perfect yet, I think that technology will have the possibility to, you know, the possibility to see exactly where you lesion, right? In anatomical coordinates and so on. So, some more. So, I know that you're working on the concept of the return of the lesion. And I think in that you propose that after lesions had a break over history, to some degree, they are now celebrating a comeback. And not only for therapeutics but also for diagnostics. The broader principle that I think is important for me and maybe for your listeners is, where have our successes come from, right? And, you know, right now it's, you know, we're trying to build a center, right? And the whole goal is new treatments, right? And academia, you know, you're motivated by papers, you're motivated by grants. But, you know, the goal really needs to be treatments, right? And so, it's spending a lot of time thinking about where have our successes come from, where have our wins come from, and how do we learn from that, leverage it, improve upon it. 01:07:01And lesions have such a strong history. Yeah. Yeah. And such a strong track record that, yeah, going back to lesion-based localization has value. And then, you know, and again, it's controversial, right? But to say that lesions might be a future neuromodulation intervention is worth thinking about. It's what is going to be that disruptive technology. It also, I think, hits on one other point that's useful to make, which is, I've spent a lot of time in my career watching where the herd is going. And intentionally going in a different direction. And the reason is that if the herd's already going there, the world doesn't need me in that realm. And so, you mentioned, you know, closed loop and adaptive DBS. The people working on that are some of the smartest people that I've ever met, right? You go to DBS think tank, you're just blown away by what people are doing with the engineering and the electrophysiology. And so, they don't need me there, right? 01:08:00Like, I'm not smart enough to compete in that space. Sure. And so, it's okay, what are they not doing or not thinking about? Or, you know, when I came out of residency, you know, I went into residency being someone really focused on resting state connectivity. Well, I came out of residency and there was a whole army of people working on developing resting state connectivity. They didn't need me there. And so, you know, right now with therapeutics, it's the same way, right? If adaptive and closed loop designs are going to get us there, we're going to get there. Yeah. So, the only thing is what is the area that's not being explored where I can potentially add value? And I like it as a principle because the more heterogeneity we have in science, the more people pursuing different ideas, different options, crazy avenues, the faster we actually get to effective treatments. Yeah. And that's everybody's goal. Sounds great. Great concept and teaching point. So, the serendipity story was actually the next question I wanted to pick your brain on. Yeah. 01:09:00So, I think you already summarized it, but I think it's worth to mention a bit further that I think it's motivated by Marwan Hariz' recent, or at least that was one paper that you really liked in that realm, where he made the point that looking back at history, there's not so many wins as one would think from, you know, coming from the animal research model where we usually think, you know, we have a mouse model, we study something in the brain, and then we derive a therapeutic answer. Yeah. And that's something that we actually target from that for neuroscience. And looking back, there's just not so many wins in that. There are some, but, you know, not so many. But there are a lot of wins in the, you know, just looking at case reports or looking at variations and interventions. Do you have thoughts on how we could capitalize on that further? Yeah. Yeah. And this was the, you know, think tank, right? Yeah. I kind of propose that. And yeah. So, serendipity and the role of serendipity is something that I've loved for a long time. I was just so excited to see Marwan. Yeah. 01:10:00I was so excited to see Marwan's article. Yeah. Where he just, he did the hard work of putting together the history, which I don't know, right? I'm not a historian. It's actually one of the reasons I love the podcast so much is I get the history that I wish I would have gotten from having all of these people as my mentor, right? That they could tell me the history. So, it's trying to learn from that history. And, you know, one of the things that you've pointed out is, yeah, okay, functional neuroimaging correlates don't have a great track record. They don't have a great track record of turning into new treatments. Mouse models of human brain diseases don't have a great track record of translating into new treatments. So, what does have a good track record? Or are we just doomed to failure no matter what avenue we pursue? And again, that's the trial and error in human patients, right? And it sounds brutal and it sounds very non-scientific and maybe even, you know, non-ethical, but that is the history of our wins. That's where our successful treatments came from. You know, it was the insulin. It was the insulin observation of, you know, James Parkinson's patient number 6 had a stroke and the tremor stopped. 01:11:03Yeah. So, what did we do? We started lesioning and creating strokes all over the place, right, trying to get tremor to stop, right? And eventually, you know, we had trial and error success where, you know, people went in, tried to lesion one thing, had a bleed somewhere else. Suddenly, somebody's tremor stopped and they weren't paralyzed. Okay, huge win. Yeah. But it was these trial and error... Yeah. interventions on human patients with careful observers as to what was working and what was not working, taking full advantage of the heterogeneity in their outcomes, and then refining their treatment over time until they converged on something that worked with minimal side effects. And so that pathway that worked very, very well in movement disorders, how do you apply that elsewhere? And so I look at, we were just on the paper on addiction. So there were a bunch of lesion treatments for addiction happening in China, happening in India, 01:12:00right? And they got shut down, and maybe for very good reasons. I don't know the history, but there were ethical concerns, there were side effect concerns. But those lesions are now all lost to the dustbin of history. And I know I tried to get them, right? But they weren't recorded in a systematic way. People didn't look at the heterogeneity, what worked, what didn't work. Oh, here's an incidental case where I actually was aiming for the anterior cingulate and I missed, and I got this other spot. Well, did it work better? Did it work worse? And so I think you had the people developing therapies for movement disorders that were leveraging the heterogeneity in their outcomes. They were embracing serendipity. They were studying it, right? And they had lesion therapies in probably the same number of cases for addiction where they weren't taking that same approach, or at least not in a way that I've been able to access. And so to me, you can learn from that history, learn from where the winds came from. And when I talk on, you know, how do we go all, you know, how do we get in on leveraging serendipity, which was a controversial question to even pose at Think Tank, because by definition, you can't plan for or leverage serendipity, but maybe you can. 01:13:03Yeah. And so, yeah, it's... And you had some ideas on how to... Yeah. Or just, you know, it's any time you intervene on a human brain, right? Especially with some type of causal intervention, whether that be a lesion, a DBS electrode, a TMS site, right? Yeah. You know, I listened to Josh Gordon's interview, right? And the NIMH has done a really good job of essentially mandating certain things, and it moves the entire field, right? So they mandated that they want some imaging biomarker of any intervention for a mental health disorder, right? Well, what if you mandated that any DBS trial has to publish their electrode locations? Any TMS trial has to record the exact site of stimulation, right? That's a way that you plan for serendipity and you leverage it, right? Because now you start building... Yeah. You know, here's what worked, here's what didn't work. Here's had this unusual effect or didn't, and then people can mine that and begin to back out and converge on, you know, a target 01:14:04that might be even more effective... Yeah. ...that was just because you got the electrode in the wrong spot. Yeah. Love the idea, right? So to make that mandatory, that would be great for everybody, right? And I think everybody would love it, right? If everybody shares their sites, you know, that would be super, yeah. I think the other thing you once mentioned would be a... Yeah. ...you know, if we have a really informative case report, it's usually there's no good journal for that, right? Or it's not, you know, valued as much as a paper, so that could be another option, I think, yeah. Well, I mean, you know, as we were brainstorming the other day, it would be amazing to have a, you know, registry of, you know, causal case reports. So a lesion, a DBS stimulation site, a TMS site, right, where you have the causal site of intervention and then you have the interesting behavioral phenomena. Yeah. And you know, you could create, you know, maybe publishing each one of these case reports in a paid, you know, journal is not the way forward. 01:15:02Yeah. But if you had an online registry of these cases that became a, you know, PubMed indexed or citable element where, you know, now residents can just submit their interesting cases, right? Yeah. And eventually you start accumulating enough cases that people can make sense of different lesion locations that did something or DBS sites that did something. Could be a cool model. Love the idea. And so on. So if you have a... Yeah. Yeah. ...you know, a... Yeah. ...you know, a... ...you know, a... ...you know, a... ...and so on. So I have one more question in that direction before we move on to some more general questions by somebody in Australia. So... Hi, Mike and Andy. Mike, my question is, who is your favorite Australian postdoc of all time? Don't answer that if you've had another one other than me. No, my real question is, what do you think of companies like Neuralink? I'm sure that you've been asked this a lot, but I've never heard your thoughts on it. Do you think that these companies are the future and that they can really help people with brain disorders? Or is the human brain just too complex? Thank you, and I hope to 01:16:05see you both soon in person. So thanks, Dan Kaur. Great to hear from you. And then with the thick Australian accent, I'd also want to know what the company Neuralink is. Is that a Popsicle company or like... No, no. So, you know, the question about Neuralink, you know, I think it's a really good question. Yeah. Yeah. Yeah. Yeah. Yeah. Yeah. Yeah. Yeah. Yeah. Yeah. Yeah. Yeah. Yeah. Yeah. Yeah. Yeah. Yeah. Yeah. Yeah. Yeah. Yeah. And you know, is it a good thing? Is it a bad thing? You know, I think any investment of resources to a, you know, to an important field is a good thing. Yeah. And you know, that company, you know, with Elon Musk brought in a huge amount of money into the field, a huge amount of PR and interest in the field. And that's almost always a good thing. It'll spur development. It'll spur other companies to be interested. VC will be interested. Other scientists, you know, that are thinking about what they want to do with their life will get interested. And so I think it's a good thing overall. 01:17:02Now, you know, whether or not neuromodulation to enhance function, which is, I don't know, Neuralink has changed their PR over time. They're right now focused on treating brain disease. But at one point, I think that they were focused on neuroenhancement. And they're not alone, right? There are a lot of other neuromodulation device companies out there that are focused on neuroenhancement. And, you know, that one. One, I'm not sure whether it's really possible or not. You know, millions of years of evolution have tuned up our brains to work pretty well. Yeah. And my guess is that we can enhance certain functions. It'll be at the detriment of others. And the idea would be you enhance certain things at certain times where it's advantageous to you and then enhance other things at other times. So I guess one is just is it even possible to do neuroenhancement? I'd say maybe, but it's not an easy bar to overcome. 01:18:01Most of our brains are. Pretty good at what they need to do. And there's always a risk of unexpected consequences of trying to make something that works pretty well better. Yeah. But let's just say we can overcome that hurdle and neuroenhancement becomes a thing. You know, I'm not fundamentally against it. It just comes with a large set of other problems that will then need to be solved of, you know, equal access. And how do you prevent the smart from getting smarter? Yeah. Right. And then other people that don't have access to that technology fall even further behind. Yeah. But in general, you don't really stop technical advancement. Yeah. You know, I'd say the only area where we think it's actually been paused is, you know, the field of human cloning and human genetic engineering, where technically we have the ability to do it. And for ethical reasons, it really has put a pause on that area of science. You know, but, you know, neuroenhancement, despite the ethical concerns. 01:19:02Yeah. Yeah. So, I'm going to move to some more general questions. And one is, yeah, hard to fit into. But I think it's a great question by Shan. I am going to play that now. And my second question is, who do you think would win if all of the people you've ever mentored were in a Royal Rumble fight to the death? Obviously, you can't choose Joe Taylor because he's the best fighter out of all of us. He would obviously win. So, yeah, he can't be in this. And I know the right answer. 01:20:01I'm going to give you a second. I'm going to think about it. Oh, my goodness. I continue to find out that, you know, people that I am mentoring are very skilled in the martial arts in ways that I am not aware of. And, you know, and so now I'm trying to figure out he's clearly thinking of someone other than Joe. And so that means there's someone else that I've mentored that's very, very skilled in martial arts. And I have to think of who it is now. It's not that. It's not that. Oh, it's not that. Okay. Well, then I don't know. Who do you pick? He wants to know who you would pick. Who would win? Who would win in a Royal Rumble that everybody I've ever mentored? And I can't choose Joe Taylor. For anybody who hasn't met Joe is very large and very skilled in martial arts, I believe. Let's see. I think I didn't mentor him. I worked collaboratively with him. Hsien Liu was actually a kickboxing champion in China. Oh, wow. Yeah. 01:21:00And so, actually, Hsien. I think I would probably put my money on him. Especially if I can't put it on Joe. But I want to hear what Sean thinks. He clearly thinks he knows the answer. Let's do that. Yeah. Obviously, the right answer is Alex Cohen because you'd figure out a way to turn everybody against each other. Maybe some clever algorithm or something so that he can step back and wait till just the right time to step in when everybody else has damaged each other to the point of no return. But I'm curious to hear what you're wrong in. I'm not sure. I'm not sure. I'm not sure. I'm not sure. I'm not sure. I'm not sure. I'm not sure. I'm not sure. I'm not sure. I'm not sure. I'm not sure. I'm not sure. Adam Dewie to Adam Dewie to Adam Dewie to Adam Dewie to Adam Dewie to Adam Dewie to Adam Dewie to Adam Dewie to Adam Dewie to Adam Dewie to Adam Dewie to Adam Dewie to Adam Dewie to Adam Dewie to Adam Dewie to Adam Dewie to Adam Dewie to Adam Dewie to Adam Dewie to Adam Dewie to Adam Dewie to Adam Dewie to Adam Dewie to Adam Dewie to Adam Dewie to Adam Dewie to Adam Dewie to Adam Dewie to Adam Dewie to Adam Dewie to Adam Dewie to Adam Dewie to Adam Dewie to Adam Dewie to Adam Dewie to Adam Dewie to Adam Dewie to Adam Dewie to Adam Dewie to Adam Dewie because he's so good at working with everybody else. 01:22:00Great. Speaking about general scientific questions, I always ask about eureka moments in science, and I think we have already covered a few of them. But even maybe thinking back to your earlier days, do you have any other great moments to share? It could also be just wins or lucky streaks or whatever. Yeah, there's been a lot of them. I absolutely love this job, and I love being able to do the programming and write a new algorithm. And I still remember the first time that we threw in global signal regression into the resting state connectivity algorithm. After pounding on it for a year and a half, all of a sudden these maps popped out, and it was just like, oh my God, that's beautiful. It's so much prettier than anything we've seen. I still remember how exciting it was that night. It was probably 2 in the morning with Avi Snyder. And I still remember sitting there with Aaron Bowes the first time we looked at the connectivity of these lesions causing hallucinations. Like, oh my gosh, this could be really important. 01:23:05And so there's been a lot of them. And I continue to have one. It was a couple weeks ago where we sat down and it was lesions causing anorexia. And I was like, oh my gosh, that makes a lot of sense. So look for that one to come out. But no, there's definitely been a lot of times where you... It's like, wow, that's super cool. Yeah. Great. Yeah. I was lucky to witness some of them over the years. So it's been a great ride. I agree. So we also should always talk about failures or missed opportunities or waste of your time. Any things you want to share in that? Yeah, that's a good one. You know, I don't know. So the short answer is no. And the reason being is I don't believe in failure 01:24:01in the sense that there's a lot of things that don't result in a publication. Sure. Right? But it's only a failure if you didn't learn something from that process. So I can think of things that we did and spent a lot of time on that resulted in a failed grant or didn't lead to a high-profile publication or didn't even bother to write it up. Right. But I don't consider any of those failures because each one of those instances I could possibly list off we learned so much from that process that then contributed to the next thing that did result in a successful grant or publication. So no, I can't think of any. I would also say that I've also been very, very careful with how I spend my time to make sure that I'm spending my time on things that I think are important. Yeah. Explicitly so I don't look back. Yeah. So I don't look back and say, wow, I never should have put my time into that, that was a failure. So yeah, the short answer is no. I really struggle to come up with anything that I sunk time into 01:25:03that I would say, no, that was a total waste. Yeah. Sounds great. And do you have tips on how to do that for listeners? Or because, I mean, how do you know what's going to turn out gold? Right? I think one thing that struck me when I started as a postdoc with you was even you said something, you know, you don't want to. You don't want to publish papers that don't change the field, which is I think, you know, it was new to me. You know, I think it was super helpful to hear that actually. Because you can't change the field with every paper obviously, but having that ambition to do it in the first place is, I think, a good one. Or any other tips or thoughts? Yeah, that's a great question. So, you know, I guess one is how do you bias things in your favor? Because you're right. You don't know what's going to happen. Right? Yeah. Yeah. Yeah. Yeah. Yeah. Yeah. Yeah. Yeah. Yeah. Yeah. Yeah. Yeah. Yeah. Yeah. Yeah. Yeah. Yeah. Yeah. Yeah. Yeah. Yeah. Yeah. Yeah. Yeah. Yeah. Yeah. Yeah. Yeah. Yeah. Yeah. Yeah. Yeah. Yeah. Yeah. Yeah. Yeah. Yeah. Yeah. Yeah. Yeah. Yeah. Yeah. Yeah. Yeah. Yeah. Yeah. Yeah. Yeah. Yeah. Yeah. Yeah. Yeah. Yeah. 01:26:00Yeah. Yeah. important. Maybe all three will, but usually not. Usually one of the three. And so, you know, leverage your bets, right? Have multiple horses in the race. You know, two is think about what the end outcome is if you're right, right? And I think that there are people that spend a whole lot of time proving something that's right that really doesn't move the needle, even though they sunk a few years into it. And so, you know, in my case, you know, our goal is translation. And so there's a lot of cool scientific questions out there, but if it's unclear how you're going to take that result and then translate it back or use it to eventually improve a therapy, I tend to put that as lower on my rank list as to where I spend the time. You know, we don't publish a lot of the smaller papers that we could. You know, it takes a really long time to publish even a small methodological point. And, 01:27:03you know, I'd rather just... Kind of mention that methodological point in the next bigger paper and spend the time on that rather than spend the time on lots of smaller methods papers that people may or may not see. And then the other one is you bias the results in your favor up front, but you're never going to get it right all the time because you're always going to sink a lot of time into a project that didn't work. And then it's just making sure you capitalize on that experience and that time spent so it's not a failure, right? So, okay. You spent a huge amount of time on this project. It didn't result in something that is worth writing up, right? So what are you learning from that? Because if you didn't learn anything that you can take with you into your next project, yeah, it is a failure. But if you bias things up front in your favor and then you learn from every experience that you have, then there isn't really a failure. Yeah, yeah, yeah. That sounds great. Yeah. That was another great piece of advice you gave me back 01:28:02then. Just saying that, you know, even the small papers take a lot of effort, right? So are they really worth your time, right? So hedging, like from the many things you could do, I think you once formulated it in a way that was something like there's always enough ideas, but being a good scientist means to choose the ones that will have the biggest impact, right? From the ones that you could work on. It was also super helpful for me, so. And remember, some of the smaller papers too, someone else will do it. Sure. And so it's also don't write the paper that you know, people always worry about competition and oh, am I going to be first? And no, that's a great opportunity to like, you know, relieve time constraint. If someone else is going to do that paper and write it up and do a good job, then you don't have to. Yeah. And so that's another one is don't, like I said, don't join the herd. If somebody else is already doing it, let them do it. Spend your time on something else because doing the same thing they're already doing, that's not going to help the field. Sure. Yeah. Although, I mean, some more replication is helpful, I think. 01:29:03Oh, yes. That is certainly also, you know, that's also some advice one could give people that have, you know, anxiety of being scooped to say that usually both papers will be valuable, right? But yeah. No, that's a good point. That's good. I don't want to dissuade replication. It's really, really important. Yeah. And actually, when I review papers, you know, I am very positive on a well-done replication that either replicated or in some cases didn't replicate it. But either way, no, that is also very, very important. Then zooming out further, like very general advice to the young people entering the field, why would they want to pick neuromodulation or let's say or lesion network mapping or why neuroscience at all in this time? Or do you have any advice how to choose your topic or do you? Yeah. No, I mean, this is, I'm sure that if you went back and looked at interviews and recordings from the last hundred years, someone is at any given decade, you know, they're 01:30:01going to be interested in that. Yeah. Yeah. Yeah. Yeah. Yeah. Yeah. Yeah. Yeah. Yeah. Yeah. Yeah. Yeah. Yeah. Yeah. Yeah. Yeah. Yeah. Yeah. Yeah. Yeah. Yeah. Yeah. Yeah. Yeah. Yeah. Yeah. Yeah. Yeah. Yeah. Yeah. Yeah. Yeah. Yeah. Yeah. Yeah. Yeah. Yeah. Yeah. Yeah. Yeah. Yeah. Yeah. Yeah. Yeah. Yeah. Yeah. Yeah. Yeah. Yeah. Yeah. Yeah. Yeah. Yeah. Yeah. Yeah. Yeah. to understand the brain and convert that understanding into neuroscience treatments gets better with each decade so as time goes on it's always the best time to go into neuroscience because we always have the best tools to to actually do things but but no I you know there's always a lot of you know poo-pooing of academia and oh I absolutely love my job I love what I do I get excited every day when I come to work and work with brilliant people and and to now is a very awesome time to be in neuroscience I would say that unlike you know people that said the same thing decades in the past the tools we have available to us today allow us to go after symptoms that are in need of treatments with real optimism that we can rapidly translate that back into an 01:31:00improved DBS you know stimulation setting or an improved TMS stimulation site I think that path to translation has never been shorter but but we all think that's the way it is and I think that's the way it is and I think that's to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to What's coming for neuroscience in general? Geez, I would say a whole lot more human stuff. I'd say that we're now doing things directly in people, all the electrophysiology happening in humans that used to be only possible in rodents and monkeys. And so I think the neuroscience of the future is going to be human neuroscience because 01:32:04we're going to have more tools for recording both invasively and non-invasively in human patients, more tools for modulating brain circuits in human patients. And so I do think the ability to directly do neuroscience in humans is really exploding. And I'm excited about it. That's great. We talked about some missed opportunities, like for example, recording serendipity, but do you think there are other things that we as a field, neuroscience or neuromodulation or neurology even, should be taking that we're not taking these days? Yeah. Yeah. Yeah. Beyond the topics we've already discussed, I would say the one that I've been thinking about lately is the gap between the animal science and the human stuff. And as I've started to wade into this, it has been very interesting to understand just 01:33:03how big that gap is and how different... Yeah. ...the approach and motivation is when you're doing a rodent optogenetics experiment, right, versus the clinical problems that we're facing. And is that optogenetics experiment actually addressing a clinical problem or is it just really cool, rigorous neuroscience? And I think there just needs to be more conversations. You have unique people like Carl Deserath, who is a psychiatrist and a clinician and takes it back and forth. Yeah. Yeah. And to get to these ideas to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to Translate or have a conversation with the clinical problems and you know your opto DBS, you know approaches, you know targeting that but but the the 01:34:01Opportunity there I think is is big but the gulf is is very very wide. Yeah It's not just not clear if it's the same region we're talking about right so say you don't know if it's the same disease same Symptoms same circuits same region. I I again, it's The gulf is so wide that skeptic could say okay You're never going to be able to take anything in eroded and translate that to anything useful in a person And I don't fall on that extreme but at the same time I think that that Many people do end up on on the the opposite side Where they assume that this? ginormous golf doesn't exist and that if you can Have an effect on a symptom that kind of looks like a human Symptom and eroded oh well gee easy peasy right just you know translate that optogenetics intervention to a circuit therapy in humans and and I think that that You can't underestimate the size of that golf 01:35:00Yeah, doesn't mean it's not possible to circumvent it, but it's it's massive and so, you know I don't want to just discount all the very cool neuroscience happening in romance, but but how we get there There's a lot more steps involved and I think It's commonly appreciated Makes sense. I want to be mindful of your time and I would have tons of more questions I want to start with one last guest question by Fred Schaefer for you Hi, I'm Mike. This is Fred one of your postdocs. I got a question for you Let's say Brigham says in six months for some reason whatsoever You can take a sabbatical, but the only thing you can do in that sabbatical is do a postdoc for two years I'm curious How would you? How would you decide where to go? Who would you work for and what would you do in that postdoc? Thanks? Oh Great question. I love it, too, man. So I get to do it to your postdoc with anybody that I want 01:36:06Man I so And you want to know how you choose Yeah, where you'd go as well, but maybe what's the rationale? to get to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to to the first thought that comes to mind is, you know, who knows how to run a center across multiple departments better than I do, right? And there are certain names that come to mind, you know, 01:37:00Helen Mayberg and others that have done this and launched these efforts, because that one's a challenge, right? And I, you know, mentioned before, you see where the herds go in and you intentionally do it differently, right? Well, I didn't become a, you know, division head. I didn't become a, you know, department chair, because I don't know how I would do it differently, right? Someone will do those jobs and they will do a very, very good job and probably a better job than I would do. But going a different direction from the herd is building a center across multiple different departments that put different people together in different ways with the idea that different things and thoughts and treatments might emerge from that. But it's very challenging. And so, you know, I would almost consider an administrative postdoc of how you learn to do that. And I think that's a really good point. And I guess on the pure scientific side of things, you know, it would probably be kind of on the, you know, circuit modulation from the animal model up to genetic side. 01:38:03Oh, really? Simply because it's, there's so much happening there and so many smart people cranking out, you know, I bet every week you open a nature or science journal, you're going to find at least one optogenetic study in there. And so, you know, I almost feel like I must be missing something, right? There's really important science that's happening there that the scientific community that does more basic stuff is really pumped about. And, you know, I know the human stuff really well, but that's an area where there are very, very good people that I think I could learn from to begin to understand, can you cross this gulf? Is it possible? And how would you, how would you even begin to approach it? And so I think that would also be really valuable. Super cool. I love that. I'd love to see that. You're the bench with mice. Oh boy. Going back to my pipetting skills. I did do mouse research actually. Oh, you did? 01:39:00Yeah. No, I did mouse research as an undergrad at Ohio State. And then I actually worked with Tom Woolsey at WashU who did the barrel cortex. He discovered the barrel cortex in mice. And we did a whole set of experiments in, you know, imaging the barrel cortex and trying to understand, you know, when you stimulate a whisker, why does blood flow actually go up? And is it trying to wash out bad stuff? And so I did actually do mouse research before I ever did imaging research. Interesting. Did not know that. It just didn't take. And you think you wanted to cover that I did not cover. I covered a lot, I guess. I think you know me well enough at this point to ask all the best questions. And no, this was a, this was a super enjoyable time. And so thank you so much for having me. Thank you so much, Marcelo. Thanks, Mike. 01:40:11Thank you.