Doris Wang 00:00They can still have trouble walking and they're still falling. And their quality of life is not good. Even though their tremor may be controlled, they can move better. But they have nowhere to move to, right? Because when they stand up, they fall. Collectively suppress pathological signals at the right moment. But then we can maybe enhance and boost the more physiological signals to the brain circuit. So that would be like the ultimate goal. So like at least in the life of these four patients that, you know, ended up sticking with the study, like we made a difference in

Andreas Horn their lives. Yeah.

Doris Wang Right. And again, this is small scale, but I think it's just a start.

Andreas Horn Welcome to Stimulating Brains. 01:00Hello and welcome to Stimulating Brains. Today, I'm delighted to welcome Doris Wang, who is an associate professor in residence of neurological surgery at the University of California, San Francisco, UCSF, and a functional neurosurgeon specializing in the treatment. of movement disorders. Doris is both a surgeon and a systems neuroscientist. Her lab studies how cortical and basal ganglia circuits generate human movement, how these dynamics are disrupted in movement disorders with a focus on gait, and how neuromodulation might restore them. Over the course of her career, she has studied how neural activity shapes developing circuits, mapped signals across the human cortex, basal ganglia, and spinal cord, and helped build implantable systems that can record brain activity, 02:00and how neural networks can coordinate the activity during everyday life. A recurring theme in her work is that restoring movement may require us to understand not only where activity occurs, but how neural networks coordinate dynamically as people act and learn. The focus of today's conversation is her new paper in Nature Medicine just out. In Five People with Parkinson's Disease, her team identified personalized brain signals associated with individual phases of walking, and used them to adjust stimulation within fractions of seconds. This was probably one of the most rapidly alternating adaptive DBS paradigms ever, really alternating with every step the person took. Three participants then subsequently tested the system at home in a blinded crossover study, and the study raised some fascinating questions. Should stimulation respond not only to a person's general disease state, but to a moment-by-moment structure of behavior, really ADBS at speed or at scale, is there one neural signature of gait, 03:01or does every person require a different solution, as was a bit the case in that study, and can a device do more than suppress symptoms, can it possibly also help restore the dynamics of movement? We will talk about these questions, about the many facets of gait and freezing of gait, and all you have there are symptoms in Parkinson's disease when it comes to gait, and about Doris's own path from neuroscience student to neurosurgeon scientist resident, thank you so much for tuning in, as always, Stimulating Brains, and welcome to Doris Wang. Okay, Doris, thank you so much for joining us on Stimulating Brains, this is a big honor. We've known each other for quite a long time, and I've always admired your work, it's really fantastic what you're doing, really cool to also be here, really cool to also build a lab around gait, at least right now, that's your focus, I think, 04:01probably your focus is a bit bigger too, but it's really fantastic to have you here. As you may know, we always ask one icebreaker question about hobbies, so do you have free time at all as a neurosurgeon, and what do you do?

Doris Wang First of all, yeah, thank you so much, Andy, for inviting me, I've been a big admirer of yours as well, and I've been listening to your podcast, so it's a huge honor. It's a huge honor to be here. In terms of hobby, actually my hobby right now is surfing, I kind of started, I guess, in med school on a trip to Hawaii, tried it, really fell in love with it, it's hard to find time for it, but I just got back from vacation in Hawaii, and had a great time surfing, yes.

Andreas Horn That's fantastic. I think I actually did see on the lab website that you, like surfing with your husband is a thing, so are you always going to do that? Yeah. Are you always going out together, or is that something you discovered together?

Doris Wang 05:00No, he has been surfing long before I did, and he's a really good surfer. Yes, before kids we would go out together, and now it's kind of like in tandem, so one of us has to watch the kids, so we hardly get to go out together.

Andreas Horn Of course, okay, great. Fantastic, well I envy you, I've been to Hawaii once, and was really blown away by the blueness of the water around it, and it's really beautiful, so I'm sure you had a great time there.

Doris Wang Yeah, it was fun.

Andreas Horn You trained both as a neuroscientist, and then as a neurosurgeon, which identity came first to your mind? Was it always the plan to be clinical, or how did that work?

Doris Wang That's a great question. I think what fascinated me and drove me to neurosurgery was actually the science, or the lack thereof. Back in high school, I just remember taking anatomy and physiology, and we kind of have a basic idea of how the heart, kidney, lungs function, but the brain was just like, it controls everything we do, 06:01but we don't know anything about it. To me, that was what drove me. It was like, well, I want to learn about it, and at the time, naively, I just thought being a neurosurgeon is the only way to study the brain, because then I can have access to the human brain. So yeah, that kind of set me on the path of wanting to become actually a neurosurgeon, wanting to become actually a clinician, especially a neurosurgeon, so I can have access to directly study the human brain. How at the time, I have no idea, but I got that idea in my mind, and I just kind of followed that path.

Andreas Horn Fantastic, yeah. It makes a lot of sense, and it was a wise decision, right? That it's really the only way to have true access to the human brain, at least. That's really fantastic. You studied at Yale, and then I think, if I'm informed correctly, you stayed at UCSF through an MD-PhD, but also neurosurgical residency, the functional fellowship, and now faculty life. What kept you there? What's so great about UCSF? 07:01Who were maybe mentors on the path or key turning points in your career?

Doris Wang Yeah, well, you know, UCSF, you know, the acronym stands for You Can Stay Forever, so... Oh, right. I think it's hard to leave, honestly. I came here, you know, when I was applying for MD-PhD programs because of the strength of the neuroscience community, and I think in the back of my mind, also the strength of the neurosurgery clinical training program. It just, you know, I felt immediately at home the first time I set foot on the campus, just talking to the faculty that had the breadth and depth of investigating the nervous system from every aspect, from the basic molecular mechanisms, cellular mechanisms to systems around science. And I think more so than, you know, what's available in terms of research topics, it's just a sense of collaboration and how as an institution, 08:01the community has just, everybody wants to push the science forward. You know, fewer like these big egos and people just really want to collaborate and work together. So that really drove me to, you know, sign up for UCSF for my MD-PhD program. And I think that's really, really important. And then throughout my training, I did my PhD with Dr. Arnold Kriegstein. It's actually, you know, a mouse brain physiology research. And I just, you know, fell in love with this community even more. And I think the proximity of this graduate campus neuroscience community with also the clinicians was really impactful. You know, I was able to shadow neurosurgeons, go to the OR, even as a PhD student. And that was really informative for me in terms of seeing this crosstalk between the clinicians, what they're seeing, you know, from their patient interactions, the clinical problems they encounter, and then bringing it to the basic scientists, or if not, like they're themselves are scientists who can investigate those questions. 09:06And then, you know, once I decided to train here for my neurosurgery residency, it was absolutely clear. I picked the right place, the right decision. Just the amount of, you know, training was phenomenal. And then the mentorship, like you said, that's what really kept me here. You know, we had Phil Starr, who is a mentor, now colleague, who is, you know, a father of kind of human neurophysiology research, basal ganglia research, and doing some really innovative stuff. And just being so welcoming and, you know, encouraged the inquisitiveness and curiosity. And then Eddie Chang, of course, who is amazing in terms of human language research work. So I think I coincided with this growth of human neuroscience research as, you know, this, what's available. And it just makes sense for me to stay because the resources we have, the infrastructure we have to perform these advanced human neurophysiology research is just unparalleled.

Andreas Horn 10:08Yeah. And I mean, it is a one of the few. And I think that's why I think we have the leading centers in the world in these things. So absolutely makes sense to me. Were there, may I ask, are you a West Coast kid or are you, because you were in Yale as well, are you, where's your, like, where did you grow up?

Doris Wang Family and base. Yeah. So I was born in China. I went through also grade school in China near Beijing. And then when I was in sixth grade, so I was 11 years old, my parents and I moved to Colorado. So that's where my parents are. So home is in the other C state. So I covered California, Connecticut and Colorado.

Andreas Horn But that's not too far, right? That's still kind of west coastish. So yeah, or not. Yeah.

Doris Wang Mountain range, but closer to the west coast. Yeah.

Andreas Horn Makes sense. And great. Were there any particular either mentors or maybe even patients or coincidence or whatever or turning points that shape what you wanted to study? 11:09Maybe like focusing on gait now. Why did you think gait is a thing that you should devote your next years on?

Doris Wang Yeah. So interestingly, I think, you know, going into movement and functional was a decision I think I made. I think I made very early in residency. It was when I went to Phil Starr's, you know, DBS case, the first one I saw. Just, you know, somebody with debilitating tremor, bradykinesia, rigidity, like turning on DBS just melted them away and can see a sense of calmness and just amazement. And that was fascinating. And of course, Phil was at the time also doing cortical strip, you know, the electrocortical strip during. And then, you know, that was just so cool. Investigating both basal ganglia. Cortical network function. And yeah, so so, you know, I wanted to kind of use that model to advance my own understanding of basal ganglia, cortical neurophysiology, because until then I was looking at single spikes. 12:11I didn't really study neuroscience and computational. So kind of have to train myself during fellowship with Phil and Cora's mentorship to kind of learn a different field. And then when I was starting as I guess it was in my fellowship. You know, as I was thinking about my own research question, what to investigate. You know, I think what I saw in clinic was that the basic things like, you know, appendicular symptoms like tremor, bradykinesia, we have a fairly good handle on if you get the lead in the right place, stimulate, you know, hair frequency works pretty well. Not ideal, but always be optimized. But, you know, Phil's lab was really focused on, you know, optimizing those with adaptive closed-loop stimulation. But there are many other aspects. It's just when I started seeing my own patients in clinic that, despite optimal programming, they can still have trouble walking and they're still falling. Their quality of life is not good, even though their tremor may be controlled. They can move better, but they have nowhere to move to, right? Because when they stand up, they fall. And it's just really intriguing to me that we haven't really figured out the gait circuit. 13:27I guess, early 2010 ish, we really don't have a system of study how humans walk, because all our technology is tied to either, you know, functional imaging, which you're doing, imagine gait, not an actual gate. Or, you know, we externalize leads, but you know, there's all these artifacts and still in the laboratory setting. And with all the wires, people are walking on treadmills, it's an artificial environment. So I think it coincided perfectly with, you know, this new wave of implantable bi directional devices. At the time, it's only 14:00for research only right with a first generation Medtronic Activa PC+S. And then it just opened my eyes to the possibility like, oh, now we can study like cognitive functions outside the laboratory, we can study, you know, things that are harder to study like gait, and take it in the real world environment, more like physiological ecological. So I think those two things like the technology being just becoming available. And the clinical question kind of drove me to study gait.

Andreas Horn Fantastic. Last question before we actually dive into your work, and how did your scientific interests survive the intensity of neurosurgical residency? And how did clinical training maybe change your interests in research?

Doris Wang Yeah, I mean, I love clinical training, I love operating room going into it, you know, it doesn't matter what cases I'm doing, you know, I'm interested in functional in terms of the impact, but I love, you know, vascular skull base tumor, even spine, everything. 15:05I think, you know, what ultimately, I mean, neurosurgical training is tough. But I actually didn't mind it, you know, I feel like science is harder. After after surveying, surviving a PhD, where you feel a lot of things are beyond your control, like the results, the publication process, grant writing process, they're extremely humbling, right, you just kind of keep going. So I think having that clinical, like immediate gratification, helped me feel like, you know, things are still moving in the right direction. So it's extremely complimentary. But then when clinical training becomes, you know, really tough, or, you know, either emotionally, or it just becomes routine, I think the scientific drive, like, okay, well, ultimately, I'm going to be, you know, chasing my dreamless dream. And I think the human brain and what are the questions I kind of always have that curiosity, like, how can we make 16:02this better? Either is it like through like a surgical technique, or through a different way of approaching things. So that kept it interesting for me. But I thought it was, I mean, yes, the time demands was rough. But also, I like being kept busy. And I think having these two complimentary, like immediate gratification versus kind of long term, type one fun, or type two fun, you know, that that, that really complemented each other.

Andreas Horn Fantastic. All right. So the clinical problem of gait, gait feels effortless when it works, what does the nervous system actually have to coordinate for us to take one normal step? We probably don't know, right? That's probably the honest answer. But maybe there are some things you know, that you could share what even happens in the brain? Or you know, how does it work? What has to happen to make a step?

Doris Wang Yeah, and I guess, you know, we don't really know. But these are kind of my hypotheses. So when we're normally walking, like once you start walking to 17:02maintain the rhythmic, just like while you're going for a stroll in the park, for instance, actually requires very little higher cognitive, probably cortical function. You know, it's just the basal ganglia scaling these, if there's, you know, you have to change your step length, change direction. Most of it is probably in the lower brainstem center, you know, it just becomes automated, the cyclical event that drives your muscle activation between the two legs while maintaining balance. But it's when there are, you know, changes to that just direct rhythmic, continuous walking, then that requires more cognitive higher order function. So things like you know, initiating walking, going from static to movement, that probably requires some type of cortical demand coming from, you know, pre motor, like the intention to move to motor cortex. So you have to initiate that drive. And then also more complex gait, right? Like if you're walking in the forest, 18:01going hiking, or you had to avoid tree roots, step over rocks, that requires a lot more attention. Again, the normal brain, I think a lot of these circuits even involving like the visual, visual circuits, sensory processing, we don't have to devote a lot of like attention to it. But I think that's where it breaks down in Parkinson's, like they can't walk sometimes, because they're not paying enough attention. But then these the automated walking process, gait process that breaks down.

Andreas Horn Interesting. Yeah, that makes sense. And then you did mention, you know, when you started your own lab to essentially there was the technology was right with the percept, sorry, the Activa PC+S and then I think the summit later, that was used in the recent study. And then, but but you know, maybe Phil Starr and others had also built some sort of system with the ECoG strips and you mentioned the, the measuring gait was harder, right? So I assume or I always assumed that when you got your faculty 19:03position, maybe there was a starter package for a gait room or how you had to set up that environment, right to measure gait. And I think, you know, gait dysfunction can mean short steps, asymmetry, instability, falls, freezing. Is that all one thing? Or should we? Yeah, can you talk a bit about that? How you measure gait? How? Which components are we even talking about? So yeah,

Doris Wang yeah, so my start up package is pretty meager. But you know, in terms of gait lab, actually, I was fortunate enough in that admission Bay campus, so that's kind of our newer campus, there is a kind of a human performance core. So it was initiated by the Department of orthopedic surgery and sports medicine, where they kind of have the capability of you know, all these cameras, motion capture camera system in a large space. And, you know, force plate to measure kind of these kind of human kinematic movements. And nobody has really utilized it to 20:01study gait per se, especially not gait in the Parkinson's or neurodegenerative population. So basically, you just kind of have to pay the fee and tell them the projects and then you know, they have a lot of the equipment available. But one thing that I really, you know, invested in, I guess it was the mobile sensors. So instead of being in a static gait lab, I was able to use these cameras. And I was able to use these cameras in the gait room with these cameras. You know, I learned about these wearable sensors. So this called the inertial movement, motion measurement units, IMUs, which are just fancy tri axis accelerometry. So I use this XSense system. So it's pretty easy to use it like small suitcase, it has 17 different sensors, we can put it you know, around major joints, and it's on patients or study subjects. And then we can also use the 21:10gait. And then to your question, you know, what is about Parkinson's gait, that's challenging to treat? And it's all the same thing? I don't think so. The short answer is, it's very different, right? Like, when you think about a Parkinson's gait, everybody's symptoms are different. You know, some people are like the rigid type, some people have really bad tremor, some people are fluctuators, some people are not. So it's not like a homogeneous disease. And same thing with Parkinson's gait. Some people it's like purely, you know, slow, it's kind of like a hypokinetic gait. So like meaning their steps are really short, they're very stiff. And those patients actually do probably quite well with traditional stimulation, because once you release their gait, you can move them, make them move more fluid faster, their gait improves. Some people do have leg dragging that 22:02could be from dystonia. So that's, you know, abnormal muscle contraction, which could also be a part of Parkinson's symptom. So for those, you know, it's more variable is the target, you know, treating the dystonia in that one particular leg. So those patients may have more asymmetry. And now we get into kind of the more complex like freezing of gait. Oh, I think that's a whole nother piece. And we really don't understand what causes that. And I think that's a more like widespread network disorder. Because I think involves a lot of cognitive processing, attentional processing deficits, visual spatial deficit. And as because, you know, many things can cause trigger freezing. It's either you know, standing up from a chair, like the initiation freezing, kind of a learning triggered freezing, or visually, right, if they're approaching doorway, or if they're under stress. So there's 23:01all these things that can trigger freezing. While the output is the final output and result is the same, it's probably kind of some like jamming of the entire system that prevents this automated walking to take place. But the trigger, I think it's much more widespread. And that's why it's really hard to study.

Andreas Horn So you're essentially studying multiple problems at once, or are you focusing on one specific form of gait or a component of gait? I mean,

Doris Wang yeah, so so with this study, you know, initially, I just want to see what happens, right? Like, with gait, what happens when people are walking relatively well, like, you know, when gait is not one of their primary concerns to people who have, you know, stuttering, shuffling, you know, slow hypokinetic gait. Right now, I'm actually focused on freezing of gait. But yeah, just, you know, kind of going back to when I first started, I just kind of want to see, you know, again, I can't 24:02sample the whole brain. So we had to pick specific areas to sample. So for these Parkinson's patients, you know, studying basal ganglia, LFP, local field potential from their DBS leads, either in the subthalamic nucleus or the pallidum, it's opportunistic, because, you know, these are patients who require surgery anyways. But then we did put in extra permanently implanted cortical strips. So for contact ECoG strips. For my study, I focus on the premotor and motor area. So again, we're sampling a very small local region. And it's probably kind of more localized to upper extremity, the homunculus. But through it, we were actually able to decode some gait events. So that was kind of one of the surprising findings.

Andreas Horn Yeah, great. Really cool. We'll get to that in one second. So if we think about you know, adaptive DBS these days, many people will of course think about beta synchrony or like beta power as a, you know, 25:05a steering signal for adaptive Parkinson's, even FDA approved now as E-marked. I think in 2016, you found if I remember correctly, that a familiar, you know, this the subthalamic beta features were also present in dystonia, to some degree. So potentially were also present in dystonia to some degree. So potentially casting doubt about the specificity of these biomarkers. Did that teach you something

Doris Wang about how to even think about these biomarkers? You know, how specific are they? And any thoughts on that earlier study? Yeah, for sure. That's work I did with Phil Starr during residency. pallidum, STN. And then again, they're in the off med state. But yeah, even dystonia patients, we saw this beta peak. And then in terms of the amplitude, it's highly variable. So that to me, 26:05taught me that, you know, beta isn't everything. It's probably not that specific. It just represents kind of motor state, movement state in general. But there are more specifics within beta physiology in terms of, you know, coupling with maybe cortical gamma, um, its coherence with other regions that that's more specific towards either states or medication states or movement states. So that to me, just pointed to, you know, beta isn't everything and everybody's beta peak, right? It's slightly different. Some people may have it, sometimes people have like double peak, single peaks. So yeah, reliance on just a single file marker, blanket, you know, across everybody is less personalized than we can actually make. So that's why I think, you know, when I was translating this into gait, I kind of went through it kind of agnostically, like, what are the actual signatures and brain fluctuations and 27:01signals fluctuations that occurred during walking? Or does it even fluctuate at all? You know, from that point, we don't, we don't even really know that. So that

Andreas Horn makes sense. In 2020, you described Parkinsonian gait as a disorder, gait dysfunction as a disorder of network oscillations. At that point, how much of the present research, program did you already envision or kind of draw in front of you, like, maybe framed differently moving to the recent Nature Medicine paper? How when did that start? And how like, when did the thoughts about that start? And then how long did it take to actually start the study and then get it published?

Doris Wang Yeah, so I think I had this goal in mind when Summit RC+S, came available. And that was, I think, during my fellowship years, I did the first implant, with Phil, you know, for the Summit RC+S and because, you know, the big difference between Activa, the first generation to this one is that we can now stream data. So we're not limited by, 28:03you know, the 12 minutes capacity onboard memory device limitation of the Activa system. Now we can wirelessly stream, which means we can almost stream like 24 hour data right from the human brain. And it's also rechargeable. So we're not worried about battery consumption. So that's one I was like, okay, all the experiments I dreamed about, maybe possible with Summit RC+S s. So I guess that was around 2018 2019. When I was, you know, that's when I was building my lab, and I kind of envisioned a program where, you know, I can implant patients with RC+S, kind of what Phil did, and then, you know, place cortical strips, deep basal ganglia activity, and then being able to measure gait. So, the 2020 paper, I think, was that a review you're referring to? So that's kind of what, you know, my thinking process about, yeah, gait has to be a network disorder. And that's why in Parkinson's 29:03patients, right, they have all these different problems, and how can we solve it? And first one is to understand what happens? And what are the actual signals controlling this process?

Andreas Horn Really cool. So let's focus on the new Nature Medicine paper just came out, I think a few months ago, a month ago. Can you briefly summarize it, and maybe also take us through the system you built, you know, from heel lift to electrical pulse. You did five patients, clinical study, I think three, then also in a randomized, you know, phase later. And the idea was to stimulate while walking, you know, in a gait, like in a walking specific fashion, right? Yeah. And then you did a lot of the work with the 30:06And in the five patients we tested in the laboratory, it restored symmetry. Left and right step length and timing were better, which we usually think of as a more stable gait. Walking speed was highly variable, but variability between left and right step length and timing also improved.

Doris Wang So it's capturing subthalamic activity using the Summit RC+S system. dynamic activity using the RC+S system. Again, these are patients still implanted. So basically, from that study, we did detect in real time walking, dynamic changes in beta or low frequency power, I think included alpha and theta as well, that dynamically change in the left and the right brain, they're off cycle, which makes sense, right? Your brain should be changing as you're generating these rhythmic movements. And using that study, you know, our idea is like, okay, what if we can dynamically mimic what the brain is doing, right, naturally? And again, those are patients without gait problems. What if we can just dynamically change amount of 31:04stimulation? While someone's walking, can we make their walking even better than what traditional DBS can offer? So using the RC+S system, our goal, first of all, scientifically is like, first of all, can we even identify personalized biomarker of left versus right leg swing from either cortex? Or the palate elites? So answer to that is, yeah, most likely we can. And then the idea is, then can we use this novel system using completely embedded algorithms, or not with an external computer, embedded linear discriminant analysis classifier on board the RC+S to rapidly change stimulation amplitude. And again, this distinguishes from what's available, because as beta changes, you know, usually stimulation ramps up and down over minutes. Now, we're talking about sub second, like within a 50 to 100 millisecond window, rapidly change how much stimulation they're getting. And the answer is that they can figure it out, you know, my postdoc, 32:04the first author in the study, so it is possible. And the final question is, you know, if we institute this change, either we're increasing stimulation while they're swinging the contralateral leg or decreasing stimulation, does it actually alter their gait dynamics? Is it beneficial? Is it worsening? And, you know, again, in the five second window, we're talking about sub second, we're talking about a lot of different things. So we're talking about, you know, the ! And also, you know, what happens to their other Parkinson's symptoms, right? Like, since we're, and all in total, they're receiving less stimulation. So can we first of all institute this 33:04in the natural world and what happens? So three patients enrolled in that the three patients with really reliable biomarkers and who can go undergo the testing? And, yeah, and the three patients, you know, two of them really prefer the adaptive, this rapidly changing stimulation compared to their closed loop. And they remained it even when the study finished. And the one person who, you know, liked it, but didn't like it as much compared to her natural setting was mainly because, you know, her gait problem is quite different from the other one. And her main Parkinson's complaint for her was tremor. So she felt she had less tremor control with this adaptive, even though her gait metric was better.

Andreas Horn That's cool. I mean, we should probably highlight this. You mentioned it in passing, right? But, you know, if you're a person who's a person who's a little bit more sensitive to the 34:16! 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.

Doris Wang Yeah. Yeah. Yeah. was built to be a research system. So every feature is unlocked. So we can change frequency. We can determine the biomarker 35:01on the Percept device. You're restricted to the beta range right now. So nothing was locked. We can do all the streaming. So it's purely for research system, which is like somebody like us. It's like our dream device. Of course. And yeah, huge props to Medtronic. Their R&D team for supporting this research device and research platform. That allowed us to use this device. Yeah, go ahead.

Andreas Horn Would it in theory be possible to, you know, just use the Percept now, do the same thing? Or is that more limited to what you've done?

Doris Wang In theory, yes. But as you mentioned, alluded to earlier, there's a lot of device on lock. So they don't have, you know, even an ecosystem to do this streaming. They don't have to even build a computer to interact with the device in that way. So there's a lot more limitations. And yeah, that's one of the things, right? Like the findings that I found from my paper, 36:01I can't really translate it to the current commercially available device.

Andreas Horn Got it.

Doris Wang Without asking for a lot of on locks. Not yet.

Andreas Horn Yeah, yeah. I mean, this is still, yeah, yeah. Okay. And then you did record from both the pallidum and cortical areas. And I think in most of the cases, the cortical, the basics was probably the better decoding or control source. Why do you think is that? Yeah.

Doris Wang And yeah, a couple of things. So, you know, again, like using just the computing, the power, yes. As a cortical slightly outperformed the basal ganglia lead. But I think one thing is like what stimulation is turned on because of, you know, artifact stimulation, artifact, and even though there are methods within a device to blank out the stimulation or artifact, you know, the computation in terms of specificity of the palatal electrode is still going to be challenging compared to cortex. It just because it's a more distant site, 37:00less prone to stimulation artifact. And yeah, so that's why, you know, that's my guess. And then did I capture the optimal site to decode? No, I think, you know, ultimately late primary motor area, which is surgically harder to access, you know, probably will give us even better decoding, right? For like muscle activation. But it's just too dangerous to go that route. So again, we just use something that's accessible. Again, these patients didn't even have extra burr holes. I used the same burr hole for their DBS, just with the electrode posteriorly. And it was good enough, I guess, for what we're trying to do.

Andreas Horn Really cool. Right. And so what you also, I think, identified a separate biomarker for each hemisphere. And maybe you can even talk a bit about what the biomarkers even were in the end, right? So what was the I think you did measure electrophysiologically 38:02when the step happened, right? And in that moment, you simulated the contralateral hemisphere and then it took turns. Which which physiomarkers, how did they look like? It wasn't beta alone. You mentioned alpha a little bit, but was it a combination or? Yeah.

Doris Wang Yeah. It varies so much. Yeah. That's a one of the things you know, again, some of most of them were within like the alphabet a range. But in terms of the actual specific frequency band that can distinguish. And again, we're not looking for just even like the sensitivity. We are also looking for specific specificity. Like there's this individual frequency band can distinguish between left and right leg swing. Right. And you know, in terms. of the actual and the biomarker we're using again a lot of them are from the cortex but there are a few from the pallidum they can be as narrow as like 2 Hertz for instance subject one the left GPI biomarker was 43 to 45 Hertz so it 39:05can be super super specific right and then yeah and some people you know like the best band is like m1 like 11 to 13 so it's actually extremely narrow so

Andreas Horn that's what we ended up using so you essentially had a first phase where people just walked around you required data with the gait lab and then you identified the biomarker in the second session then did the stimulation and had to also put in the you know the logic of when to stimulate it onto the device

Doris Wang right yeah yeah so we did many many sessions you know like before patients DBS was activated to see what those biomarkers are and then again just have them walk back and forth in the lab and then doing it both in the lab and the on meds day as their meds is wearing off and then once their stimulation is activated when we had to do a few more of these sessions and then we even took some of these recording sessions at home so we measure their 40:00like walking when they're just walking navigating home and we have the patients actually activate their own device we trained them you know for an hour each day they would just turn on stream their device where the communicators and then walk in a certain path around their house while we stream their data

Andreas Horn yeah so we had multiple ways so could people notice you know the fast switching normally if you ramp up the current pretty quickly in let's say monopolar review you can sometimes sense it was that a thing did they ever complain about oh I feel tingling all the time or you know the rapid alternating switching on and off

Doris Wang yeah surprisingly not I think this is due to a couple things you know again we didn't shut off stimulation altogether we kind of vary between their clinical setting and then half of the clinical setting so like you know four milliamps to two and then also because I think it's pallidum instead of you know in the STN which you might be more sensitive you know there might be like a 41:00longer washout effect and and so yeah they actually didn't feel the

Andreas Horn stimulation change make sense what was the hardest part of making this system work the neuroscience the engineering the surgical configuration probably not so much your routine there yeah or even regulatory path using daily life what was the most challenging part of that that's stressed

Doris Wang study I think that's to implement so I guess it's a combination of the neuroscience and an engineering you know identify like once we have the biomarkers how reliable it is so it's like the biomarker validation that's the most challenging because I think you know this is huge kudos to Ken Louie and the rest of the study team. You know, he to change configure a lot of things like the ramp rates within the device the sensitivity of the lda classifier to optimize you know this is just like watching the tablet stream and then seeing stimulation going up and down as somebody's walking whether you know doing it in real life 42:05and then also post hoc to figure out like what we can change to optimize the actual classifier function to make it reliable and robust did the

Andreas Horn heterogeneity between the participants including also the bio biomarkers did that surprise you or were there other surprises that you didn't think about when you know before starting the

Doris Wang study yeah i think um couple things yes the heterogeneity definitely made the study more challenging as i mentioned alluded to you know uh two of the subjects are more of like the near freezing rigid economic genetic type whereas the others are asymmetry and then someone was like dyskinetic and so they all had some subtle differences um you know ideally i i would recruit the same patients but you know it doesn't always work out and also with the time limit of the 43:00availability of these devices i kind of have to you know implant them in a certain time frame um so that made the analysis uh more challenging you know the biomarker as i mentioned is highly individualized so we couldn't really generalize you know something that worked for patient a may not work at all for patient b so i think individually customizing and pouring over that and then also making sense of analysis like you know why does adaptive stimulation improve speed walking speed in patient a but actually worsen it for patient b and what are the common themes um so teasing through that heterogeneity in a small end that was very challenging so your study came out of this study with this Something that worked for patient A may not work at all for patient B. Adaptive stimulation might improve walking speed in patient A but worsen it for patient B. Teasing through that heterogeneity in a small sample was very challenging.

Andreas Horn i think back to back with in Nature Medicine with another study very similar from geneva or uh is it zurich uh switzerland Eduardo Moraud's lab yeah i'm not 100 sure i saw him in geneva just now at opto dbs speak about the study that's why i said geneva but i 44:01think he's probably in

Doris Wang he's from switzerland yeah with uh yeah can you

Andreas Horn and i i'm planning to also invite him on the show at some point too um but you know this is of course first of all a great success for the field of dbs of having two studies coming out in that um uh really high impact uh highest impact journal um was it a back-to-back thing uh did you know about the other studies or was it more the editors kind of lumping these into the same issue how did that

Doris Wang work yeah so ironically two years ago at the dbs think tank you know i was tasked with organizing the session for the adaptive db or a gait right uh for the think tank and then i invited Eduardo Moraud as one of the speakers so it was during that meeting i learned about his work which i thought was amazing so his work is using the commercially available percept and i think it's only pc in europe they don't have the rechargeable yet but you know again he does they have some unlocks but using kind of the beta power so he was a lot 45:05more restricted using the beta to determine like um turning up stimulation and i think it's a key but it's really awesome work so i learned about it during that and i know at the time he was submitting uh to Nature Medicine so we didn't plan it but um it just happened i think our papers got accepted and came out at the same time so it's very it's very cool to see his work yeah

Andreas Horn yeah as well and and so so i think maybe you can briefly if you if you don't mind summarize also the differences or the um you know uh for just for the listeners to kind of two of these big studies similar topic or same topic even um i think what they did differently was they had kind of two settings right they probably a priori tested what is a good gait setting what is a good other setting right and then adaptively switched to and from the gait setting is that correct

Doris Wang 46:00roughly yeah well i think it's uh maybe the amplitude for the stimulation as stn because against with persep it's a little bit more limited in like in terms of switching like frequency and other parameters but yeah uh to yeah so then they basically again um their study is restricted to only stn dbs leads and then in terms of identification i think they um um you know study and recorded from many more patients uh in terms of what uh STN beta encodes in terms of gait dynamics state like walking versus sound walking walking versus sitting versus standing so that biomarker is pretty robust and then again yes while they patients were walking if they're actually walking then they use this biomarker to change to the optimal i guess stimulation amplitude while patients are walking so pretty different because it's you know not alternating between the two hemispheres which our study is doing so r is very specific to actual movement there's more like a generalized movement state 47:05but i think it's you know phenomenal that they they were able to achieve that with the limitations of the Percept device. think you know what whatever they find could be directly implemented in all the patients

Andreas Horn or you know who have the current device interesting okay cool so um maybe zooming out a little bit more towards these circuits and the idea of you know dbs as a circuit neural modulation thing you i'm not 100 sure now but i know it's a Phil Starr study where they did a lesion in the palliadome and the paludotomy and then also record it and saw that there were changes in the cortex and you know from after performing a lesion was that your study or were

Doris Wang you a co-author on that I was yeah I was co-first author with a Cora on that study like thalamotomy and paludotomy and effects on cortical gamma instrument which is a you know dyskinetic yes

Andreas Horn 48:01right so thinking about that right um I often have the you know when I simplify things at least for the audiences I sometimes say you know DBS mimics a lesion roughly it's not a lesion but it mimics it right probably not completely wrong um and and then if you would uh talk to you know other people in the field they often say well but that's not true because you know DBS fires action potentials and so on and I said well lesions can lead to added activity too right just by loss of inhibition or you know complex effects like and I think your study there with Cora um and yourself as first authors that you just mentioned is is a core example for me right where you have more activity of something in the cortex after performing a lesion right that's um uh contra in intuitive in a way a counterintuitive and so with adaptive DBS on the rise especially your study now with a very rapid and very kind of um uh highly adaptive way of stimulating right very um uh so then you I think 49:05we go away from a lesion or at least it's a it's a scattering lesion right it's a it's one that you can switch on and off all the time um how do you think about that maybe also in the advent of you know focus ultrasound surgery coming around or like um will these very adaptive or very smart ways of DBSing the brain maybe become a new thing that gets ever more complex and goes away from this simplistic kind of virtual lesion idea or how do you think about that whole ballpark

Doris Wang yeah that's a great question um I mean right now I kind of think of DBS as kind of in training circuits so you know you entrain in certain circuits you suppress certain you know signals like right high frequency typically suppresses like the beta range Alpha range uh LFPs but you again entrain them into different rhythms 50:03so that's why stimulation can entrain high frequency stimulation can entrain into gamma can train into gamma. So that's really nice work done by Phil Starr, that can lead to dyskinesia. And that's kind of the limitation of DBS. If you're in a static, constant frequency, then, you know, while you help some symptoms, you may create other problems by making the system more stuck, right, you're actually reducing the flexibility again, by entraining them into a certain rhythm. So I think, you know, to your point of the stuttering lesion, that's kind of exactly the motivation for my study. Like, what if DBS is actually causing eight issues, because it's reducing this network flexibility, by entraining them into this high frequency, and by kind of just varying the amplitude. And, you know, furthermore, like we can vary the 51:01frequency to like all the other difference of parameters. We can vary the frequency and all the other different parameters. So we can selectively suppress pathological signals at the right moment, but then we can maybe enhance and boost the more physiological signals to the brain circuit. So that would be the ultimate goal for any type of precision neuromodulation. Now, compared to focused ultrasound, it's really interesting. procedure for essential tremor and now Parkinson's disease. We are collecting a lot of, you know, resting state functional MRI data and DTI data. So it would be really interesting, you know, again, to look at least initially post hoc at what are the changes after we create this acute lesion to the resting state functional network? Are we, you know, restoring some motor activity, suppressing others? And you can do think about all the different ways to do more sophisticated 52:01analysis, right? With like neurophysiology. So and the question is, like, how long does it last? I think, you know, one of my guess the caveats, I think, with the current therapy of constant DBS or the permanent lesion, it's not dynamically changing. Once you do it, it's there. That's why I think for me as a neuroscientist, I think this idea of being to modulate in dynamically change the circuit by whichever way coordinated reset, you know, different frequency, different contact. It's a lot more attractive.

Andreas Horn Of course, yeah, no, I very much agree, right? With, I mean, we probably both agree that this is a very big discussion. There's a lot of pros and cons towards, you know, lesion, DBS, one set and done and all that. We've all heard these discussions. But of course, there's a huge advantage of this flexibility in the 53:01future, right? With with more sensing and more smart decision making in when to stimulate how to stimulate where to stimulate. Right? Exactly. Fantastic. Yeah. Okay, so so maybe

Doris Wang thinking next

Andreas Horn steps right from this fantastic study now, which is still a feasibility study, I think, you know, powered to essentially show show this is possible. I think you write that very prominently and transparently that it's not, you know, power to show superiority or so. Yeah. Yeah. Yeah. of maybe size, duration, endpoints, to compare it to CDBS, or maybe you're planning something like that, but even if you're not, what would be a good next step in terms of, yeah, a real bigger trial?

Doris Wang Yeah, you know, I think if we can expand it, ideally to, I think, on the order of somewhere between ends of 50 to 100 subjects, and then if we can do it in people who have 54:06existing devices without the need for additional electrodes on the cortex. So, yeah, one of the things, actually, I mentioned is the differences between Summit RC+S versus the Percept, again, limiting to within that five hertz range. The other thing with, I should mention, for Percept, the current adaptive DBS, Their biomarker cannot be two Hertz or fifteen Hertz. It has to stay within that five-Hertz range once you pick a peak. We are about to submit a manuscript in which, with our data, summit data, if we restrict to a five hertz band within allowable range within Percept, like how good is that in terms of determining walking versus non-walking state, which is pretty good, but in terms of leg swing, that's a little more. It's a little more challenging. So, yeah, so it may be possible to replicate this within the Percept, you know, cohort. 55:04And I think, yeah, if we can have, you know, a lot more patients, identify maybe personalized biomarkers of walking, I think that's going to be the biggest challenge. Can we do that reliably? So, I don't know. The short answer is, yeah, I think this is going to be really technically challenging, but we're going to have to see what we're going to be able to achieve with the current system. But we do have a study we're submitting soon as a kind of like a follow up. So using Summit RC+S again, just like Eduardo Moraud’s paper, we picked We picked out biomarkers of walking versus non-walking. At least that biomarker story has been published in Science Advances just this year. And then, you know, we again using the Percept or sorry, Summit device figured out what is optimized gait setting. And we're not restricted to just the amplitude of the stimulation. 56:00We actually change frequency, amplitude, pulse width as well. And then once we identify the best gait optimized setting again, when they're walking switch to that setting. So it's a state dependent adaptive VBS. I think that probably is a lot more achievable in the larger cohort. I think it's easier to determine walking versus non-walking using existing electrodes and then, you know, switching to a different gait optimized setting. Then like this study, that's. In Nature Medicine.

Andreas Horn Do you think your study or studies like this one could, you know, bring companies to, I don't know, extend the hardware that's needed or to even reopen a run with Summit? Is it, I don't know why the Summit was so such a short time window. Probably they didn't have a full clinical approval or something was just for studies or. Yeah. Yeah. Okay. Yea

Doris Wang h.

Andreas Horn So it's unlikely that it comes back, I guess.

Doris Wang Right. I don't think it will come back. 57:00I think it was a huge effort, you know, investment by Medtronic to kind of figure out whether it's achievable and, you know, we learned so much from it and yeah, it's really a shame that we don't have access to it now, but I think, you know, with other companies, right. And even with, you know, with Medtronic percept device, right. Like with the accelerometer, like you might not need a complex, like in internal, you know, neurophysiology biomarker to determine whether someone's walking. Just look at the accelerometer, right. And it's probably going to be very good. So yeah, I think if, if we and other groups like, you know, Eduardo's group that demonstrate like these state dependent and movement state dependent ADVS really works, it may urge these companies to, you know, be more motivated to unlock these features for future studies, larger studies. Yeah.

Andreas Horn And I mean, a lot is going on in China too, right. I think Pins Medical has a device. That's even 3D compatible and has a streaming capability. I hear scene rate, like the, there's a lot of like, maybe the innovation is faster there. 58:05There's really a lot going on. So you know, possibly you know, something like that could also be possible to, to do studies there or collaborate with people there. So yeah, really, really fascinating. You did mention the accelerometer and I realized that was on my list too. So I wanted to ask why did you use the maybe harder way of EFIS decoding? And if you compare that directly to just simply in quote unquote, having accelerometers on the shoes or something. Yeah. Yeah. So for this study, yeah, exactly.

Doris Wang So I think if you have an accelerometer like built on this device, I think, you know, the Circus is a lot easier. Then you don't have to stream to an external device to kind of feedback. Yeah. Yeah. So there's always a lag. So for this study itself, you know, accelerometer just picks up and down, so you won't be able 59:02to distinguish between left and right step. So and plus we can't use the accelerometer output as one of the biomarker to set the LDA classifier. So that in terms of the, yeah.

Andreas Horn You can't right now with the hardware, right? But technically it could be pretty easy in a way if there's a Bluetooth connection between, you know, all of these devices. Yeah. So, so if we, if we had the hardware necessary to do that. True, true. Yeah.

Doris Wang So I think

Andreas Horn your, I think what you mean is for that, like,

Doris Wang you know, if you had an accelerometer on the left ankle versus right ankle, like external accelerometers to feed into the device, right. So we'll need a Bluetooth connection. There's a lag, but yeah, technically it will be feasible. But just like the technology wasn't available and probably won't be available for this study. Yeah. Yeah. Yeah. Yea

Andreas Horn h. So it won't

Doris Wang be available to do that for this study. But the Summit RC+S itself has a built-in accelerometer. 01:00:03So you know, in terms of synchronizing home recording and home streaming, that's how, that's how we synchronize the neural streaming with their activity.

Andreas Horn Oh, interesting.

Doris Wang That's cool. Yeah. So we use a onboard accelerometer. So I think, you know, if they change the circuit, again, that would be a hardware slight change to use accelerometer signal to trigger. Yeah. Yeah. Yeah. Yeah. Yeah. Yeah. Yeah. Yeah.

Andreas Horn Yeah. Yeah. Yeah. Yeah. Yeah. Yeah. Yeah. Yeah. Yeah. Yeah. Yeah. Yeah. Yeah. Yeah. Yeah. Yeah. Yeah. Yeah. Yeah. Yeah. Yeah. Yeah. of the accelerometers or maybe sometimes you want to wear different shoes and then they're there you know so it's cumbersome it's like hearing aids for example that people would complain about always 01:01:03having to maintain them to wear them to charge them and all that right so i think there's this beauty in having it all under the hood all kind of uh implanted and you don't have to worry about it so plus of course there's the neuroscientific insights that you found out how to decode these things and um the study so yeah fantastic okay so um maybe can you talk a little bit about beyond gait right so motor learning and agency is also part of your broader scientific program but i must admit i'm not an expert on you know what you've done there and um maybe just uh about what you what you want to share uh past studies potentially future studies or things that you're interested above and beyond gait yeah

Doris Wang yeah so um you know actually before gait uh yeah one of my interests is uh kind of like impact of dbs and cognition and like other functions right we know from animal literature 01:02:02and some human studies that the basal ganglia striatum pallidum stn is involved with like motor learning and um like sequential motor learning skills and something that becomes internalized and automated and um disruption in that could lead to you know motor learning deficits of course like for patients with parkinson's not their top complaints like oh i can't learn like how to play the piano anymore with my parkinson but i think from a human neuroscience perspective again it's opportunistic because we have these you know lesions and deficits in a circuit um and we can tune it and alter like uh try to causally probe the circuit with stimulation or medication so uh one of the one of the studies kind of like studies i did again with the same patient cohort is like have them now perform a multi-day learning task like just typing on the keyboard learning a motor sequence and then we actually 01:03:00recorded some of their sleep data too you know so what happens offline learning so um yeah that data like unfortunately we didn't you know have a lot of resources once the gait project took off to really analyze it. But Kara Presbrey, who is a graduate student with me, analyzed that data and just looked at, you know, in terms of even before we start the movement, over a span of a multi-day course, does the neural circuit change? over a span of multi-day course? Does the neural circuit change? Does the signal, especially motor cortical signal, kind of tune or predict the accuracy or improvement in motor performance? So again, it's a small study, but it's really interesting. It just kind of gives a glimpse of what this device is capable of. We do have a manuscript about to be submitted in terms of learning, but to do with gait. So it's a gait adaptation task. So it's kind of almost like a Dance Dance Revolution video game. So we have patients walk on the treadmill, and then we put markers on their shoes with an infrared camera. So that shows their shoe 01:04:04and foot position. And so they're walking along this path, and they basically have to alter step length to step on target. So this was developed by my collaborator, Julia Choi from University of Florida. So we developed and implemented this task in the same patients with the device. And again, I'll just give you a preview. But the key finding is interestingly, during the planning phase, so the pre-leg swing, before they lift their leg, the coherence between, so kind of communication between the premotor cortex and the pallidum seems to predict errors or hits versus misses. Whether they're able to step on target. And then during the swing phase, the coherence between M1, so the actuation, right, the M1 and pallidum. And this is again, low frequency predicts hits or 01:05:01misses. So you kind of have these two different processes. And so as somebody who's processing visual information, they're actively planning. So again, this is a highly engaged task, requires attention, and then they're kind of making finer adjustments as they're stepping and altering it, maybe they're stepping. So, yeah, so it's kind of a more complex task than walking. But again, just showing that these circuits are actively engaged in something like gait adaptation.

Andreas Horn Very cool. And I assume, could like the same, you know, behavior contingent principle that you used in the study also extend to speech or upper limb function or other dynamic behaviors? Is there any interest of, you know, using similar setups that you've now set up for gait to look into the, you know, other rapid movements or dynamic movements?

Doris Wang Yeah, I think you can use this system like for anything, honestly. And then especially if you recruit the right patient cohort, 01:06:02especially the ECoG strip or the surface electrode, you know, we can access many different areas of the brain. I think one thing, as I mentioned before, it's like channel count. Like how many? Like, how many? How many electrodes? And also the invasiveness of it, right? So, you know, Phil Starr has now and Simon Little, they have a program where they are implanting electrodes not on the surface of the brain itself, but just overlying the skull. So, again, it's a paddle electrode just over the skull. You can get, again, the signals dampen, but you can still get some signals like EEG, right? But higher resolution. So, yeah, that might be a way of probing some of these harder-to-reach cortical regions and figuring out the network dynamics and how it controls more naturalistic behaviors, like speech and motor learning. Very cool. All right, a few questions on being a surgeon-scientist. You operate, see patients, run a human neurophysiology lab, mentor trainees, and help lead a focused ultrasound program. How do these roles inform one another?

Andreas Horn 01:07:14how do you get it all into one day?

Doris Wang Yeah, you also have kids. Yeah. Yes, young kids. Honestly, health, lots of health. So, you know, like my lab, I'm completely reliant on, you know, again, my lab is actually very small right now, might be looking to hire soon. But yeah, so it just, I pick people who are collaborative, who work well together, who are good with patience, highly smart, motivated, but who are also very independent. So I just, you know, make it a point, I'm always available. But, you know, they're almost like fully independent, they can run these experiments, they just need a higher level input from me in terms of like, what settings to try and what the overall goal is. So my job is to support these young, talented people and let them 01:08:05run and provide funding for them. And then same thing, you know, the clinical programs I set up, you know, obviously, with surgery, I don't offload, I do everything myself. But I've just tried to develop a way to do things more efficiently, and train people and encourage them to take ownership. So that lessens my need to be on top of every aspect. Same thing with home, you know, my husband, even though he's a surgeon, he is so involved. And we, you know, share all the childcare responsibilities, home life responsibilities. So it's just all about sharing. And then, yeah, I mean, it's tough. But I also like through my training, I guess, learn to function on little sleep. So when it's needed, I can still pull all nighters if needed.

Andreas Horn Okay, well, yeah. Got it. Really fantastic. And you did mention very briefly mentorship, 01:09:03and that you let people run and provide funding for them. That's fantastic. Any other things you provide for people that is important for you now as a mentor, to young trainees? And what's the most important for them?

Doris Wang I think opportunity, like advocacy, and sponsorship. That's what one thing I learned from my mentor, like Phil was phenomenal. Like, right, as I was starting up, I didn't have my own patients to influence. So he let me run experiments on patients. And he still do. And we kind of still collaborate that way. You know, if Simon's running a sleep study, he can recruit from my patient cohort. So I think having this collaboration, and I think it's really important to have that collaboration. And then just always like, you know, if there's a good meeting, tell your mentee about it, provide resources, promote them. If I'm invited for a talk, I can't give, but it's a good opportunity, you know, promote the next generation. So that gets them the buy in. And also, it trains them in a different way. Because, you know, doing experiments, 01:10:04analyzing data is very different from communicating about science. And being an effective communicator, really is like 80% of the game. Yeah. And being able to write well, present well. And, and ultimately, what we're doing is serving the public. So I think it's our job as mentors, to promote that in the next generation. And like, you know, when I'm writing grants and stuff, I always involve them. Like, yes, they help me like get my prelim data figures together. But I always share my grants, how I write, how I think my process. And, you know, mentorship, it can come from different ways. And just, yeah, it doesn't, even though I'm not physically in the lab a lot, and my lab is very dispersed, a lot of times, they're doing patient visits in people's homes, right? And the gait lab, which is the Mission Bay, where, where it's like my physical space is in the other campus. Just being available, like, always available by text, by phone call, they can reach me for anything. And, 01:11:04and I'm just open for discussion and being open to share ideas, you know, being told, that I'm wrong, that some, there's other ways, better ways of doing that. I think that just promotes, you know, collaboration and just a healthier environment.

Andreas Horn Great. If I invited you back in 10 years, and you would still say yes to another episode, what would have to be true for you to say that this line of work that you did, has changed maybe clinical care or has led to insights that truly changed how we think about gait or about the brain? Yeah, what, what needs to happen? Or did it already happen? That, yeah, you would say in 10 years, this really changed something?

Doris Wang I think so. Just because, you know, we know so little about, like, human control gait, and just, like, any insight. And just, like, you know, thinking back about even when I was, 01:12:02like, in high school, what I dreamed to do, right? Like, study the human brain. Yeah. And then, like, maybe develop some therapy that can help it. And that was, like, a lofty, goal. And I think even in this small cohort of patients, especially after the study that's going to be submitted soon, like, all but one patients are now in active, like, adaptive control. So, like, at least in the life of these four patients that, you know, ended up sticking with the study, like, we made a difference in

Andreas Horn their lives.

Doris Wang Yeah. Right. And again, this is small scale, but I think it's just a start. And I think in the next five years, we're going to be able to do that. And I think it's going to be a big, big, big, big, big, technological development. As you mentioned, other companies developing better devices, more robust devices. I think it'll become a reality. Like, I really do think that, you know, our work collectively, not just my own, but as a field, will really transform how 01:13:01we approach a disease like Parkinson's disease. And then perhaps extend it to, like, stroke, neuropsychiatric diseases and others, really transforming the way that we're doing it. And making it really precise and more personalized.

Andreas Horn Yeah, fantastic. I mean, there's so much going on with lots of startups, and I think also new hardware coming likely out, not just from the big players, right, but also from smaller companies, so much going on. I think we'll have a raise of possibilities in a while. Talking about that a bit, and we have talked about hardware limitations and hardware opportunities before, what would the ideal neuromodulation? system look like for you, right? If you could dream now? There's lots of lots of channels, lots of sites, more electrodes. What would it need to really become the dream tool? Maybe for research, or also for clinical? Yeah.

Doris Wang I think more computational power, actually, like the channel. I mean, like, for three decades, we 01:14:04dealt with like eight channels, like for brain side, and it works pretty well. I think it's more like, how do we change these channels and being able to communicate with external devices, as you alluded to, or later, you know, we might not need that many channels to figure out like what someone's doing at all times. But if we have access to, you know, their heart rate, that can like denote something like accelerometer, like where they are in space. And, you know, like our Apple Watch, like the things we're wearing, right, almost every day can give that information. So for me, it's actually being able to have a system, I think, number one, the requirement would be ability to stream this data 24 seven. Yeah. And so first, we need amount of data, large amount of data to figure out and decode what's going on. And then having a sophisticated enough system to not just limited to like spectral power. What if 01:15:04we can compute things like network, coherence, physiology, fake face and cranny? You know, like simple thing, again, without an external computer, right, like on board computational power? I think that can really drive innovation.

Andreas Horn Yeah. Interesting. Makes sense. And I mean, thinking about the size of an Apple Watch, right, and the size of a simulator, it's not that unrealistic to have some even simplistic neural networks processed there for a full day, probably, right? Or Yeah. Interesting. So you put your nickel on, we need to walk around with a computer in our head, in our chest. Right? That's the Yeah, okay. Yeah. Interesting. No, no interest in multiple electrodes. What do you think about these? Oh, yeah, your studies are like, with that, I mean, multiple sites in the brain, right? Which you're already doing with ECoG, but possibly also multiple kind of depth electrodes across the brain. Is there any?

Doris Wang 01:16:06Yeah, I mean, the more we can sample the better, but I think as the clinician, part of me is like, you know, even though they're not that invasive, in the grand scheme of things, I think, to the general public, you know, we have this life alternating technology, and still, like 50% of my patients don't want electrodes in their brain, even though it's tried and true. So I think to make it really palatable to the general public, you know, we don't need to, again, like selected electrodes, we do have to put some in there. electrodes. But in terms of the overall coverage, maybe I'm not Yeah, thinking about, you know, putting like 500 electrodes in the brain implant that, but you know, selected, be you to make it less invasive, but just enough information to serve its function. So do what's needed. But yeah, I guess my wouldn't put my money right now into just 01:17:08more channel accounts more coverage, I think we're gonna get coverage from other things like, right, like electrodes underneath the scalp less, less than directly in the brain itself.

Andreas Horn Yes, yes. Very cool. Okay. So I want to be mindful of your time a bit. I always close with a few rapid fire questions. Okay, pretty standardized. If you have, you mentioned you have listened to a few episodes, so maybe you rECoGnize some of them. What was the true eureka moment in your career?

Doris Wang I think it was the ability to capture signals like stream neural activity in somebody just freely moving.

Andreas Horn Fantastic. Okay. What was a maybe disappointment or wrong turn that taught you something important? You know, it's good to talk 01:18:01about the successes. But if young people are listening, also sometimes good to talk about the failures a little bit, at least if you want, that would be

Doris Wang Oh, I have plenty of failures. Yeah, I always when I talk, right, like in terms of grants, like, you know, when I was starting off, I didn't really like know how to write well. So I had many field grants. But I think one of the biggest lessons I learned is, you know, in more than in terms of like personnel and hiring people, especially as a young guy, just, you know, listen to that inner voice, if something doesn't seem right, or seems like not a quite fit. Either is like patient selection for your study, or, you know, hiring somebody, listen to that voice, we try to kind of trudge through things like you can fix things. I think ultimately, it will save you a lot of heartache and time and resources, if you just listen to that inner voice. But I think making sure again, like, don't overlook things, listen to that inner voice when something doesn't feel right, whether it's the idea person, you 01:19:07know,

Andreas Horn yeah, yeah, no, I totally agree. And hiring is such an important thing to do. Yeah, sure. Yeah. Yeah. Yeah. especially in the beginning when setting up a lab but also probably once the lab is already established i think um yeah i've learned that lesson too um yeah yeah very important point um which person has most stimulated your brain oh Phil Starr okay i would have guessed that but could have been also your parents uh or is there anybody else on you know maybe the non-professional side or um i don't know high school teacher husband whatever um kid that influenced you a lot

Doris Wang stimulating yeah my husband honestly because you know we kind of met during residency he um you know taught me a lot about perseverance he's a spine surgeon even though he has a phd um but you know it's just a very different path in terms of navigating all the challenges and you know doing this together so um yeah he's the one 01:20:05who has stimulated my brain the most outside of

Andreas Horn work was there any paper or maybe idea outside our field that has somehow changed how you thought or gave you inspiration or some historical idea or um yeah anything in that direction where you maybe have translated something from you know somewhere else or yeah realize you could maybe adapt that from here to there um any thought

Doris Wang on that in terms of um i wouldn't say there's like a necessarily a single paper that i can point to or like a scientific discovery but i think um i mean i could say something right now again not related to what i'm doing like in terms of neurophysiology but i think something that really excites me right 01:21:03now is um actually just a body of work about you know cell transplantation and i think that's a lot of work that's going on right now and i think that's a lot of work that's going on right now and i think that's a lot of work that's going on right now and i think that's a lot of work that's going on right now stem cell therapy to treat parkinson's i know something like it's not probably what you expected but i think it just gives me hope that you know what if we can now like really directly precisely determine how these connections are made um instead of just general you know really restore function so like instead of just helping um but restoring function yeah so not not not not only from the physiological perspective but from like a cellular perspective

Andreas Horn maybe let me take that segue into something a little bit still controversial but there is a i think growing voice and i think five years from now nobody would have thought about that seriously but i think there's a growing voice of potentially if we did dbs very very early like maybe directly after after um diagnosis it might have and you know for the listeners might 01:22:03have there's no evidence forever but whatsoever but might have potentially disease modifying properties and i think that's a really big question and i think that's a really big question do you have any thoughts on that like do you believe in that could be a possibility that this could potentially slow progression of parkinson's if we did it very early

Doris Wang i think there is some early evidence that it may slow down the progression to motor like fluctuations that's kind of the on off phenomenon right it just but i don't think it ultimately changes the rate of dopamine cell death per se so and and i think you know right now we're playing catch-up right it's because the dopamines degenerate you're not your brain's not producing the same physiologic amount of dopamine in the brain circuit so that's why we need dbs kind of as a rescue but i think yeah if we implant dbs earlier then perhaps place maybe 01:23:00there we place less string on those neurons to produce as much dopamine to function yeah it could work i guess that's what some people are saying i'm not sure i'm not sure i'm not sure i'm not sure

Andreas Horn That's what some people are saying. I had Jens Volkmann on the show, for example, who has some, you know, evidence for that in The Rat. And I think there is. Yeah, but it's just nobody knows, right? I felt just when talking to colleagues that there is a growing interest in that. And that just means potentially we'll study it as a field one day soon and then find out. But I think right now nobody knows. But it's exactly that idea that if you did it very early, it could relieve some of the stress from the already over, you know, working dopaminergic cells that remain. Right. And it could. But we'll see. I mean, it's speculation.

Doris Wang Yeah. But I think it would be cool if we end up doing that just because then we can study, you know, disease progression. And if it's so early, we can study like what we're studying and recording neurophysiologically puts me more of a putative normal state. 01:24:05Yes. So we're closing to discovering like normal brain function before the disease process really takes place. So I'm all for it.

Andreas Horn Totally. Great. Me too.

Doris Wang Yeah.

Andreas Horn So one question I always ask is what advice you would give a young person? And that could be either entering neurosurgery or neuroscience. You can choose or both. Yeah. Any advice?

Doris Wang Yeah. I think it can be a really incredibly rewarding field. It is hard no matter which field. You know, I feel like it's it's challenging mentally, psychologically, and it's always ups and downs. Even when you look at somebody who's senior, who's on paper very successful, we're still struggling with the same thing, like getting grant funding. You know. Things like that. It's a never ending struggle. But if you find the right balance, and I think it requires a lot of insight into what makes you happy. 01:25:01And for me, I like the short term gratification, which I'm getting from surgery, treating patients, helping their daily lives. But I'm also getting a long intellectual gratification. And that's long and far in between. But when it happens, it could be ultimately a lot more transformative. So just knowing. What drives you and makes you happy and try to the best you can to model your career. And it's never too late to pivot. If something doesn't work and it's just not making you happy, your passion is not going to be in it. And you're not going to produce the best work that you can. So being happy and passionate about what you do is really, really important. And you should always constantly self-evaluate. And this may change over time.

Andreas Horn Sounds great. What is a missed opportunity? What is a missed opportunity for our field? So something we should be doing more maybe, but are not doing enough or something we should be doing?

Doris Wang Collaboration. Collaboration. I think like, you know, as a field where, you know, like pockets, right? 01:26:03Like five patients here, even like larger studies, 20 patients there. And I think, you know, as a field, neuromodulation, like there's so many people doing, putting electrodes in all parts of the brain, following outcomes in different ways. So having a consortium where we can, you know, track outcomes in a standard fashion, getting images in a standardized, semi-standardized fashion, and then just being able to share data, you know, share knowledge. I think that's a huge missed opportunity. And I know like many centers and individual pockets of people are collaborating and working towards that.

Andreas Horn Yeah, no, but it's a, I couldn't agree more. It's, it's still hard to collaborate, right? And to even to share data and to, you know, collaborate. Yeah. You know, sample data. So, but it, it, as a, as a sum for the field, it would be massive if we did it more. I very much agree. Yeah. We'd be so much more efficient, right? Not everybody had to acquire all the data themselves. 01:27:03Yeah. That.

Doris Wang Right. And not repeat the same thing. Right. Like not, we don't have to reinvent the wheel. Yeah. And then, but there, I agree. A lot of things seem to happen, like politically, also ego wise. So, yeah, that's, there's a lot of hurdles. But yeah.

Andreas Horn So Doris, this was really fantastic. Thank you so much for being on the show. Is there any topic you would have liked to discuss that I missed that I know I asked a lot, took a lot of your time, but was there something that you really wanted me to ask and I didn't? Are we all done?

Doris Wang Oh, no. I love your questions. And honestly, it's so fun for me to talk about this, like in depth and, you know, that pontificate about the future of our field. So I really enjoyed it. So thank you.

Andreas Horn Thank you so much. Thanks. Again. Thank you. 01:28:08Thank you.