Opportunities of Connectomic Neuromodulation-Machine Medicine Interview Series
okay we're live um we are privileged to have Αndreas Ηorn uh here who is um among other things a clinician and an expert on connectomics in Βerlin is that right Αndreas where you're based I know you spend some time in Βoston Boston now you are at the moment is that right okay but you're you're sort of uh your origin was in Berlin was it yeah great and uh well thank you very much for being here we're looking forward to talking about connectomics and neuromodulation in particular DBS a prominent and well-proven form of neuromodulation so but perhaps you just kick off by uh telling us a little bit about your background and who ended up sort of in this area in particular yeah so i um uh thanks for inviting me uh first of all i i started um i studied medicine in in southern Germany in Freiburg and um uh became interested in neuroimaging during my MD thesis there and then went to Berlin to pursue that a bit further like essentially started with connectomics so with you know using non-invasive methods like functional MRI but also fusion-based imaging-based tractography to study non-invasively the broader connections in the brain of living humans and then was fortunate enough to to be able to join Andrea Kühn's group at the Charité Berlin where they also did you know deep brain simulation localization so um already back then i thought it would be great to combine the two fields to some degree and did focus some work on you know making precise reconstructions of where exactly the electrodes would line up in the brain with her and then um only once we had that settled uh looked into what are they connected to and that coincided with me going to um for a postdoc to Boston to Harvard with Mike Fox where we looked at essentially that right in the first paper you know what differentiates the patients that would do great after surgery have great motor improvement for example in parkinson's disease and their connections to which sides in the brain versus the ones that would unfortunately not do as great and yeah I went back to Berlin founded my lab there um 2018 and then um now was recruited back as an um faculty member here at the Brigham and Women's Hospital with Mike Fox in the center that he's leading so the center is quite perfectly poised for that or investigating that so it is even called a circuit center for circuit brain therapeutics so essentially with the aim of looking at which circuits do we need to modulate to derive good outcomes or even you know extending that further also you know which circuits would lead to side effects such as i don't know depression in parkinson's disease or cognitive decline or you know um speech problems or so so um trying to you know see the world of neuromodulation in terms of networks and then also using that to help patients in the long run cool all right well let's let's back up a little bit and just talk a little of some of the sort of basic terms here because some people might not be that familiar with them so what do we what do we mean by connectomex this is a term that i think was coined in like 2005 by Olaf Sporns but um but what is it what does it was it come to mean does it have several meanings or what are we talking about yes great point so because i think you know the connectome term is a great term uh or was a great term from Olaf Sporns in 2005 and then now i think it's also a bit used as a hype thing so essentially we have to differentiate just brain connectivity which goes back to at least 1920s of the folks in germany or whatever anatomists of course have always looked at networks and you know people have treated parkinson's disease for example as a network disorder since that time right essentially since 1920 at least or probably longer than that um so not at all new so brain connectivity just meaning two things are connected in the brain and maybe they form a network and a circuit and um really even back then people thought about modulating networks so the connectome term is new and was originated in the in the imaging world and essentially means that we break up the whole brain into parcels and then look at their interconnections and often that's visualized by a graph by a you know mathematical structure from graph theory just you know nodes and edges across them i think that you know probably so and and that has powerful consequences you you could look at you know um specific properties of that graph say um how modular is it or how you know each edge of it or each node of it how central is it to the network you know is it a hub is it and so on so that's really the connectome versus connectivity being a much broader term which just looks at brain connectivity and i'd say that for most of what we do and most people do in the DBS field currently it's more like connectivity DBS it's not connectome DBS although there have been people that looked at it in that connectome fashion where where you would look at really something like how does the centrality of this node decline with DBS and so on so that would be true connectomic DBS let's say I think we're personally my lab is you know looking a lot in connectivity and there are some things that widely overlap with the connectome fields which in the imaging domain where you know for example we sometimes look at similarities of whole brain connection profiles which you know could deserve this connectome term but I think that the more important thing really is what we're interested in is which connection so connectivity you know which connections are crucial for clinical outcomes okay well let's ask another basic question which is like why do we need to worry about this i mean we probably most of us have seen the videos on youtube a patient has their DBS system switched off they have terrible tremors they switch it on and abracadabra uh they're much much better and able to drink a glass of water so why do we why do we need connectomic uh as a dimension of DBS absolutely great question so so for me there are two levels you know we could look at the local level where to stimulate which coordinates like which sweet spot we often call it and which network and you're right that we can explain a lot of variants just based on the local level where we can say um you know hey stimulating this spot is great why do we need connectomics so a few reasons one would be more from a basic science perspective it's really helpful to know which networks are associated with that sweet spot right so just to learn let's say it's not really that we need to modulate the STN but we also need to modulate its connection with pre-motor cortex so so that will help us better understand let's say parkinson's disease or obsessive-compulsive disorder as a disease so we can essentially even use DBS to investigate the functional connectome of the brain right as a tool because if you have such a graph and you can modulate one node of it you can see what happens in the network that's really a powerful tool for basic science reasons now for clinical reasons i think there are there are also great examples where the connectomic part really shined you know for example we could show in obsessive-compulsive disorder that that's a disease right um from the neuropsychiatric domain and um and there are different target sites that have been discussed which would need to be stimulated one is the subthalamic nucleus again one is the anterior limb of the internal capsule we could show that in both of these targets the same tract that connected to two was important to modulate right we could even use connectomics to cross predict across the two targets they built a model of optimal connections just based on one target one cohort operated with that one target and then predict the ranks of of the other arm right of the other target so so that way we could you know that was maybe the best so far the best demonstration why connectivity could matter because we could not only show this is the spot for one target but we could show this would be the network and maybe you can stimulate it at different sites but let me be a skeptic and say what you're describing sounds like a result in which you demonstrated statistical significance what would you say to the skeptic who said you know we've yet to see evidence of clinical significance so but i mean it is i i don't know what what do you mean with clinical significance so it was a retrospective study that's true but we did predict ranks and clinical outcomes right so it was yeah it was not just that you know the same track is modulated but we could show that it's crucial to modulate that track these patients we have investigated right i see so you see we need prospective trials to validate that and i told people yeah i know i think you're you're inferring i'm rather more sophisticated than i am no i was just i was just trying to push on that uh question of i know um i guess it's a more general point that it really is true that we very frequently still see patients that have poor outcome from dbs right in OCD but in movement disorders as well i mean the guy the guy that i mentioned on the youtube videos right they never show the patient that had a bad response right they always show the guy that had a miraculous you know a miraculous recovery of function so so and that's good yeah so so some hope is that we can you know so you're totally right maybe maybe that that is your last question some people already do great why do we even need to investigate it more but so i think you know maybe right now probably in parkinson's DBS just gut feeling why is uh 90% will respond but only let's say 30% will be these excellent responders and then my aim is to make that 30%, 90% you know or try to and that that could be done with connectomics but also just with good imaging so so it's you know both levels the local one that's important but then also the network level and um trying to get more deliberate and and trying to probably we we won't be able to improve the 30% percent even further with what i'm doing but you know we could make more people top responders essentially do you think do you think uh clinically speaking you think those 30% are are getting as as good a response as it's possible to get yeah so i would think as as as as good as possible with what we're currently doing with this classical 130 hertz always on DBS i think to improve those further we would need something new disruptive which could be adaptive DBS so far that you know it doesn't look like that either that you know patients would really get better than with continuous so adaptive DBS it would be that the system would listen to the brain and then only modulate based on specific rules that and we've seen we've seen the the now the release of the first commercial device right the Medtronic uh percept system so that's a that's a reality now but you're you're saying that you know we've yet to see real evidence that's the game changer or is going to be the game changer that we had so even there i would i would hypothesize it won't change these top responders even further in terms of their motor outcome but it could maybe help reduce side effects right that that the system isn't always on could mean that it's only on when really that effect is needed but then um it's often maybe not detrimental to some other function like speech or or whatever um if if it's adaptively off when not needed essentially so so yeah i think with the current technology we have it will be hard to improve these ultimate top responders because they often already go down to nearly no symptoms anymore right for for a while and then parkinson would progress further unfortunately and uh that is probably you know the nature of the disease is neurodegenerative so we can't restore that with DBS at the moment yeah yeah okay that's really interesting um this is another i think important distinction that we haven't really gone through and that is the kind of you know as i think you say in this very good review paper opportunities of connectomic neuromodulation and you mentioned a little bit in your sort of preamble actually people have been thinking about networks uh for many decades if not centuries um but but what's really changed is the is our ability to characterize them and there's i think you mentioned in that paper two main modalities i'll let you sort of explain is that right functional and structural might be there yeah so so now that there's video here i can even show a book um from uh that's from 1950 and this is um from uh creek Edward Creek now and and and so you know this is this is actually from that time and this is a connectivity of the brain right in in colors even with you know it's beautiful stuff so yeah so it's really important to always highlight that that brain connectivity as is as old as neuroscience so so it's nothing new under the sun there i think what we came up with or people came up with um in in the last two decade decades would be non-invasive ways to characterize connectivity and in the living human brain and that is very elegant in you know from a physical physics perspective but then also i would even say you know these old methods were much much better but they are limited to post-mortem dead brains especially or animal studies and so on so so so um you know i think now we are at the situation where where the ground truth is still in the anatomy books definitely and we have a means to to derive some sort of personalized poor man's version of that anatomy textbook that fits our patient and that that's a great advantage i think and now the other thing with really connectomics right with the idea to parsilate the brain into like a whole brain connectivity profile that is also new and it might just be new because of the computational resources we have right often that involves having you know 20,000 voxels in the brain and having a 20,000 by 20,000 connectivity matrix so that was just probably not feasible way before the 2005 Sporns um paper at least not on a laptop you know so so it um I think I think that is that is probably just new because of that that we have the computational resources to create connectivity profiles across each and every voxel in the brain and I think you mentioned the two methods so one is functional MRI that is an indirect method of brain activity uses um it's called the bold contrast the blood oxygenated um level-dependent contrast and it essentially uses the fact that haemoglobin is slightly different in in terms of its magnetic properties whether it's oxygenated or not so it essentially uses you know um brain uh so brain oxygen and blood flow ratios to derive it which area is active in the brain so if you have a you know human seeing seeing something a flicker board of light then the v1 reach in the primary visual cortex will light up and we will be more have more of that bold signal there and then people have come up over the years with with good risk resources looking at that now resting-state function MRI would be to look at the co-fluctuations right if the two parts of my to motor court disease of mine would co-fluctuate and and their signal would go up and down in the same way they would be just correlated in time that would mean what we call functionally connected right it's very far from a real axonal type of connection it is more a statistical rough guess that these regions do something together but what they do together we don't know right so the temporal resolution is also really slow in fMRI so it all boils down to you know it's a poor man's version of connectivity um but it works in living humans which is amazing and um you know we can we can use that to estimate things right so yeah very derived method the same same way derive this diffusion imaging which measures water diffusion in the brain and has the aim to rather look at accents or like anatomical connections again we won't see the axons they are way too thin we will see the very big highways in the brain right so the i don't know um state trooper highways um that that you have here in the US so that's what we see and again so so so there's and again that so so maybe just briefly explain that would measure that diffusion or exploit that diffusion would diffuse slightly more perpendicular to the axons rather than sorry you know in parallel to the accents rather than orthogonal to them right so so it would exploit that fact and based on that there are algorithms and tons of algorithms to reconstruct the tracks um but yeah it's an important point that especially if you just scan a single patient in a clinical setting and you derive their connectivity either with the two one of the two methods i outlined it will be so coarse and you know poor resolved and um and so on that it's even a good question does it really help us for deep brain stimulation because in deep brain stimulation millimeters really matter being two millimeters off target might you know result in an only 30% improvement rather than 80% improvement right so it really matters and with the two methods we we have trouble seeing these fine details and we also have trouble fractography with the diffusion one to to see the very thin bundles in the brain and so on so so my lab has focused a bit on rather than using patient-specific data to use big cohorts of of brains essentially that were scanned plus also histology data or you know other types of like more textbook connectivity data to infer with our patients we know where the electrode is we know where the structures are that way we know what they are usually connected to when we exploit that yeah but we've done both i mean looked at patient-specific data but also then more high-resolution atlas data and is it possible to combine the two the two forms of data in the same patient so i would love to do that that's one of my you know focuses for the next year is to find good ways of emerging information from a high resolution even the postmortem connectome that was scanned in the best centre in the world probably here in Boston Martino Center also with the patient-specific scans and then you know individualize the high resolution data so i think that's a definitely future topic that's fruitful yeah so then you have best of both worlds you have a patient-specific high-resolution somehow but so it's so sort of interesting so with this structural connectivity we get a kind of like you said a highly imperfect but um sort of as it were solid kind of direct measurement of the connection albeit a giant highway um uh whereas but it doesn't really give us any kind of functional information about you know that highway could actually have no cars on it right nothing it's possible that nothing's going up and down it um but yeah possible yeah i agree with you yeah absolutely but that's kind of we have it gives us it gives us anatomical kind of uh a reconstruction rather than a functional reconstruction whereas with the fMRI we've got these two areas um and we're looking at the bold signal if they co-vary and we can say there's a statistical relationship with them that could be because there's a highway between the two but it could be because there's another area that they're both connected to or even some other you know two but several moves so that a and b are connected to c and d by each other and then c and d are connected to uh e and that's where the so yeah so that but that maybe to jump on that i'd even say that the word functional connectivity for fMRI is probably a misnomer they see them more as co-activations and i think that's the better term you know they likely are in the same network if they're co-activated but yeah it doesn't mean they are they have to be connected and don't aren't there other techniques like for example if you're able to predict the the behavior of one area from the behavior of another area immediately preceding it it kind of maybe gives you a slightly stronger kind of uh yeah so few people have brought that like Granger causality or modeling which is a more complex model based derivative to with a model inversion step where you would solve these things on a on a neuronal code level so essentially trying to from the blood level um go back to the neuronal level infer that and then on that level create the um the graphs or the directionalities of the connection and people have then termed that effective connectivity yeah and i'd say for for tasks and especially in the case of DCM this has been worked out quite uh quite impressively and is really nice i think what just with the lag stuff like Granger causality where one signal slightly precedes the other one you know i'm no expert in it but i'm i'm not convinced because the signal is so slow and i've just heard experts actually in the in the last uh i think OHBM podcast that we're all so all not convinced by that so so i'm not alone in that i think i think um you know they're such a slow signal so we're really thinking about you know each data point in an fMRI signal is usually two seconds from each other right so a curve would be 10 seconds at least right so so measuring the the lag between two of them will result in a lag of maybe a second or so that just isn't brain connectivity right that so you don't see any of that these techniques like Granger causality or or dynamic causal modeling you don't see these things impacting clinical care in the near future so so i think DCM more than so i wouldn't you know lump them all together as one but in DCM there's one really nice study published in brain from the london group where they had deep brain simulation uh like scans on and off deep brain simulation i think it's quite old already from probably 2012-ish or so 14 maybe um that where where where you know they then looked at um DCM to look at which pathways of the you know indirect and hyperdirect pathways on would be modulated by deep transformation right um George Carharn is the first author and as well so it's a nice paper and i think that type of analysis could could help us probably not clinically but to understand what's happening in the brain um so i think you know they have so the whole group with DCM they have such a you know vast body of evidence that this seems to be working if you really know what you're doing and so on um so i wouldn't say DCM is you know not useful but uh yeah no i don't want to speak ill against the great Priston um but um yeah but yeah okay that's cool so so how do you kind of i mean as you as you say in your in your paper as well you know this it does seem like a very promising uh sort of plank in in um uh for for sort of in the overall sort of support for clinical what do you see like i say five ten years down the line if kind of we get a what your work in connectomics and other colleagues in connectomics um works out how do you see it kind of affecting the as it were the patient experience from being referred to a movement disorder specialist to actually getting DBS how would it kind of impact what is this just going to be 15 minutes in the scanner or they're already having scans right so an extended scan or or is it going to be more lots of great ideas so so i think what what what we're trying to do here next um which is not the distant future but sooner so so Mike Fox has just published one paper in brain just came out a few days ago with as a first author looking at networks that would impact cognitive decline based on deep brain simulation and then we had a paper and annals i think in a few years ago which would look at depression as a side effect following in parkinson's disease so which connections would lead to depression so looking at then you know now maybe inviting looking at our database here locally and looking at which patients might have that symptom and then reinviting them to the center and trying to reprogram them that would be one like sooner avenue that we're trying to look into here but i think long term you know my vision would be to have symptom-specific network profiles let's say one for tremor one for bradykinesia one for rigidity but then also for depression and cognitive decline and all that mapped out in a normal brain as a library right so not patient specific but they are essentially maps yeah that we can use and then if a new patient comes in we could check their symptoms scores let's say they have a lot of tremor we would then wait the tremor network more strongly to to to plan their surgery and then also to program them and let's say they you know they they um we don't want any patient of course to get depressed after surgery but maybe for patients that already have that type of problems before that's an even higher you know um thing we want to avoid even more than in others that maybe are really you know so so we could wait for each patient the profiles we have and find their optimal mix or blend of networks call that network blending and then and then I think based on that do surgery but also DBS programming and i think that will be the future um i really think so there will be a lot of other things in the future as well like adaptive DBS and so on but i think this will certainly evolve to be ripe for clinical practice at some point and then i think an intermediate step that we haven't talked about now would be we have the libraries and we have what the patient has in terms of symptoms we want to to scan also their brain and their networks and match their networks to the library networks right we want to see we we know the library network but we want to see that in the patient in individualized form so we would scan them before surgery we would you know they would get a resting state fMRI and a GPI scan best we can do in a single patient and then we would segregate their brain into the networks we know would respond to tremor bradykinesia and so on then do the network blending on in their own brain and of course have it all individualized i think that will be likely my program for the next i don't know five to ten years to develop that further and then at some point also go into you know clinical trials with that hopefully but um and would you see this being a tool that's kind of delivered through with the kind of the DBS providers kind of a thing or long-term yes yeah at the moment at the moment it's research right of course but i think i i would really see value in that becoming commercialized as well and um you know but first of course we have to show it's robust it works and you know it's a long way probably five to ten years is probably wrong it's probably my whole career or whatever but um it's going to take a while to make that transition to a truly you know connectivity-based DBS uh can connectivity inform DBS do you think it will be sort of what other what other i mean i think it's always interesting to consider kind of like how a disappointing kind of genomic studies in this area have been right and maybe that's because genomics have nothing to tell us about people's subtypes of disease or maybe it's because we don't know how to analyze the data what do you think yes so i think the audio glitched a bit but what i understood was that the the omics or the genomics um wouldn't be helpful or uh or so far if you look at the literature you probably know it uh better than me but i looked at it relatively extensively about a year ago and the results generally seem fairly disappointing a couple of weak associations with sort of this gene of that that gene yeah um but really kind of you know the idea of kind of you know because one would hope you know maybe following a similar logic to that that you're following connectomics there would be a rich seam of information that would allow us to subtype patients genomically and it wouldn't be the whole story we might have to but there would be whereas whereas really what i found was a couple of weak associations with a few genes much less impressive as a story than what is going on in in connectomics and i wonder you know why is that is that because genomics really is is not that relevant for DBS or is that because how to analyze the data yeah it's a great question so so so i i am totally on board that you know no patient is alike especially also in their and their um genomics and so on so that that each parkinson's patient will have to some degree a different cause so there's great work by alberto espey that i interviewed for my podcast as well and um and ben stetcher who's a patient but also patient advocate that's really knowledgeable about these things so they claim this you know if you have twenty thousand parkinson's patients you'll have twenty thousand diseases right so everybody's different and so on and i i i you know i am not an expert in that and i wouldn't even disagree with that i would still think though that there will be especially with something as simple as DBS be some sort of common end pathway for for a lot of patients maybe not for all of them but for maybe 90% at least that would still you know the same networks would be affected likely because of different causes or you know in some patients especially would you know one network would be affected more than the other and so on but to that that's exactly what we want to do anyways but but then you know um how much the genes play a role in that i would say that in in parkinson's at least you know um that hasn't been investigated as profoundly in the big cohorts we had for a good reason because there wasn't any evidence that some subtypes would not respond to DBS right i would say most that would respond to levodopa as well and most do right most let's say genomic variations of PD um they they would also usually respond to DBS if it's placed well yeah it's very different than dystonia i think where people have really thought about or looked at you know some subtypes would just not respond too well to DBS right but in PD that didn't seem to case so far in history so so i'm not saying it doesn't matter at all i i think it will be it will certainly matter to some degree the question is how much variance will genomics explain and you know then the other point is if we wanted to include that as well as a certain like additional thing into these types of analysis you just also again need more and you know in bigger studies more deeply phenotyped studies and more expensive studies and so on so um I see practical problems there but rather than it should be worth it what are what other data types which do you think will be relevant that are not being i mean it's it's amazing that kind of there was a i'm just trying to remember where it came from but it was a it was a study based on a questionnaire that was sent to several hundred DBS centers all around the world and basically they were looking about looking at um the pre-surgical workup and and what procedures are being done and what's not being done and and even even the dopa challenge is only being done at about about 90% of places but that's far far and away the most common one the next thing is like 30% of places or 40%of places are doing you know a formal cognitive screen and then it just goes down and down and down right the psychiatric screen a bit less and so there's it seems like there's enormous heterogeneity and and there's like you know reasonably good element evidence that any of these things are somewhat informative but but no apparent kind of consensus about how we should actually work up a patient great point i i still think the answer is somewhat close to what i just said that it works so well in most patients right so they're usually if things work well there's no not as much incentive to change in uh or to even sometimes to better understand them because it you know um yeah of course we want to understand it but yeah you know if it works you know what never change running uh winning system still that being said i think there's some cool work for other biomarkers pre-op um by the group and also others but uh looking at atrophy patterns so you know you could i think they could show that if you have a lot of atrophy um in the SMA in the supplementary motor area then patients would not respond as well to DBS on average and the reason could be that that's where the hyperdirect pathway originates that goes into the STN and maybe if you know the disease has progressed already as as far um as that that there's a lot of atrophy it's just not you know we can't revert the network changes as easily with DBS right so if i were to include additional things into such an analysis i would probably use atrophy patterns as well and that would essentially just be MRIs right so we would use again the structural MRIs for that um right yeah and then i think yeah uh levodopa response there is some doubt yeah it is certainly a good predictor or people think it's a good predictor that's the consensus there there are still there are some papers that showed you know that um uh you know it might might even sometimes be a statistical artifact so i'm yeah i i don't even know how how informative it is personally even definitely word on the streets as it is very informative and i think there's there's been one paper from from Grenoble where they looked at like really long really long term outcomes just out in recently um i and i think they also came to the conclusion that at least for long-term outcomes the levodopa challenge wouldn't even be that predictive but there were different metrics um and i i can't tell you which ones of the top of my head but they analyzed this in a i think so far the longest period so sometimes going up to 18 years or so yeah um yeah i see so um yeah very interesting so um so finally i i just want to push push up on one other thing which is i think is also very interesting which is that um in your paper there was a very interesting discussion of uh you know particular networks being associated with particular conditions such as PD or essential tremor and then you know there being multiple possible sites that you can that you can address and they may have slightly different symptom profiles but basically you could treat a disease in principle i guess by by addressing any part of that network because it would have a global effect as it were on on the rest of the network and then one of the things you pointed out was that well for many many sites you could access the network either transcranially with a non-invasive approach like transcranial magnetic stimulation or invasively with an invasive approach like deep brain stimulation so do you think do you see the the future of neuromodulation more generally sort of ultimately emerging from intracranial work to essentially being sort of programming nervous systems from outside yeah great point so that that is something um my mentor Mike Fox has explored way more than myself so he has a great 2014 PNAS paper out of that that where he really looked at TMS sites versus DBS sites in different conditions i think in 20 conditions or so or around that number and look it could show that resting state networks would always in each single one you know case connect the two sides right so what we modulated up here would always in all of these cases being you know connected to the to the side that we would modulate in the subcortex with DBS so that just you know made clear or at least suggested quite strongly that we seem to be modulating resting set networks so our brain networks and we can do so from outside or inside um so i think it's a very powerful approach i think practically speaking one downside or one difference between something like TMS or tDCSor the outside methods with the brain simulation is that the effect sizes are usually smaller in most conditions at least and then it's always also sometimes not long lasting right so that's that's so far unresolved there's great work now also from our group so from Mike Fox's group here with the crash and TMS size but also Nolan Williams has a founded company and has a good trial where where they have found ways with TMS have you know much less stimulation um time points needed for the same effects um with you know a novel paradigms and they could also use the fMRI networks to to inform where exactly they would need to to modulate so they've really tried i think had great results in the depression field so i do think you know we should do that more for parkinson's as well um look at you know which sides um we could modulate there has been some work in the non-invasive fashion as well sometimes even with good results but usually just not long-lasting right so bradykinisha would come back after i don't know 30 seconds again right so you would see some effect but it wouldn't last so yeah there's some some interesting work on the horizon i think uh there's a trial that we're involved in with uh called the STEM-PD trial that um is looking at uh caloric vestibular um uh yeah with the scion company uh yeah that's that's that's very interesting as well yeah that's really not understood as far as i know right so so far but but it's really promising well like like the whole field right well you know people always say we don't know how DBS works but I would think that's not true we do i mean we at least know you know a lot of things about DBS we know it it essentially stops information flow right and it's probably it has to be pathological information so it will you know all the literature on the beta power signal um you know there's a lot we understand it's not that we you know just put electrodes in the head and it works and um i would say even back in the day that you know the lesion surgeons that did it in the 50s or so even um they they had really good theories of why they lesioned exactly palliative and so it's not that we don't understand how it works right and with the with the um cochlear simulation so far at least i've talked to the companies well that um so far it's great results but it there i think their model currently is that it would modulate the whole brain somehow right so it's a wide connection of the vestibular nucleus and um that that is you know it could be that's the whole all it takes to modulate you know everything essentially but but i think there's no better or no DBS level of understanding right so and that being said i still think there's a lot to learn in the DBS field as well so it's not that we understand it but we it's not that we don't know how it works at all right i would say yeah that's good do you agree with europe you'd agree that we're a long way away from able to be able to build a sort of uh in-silico simulation of a patient and and predict how they'll respond but i guess you know your work is contributing to that to that process um yeah there's a very well it boils down to you know we we don't understand how the brain works right that's the main issue here yeah and i think with that you could say we don't know how anything in the you know like MS or whatever you name it we don't we wouldn't know we we we we also don't know how parkinson works we could say dopaminergic neurons degenerate but yeah does that you know generate parkinson's so so i think we we have a similar level of understanding of of how DBS works as how parkinson's work right so i would say so and it's not a very detailed one but we have some clue so yeah now there still remain fundamental mysteries right about how neuronal communication and computation takes place so that's absolutely not not that surprising at all Andreas it's been a pleasure um and very interesting and you know i really enjoyed your uh review paper with um uh Michael Fox so you know thanks from me and for to everyone else because it's a great great contribution and i'll i'll look forward to following uh your work in the future with interest maybe we'll be able to talk again in a year or two and and see where we are in terms of uh progress great thanks so much all right cheers
2022-02-03 07:38
