Chapters Transcript Video Clinically Relevant Biomarkers and Outcomes for Motor Recovery after Stroke David Lin, M.D., presents at the Johns Hopkins Department of PM&R’s Grand Rounds on September 15, 2020. Thank you very much Tracy. Um uh can everybody hear me? I'm I'm gonna take the silence as Yes. No, there's no shaking heads. Yes, so thanks a lot Tracy for the introduction. Thanks to everyone for inviting me to give this talk today. Um I'm a newly minted faculty member at MGH I started in july uh and um I did a neuro critical care fellowship. I'm as well as I know rehab fellowship. So I'm three years out of my neurology residency which I completed in 2017. Um I'm gonna tell you a story today about sort of work we've been developing around stroke recovery over the past four years at MGH We've been collaborating with folks at Hopkins specifically um on pretty Raghavan who I think uh we'll join the call when she's at a clinic. Um And so the story sort of I'm going to tell you is as much sort of an origin story as it is a science story. Um I have no financial disclosures. Um so the outline for today is I'm first going to talk about stroke recovery as a health systems problem, which I think many of you in the PM and our field will be familiar with. I'll talk about outcomes of them as well as biomarkers um for for upper extremity store recovery. And then I'll talk about sort of how we're thinking about the future of stroke recovery. I'm in terms of treatments. So uh this is uh the view from the MGH hella pod. Uh This is a helicopter out in the distance. Um This is a helicopter landing on the helipad And this is me while I was a fellow in acute stroke call, this is a patient who is being brought from New Hampshire to MGH. Um And as I'm sure is also the case that Hopkins MGH is a is a major coordinating center for acute stroke. And so many of the patients that we receive um have received T. P. A. Actually the outside hospital and are flown to MGH for consideration of endovascular therapy. And so um that's where a neurosurgeon or neurologist goes in with a stent retriever to try to retrieve large M1 clots. And so as this crowd, I'm sure is familiar with the last 10 years. Even really during the time of my own training has been transformative for acute stroke. From the advent of T. P. A. In 1995. The N. I. D. S trial to more recently in the last in the last decade really 2015 and forward endovascular treatment where large strokes can be removed has been really transformative for this field. It's changed the way that we think about stroke systems of care. It's changed our metrics etcetera. But the truth of the matter is as as you guys know, only one in 20 patients are treated in time for acute stroke therapy and um as a result of that stroke remains a leading cause of disability in the United States. Despite rehabilitation, a significant number of stroke patients remain severely disabled. And I'm going to talk about our motor impairment as a sort of a model for thinking about stroke recovery? Um because it's seemingly, well the first is that it's a major source of disability. And the second is that seemingly it seemed like a relatively straightforward quantifiable model for thinking about recovery after stroke. It turns out that it's it's complicated as we've learned in the last many years, but but we thought it was that, but we thought it was a good starting place. And In contrast to the sort of transformative advances in acute stroke therapy really are treatments for stroke rehabilitation have remained limited. And and really in the last 50 years, the kind of cornerstone of rehabilitation after stroke treatment remains pt Ot slp and and novel therapies sort of have not crossed the borderline. And I'll talk about reasons why we think so um in in later slides. And so the fundamental question that that drove uh much of my research program over the last four years is the question that that most that that all patients ask when they come to the acute stroke ward. But as neurologists, we feel that were least equipped to answer, which is how am I going to recover after my stroke. So for patients that come to our miller fisher service, which is our acute stroke service um and and on the ward, they asked the question, you know, with regard to arm function, for example, will I be back to everyday life? Will I be moderately disabled, or will I be severely disabled requiring help all the time and and and and in a nursing home. And I would argue, I mean there have been uh substantial efforts over the last many years to be able to predict to be able to predict recovery after stroke. But by and large our ability to predict recovery in stroke and more broadly in acute neurologic disease remains a black box. We're not very good at predicting and prognosis in our field. Um And I think, you know, to talk about the health system challenges that this is a challenging scientific problem. Inasmuch as it is a challenging health systems problem because as as you guys know patients after acute stroke have many transitions in care. They go from acute acute stroke hospital like MGH two inpatient rehab like Spalding, which is uh which is our local rehab across town uh to home and then to outpatient clinic and traditionally across these different settings of care. Uh there isn't a quarterback that really follows these patients. Um And so to address that we we um in fall of 2019 launched in a AJ sponsored an AJ sponsored workshop in which we brought together a multidisciplinary cloud crowd including fizzy interests and neurologists and speech occupational physical therapists, we all met at the boston um at at the Copley Marriott in boston. And basically we went through a series of exercises to brainstorm collectively what are challenges that face the stroke recovery system today. And one of the things that came out of that which we're currently writing up is is this word cloud and as you can see many of the words that appear large, which means that they have a greater number of hits relate to health systems issues rather than science issues. People identified access and insurance and fragmentation of the post acute stroke system as a major problem. And so really to address this um at MGH starting in 2017 when I was finishing my training as a as a neurology resident. Uh we we basically launched a research study which became kind of a flipped care model for for recovery after stroke, in which um we we we we understood that before we could um create new therapies to help people recover. We first needed a framework to understand how people recover and why people recover after stroke. And so what we did was we launched a clinical research study in which all patients who enroll in this study are enrolled in the are enrolled in the acute hospital setting. We focused on patients with armed weakness after stroke as a start and we brought patients back to our research study at six weeks, three months, six months in a year. And in parallel with that, we also developed a new clinic which is uh which which became a model of care in which we're seeing our stroke patients to provide what they need from a stroke secondary prevention standpoint as well as a focus on their recovery. Um And so and so we see patients in the hospital six weeks, three months, six months in a year. And uh what the research study does is that we collect a series of outcome measures that quantify their recovery over time. Um and it took a multidisciplinary team to do this. And so so that's our first kind of major point is that this this kind of clinical research effort really takes a lot of buy in from a lot of different divisions and departments and I think at MGH one thing we've been successful at doing is is getting people excited about the mission of taking care holistically of stroke of of of patients with stroke and focusing on their recovery. Um And so we launched our research study which is called Smart for boston with the H. Um and essentially what it was was that we're focused again on collecting outcome measures over time for people with upper extremity arm weakness starting during the acute hospital stay to date over three years with a little bit of a break because of COVID, well, we've enrolled 140 patients. Um I will say that the measures that I will show you in a second are time intensive. They include the feudal meyer upper extremity and a lot of this has been done in in person to to date. Uh And we're now sort of rethinking in the in the covid era, How do we collect a set of outcome measures remotely that would be meaningful for the field? So uh the the the primary goal of this study which was broad to start and this was at the time which I'll talk about later when um uh when Jon Krakauer and others were also thinking at Hopkins about how people recover after stroke. And uh and so we we at a very broad level, we just wanted to understand what our trajectories of recovery after stroke, what determines whether someone is going to start out nearly, please tick and recover very well versus recover moderately versus have no recovery at all. And so uh to do this, we we started to think about well what outcome measures do we need to collect and as a neurologist uh sadly speaking, um you know, as a neurologist, we we don't we don't spend a lot of time thinking about outcome measures after stroke nor about recovery really traditionally, and I spoke to one of my mentors at U C. L. A. Steve Kramer and he said, well, you know, if you want to collect outcome measures across stroke, one thing that would be relatively novel is if you started to collect outcome measures that span the I. C. F domain and I said, well what's the I. C. F domain and that's that's sort of, you know sort of embarrassing from a neurology perspective but I think that's one of the shortcomings of our of our field and that, you know, as a neurologist in a very different way from P. M. And R. Um we don't learn about structures for thinking about disability and so when we designed this study we collected outcome measures that span the range of the international classification of functioning, going from impairment to activity to participation. And this is just a list of the outcome measures that we started with with including the few Ghimire grip strength and I stroke scale for impairment blocks and blocks and nine hole peg and Barthel index and the ranking scale for activity as well as for participation. We ask questions how they were doing in their everyday life over time, over time. This study has broadened as we've gotten continuing input from our occupational and speech and physical therapists and we've brought in our outcome measures to include measures of gate as well as measures of cognition because we realized that upper extremity motor recovery doesn't occur in a vacuum. Uh and that in order to really understand upper extremity motor recovery, we really have to start thinking about cognition and and and and other aspects of recovery. So uh this again this research has taken a lot of time to build the team to be able to do this and I think that's that's one of the first points is that people collectively at MGH have recognized that this is that, that stroke recovery, the health system problem. And that if we could collect outcome measures, we could start to address that. Um And we wrote up this paper in um in a journal that was published in 2019, that basically just talks about the challenges of of collecting standardized outcome measures over time, specifically on the acute stroke board when um when patients are, you know, cognitively impaired and sick and when occupational therapists and speech and physical therapists have demands on their time. Uh and so this is a person in our clinic who who we did the feudal Myron. Okay, so uh that's broadly speaking, just to say that, you know, health systems is a problem in recovery after stroke is as important as challenging as the science. So now I'm gonna talk a little bit more about the science, outcome measures and biomarkers. So, as as you guys know, there's a rich science behind recovery after stroke that spans cells two circuits to behavior. So, in acute stroke, uh there are significant changes that are happening within minutes to hours after stroke that spans cell death and inflammation to changes in network function to rapid changes in behavior in stroke. And and this kind of uh we learned a lot about as neurologists in stroke recovery, there's also a rich biology and I would I would argue that at a fundamental level when we're talking about stroke recovery were fundamentally I hope talking about biology. And so at the cellular level there's changes in blood vessels and neurons and synapses at the cellular level there's shifting cortical activity and and at the behavioral level there's a longer term changes in behavior. And uh one of the things that we have to stress neurology to sort of our colleagues is that uh stroke recovery and acute stroke are different and stroke recovery is different from uh is is really if you're targeting the biology of certain recovery you're you're targeting an inherently different biology. Um And so this is a this is a diagram from Tom Carmichael's lab at U. C. L. A. Who who runs a stroke recovery animal lab. And this is just there's a lot going on in the slide. But this is just to say that in the biology of stroke recovery which this is not to talk about, there's there's rich changes um that that occur um and at the cellular level at the blood vessel level that are that are representative of what happens during recovery as well as what happens in in in terms of cancer programs as well as what happens during um Mhm normal development. So uh this is one of sort of my favorite studies to talk about because I think this is one of the unequivocal demonstrations that plasticity um occurs in the brain after stroke. And uh and one of the unequivocal demonstrations that rehabilitation actually affects plasticity. So, these are classic studies from Randy Nudo that were published in science in 1996. And and when I think that when we when we think about neural plasticity as a field, many of us think about these papers. Um And so what they did in these papers was uh they induced a stroke in a monkey in the middle cerebral artery. This is showing the artery, this is pre infarct. They included one of the arteries and then they mapped out with electrical stimulation. Um The the the different parts of the primary motor cortex that represent different parts of the upper extremity, including the digit, the wrist and forearm etcetera. And basically what they showed was that if you compare pre to post infarct, there is a change in the representation of upper extremity after the infarct, meaning that parts of the upper extremities that's observed the digit shrink and parts that's observed more approximate parts of the upper extremity grow. Um And then what they, what they went on to show in a separate science paper that same year was that if you give the monkey rehabilitation increasing, which which in that was basically moving the the monkeys affected arm that you could actually preserve parts of the motor cortex that that that that are subservient um by parts of the upper extremities. So the parts of that the primary motor cortex that uh that are used for the digit after rehabilitation did not shrink as much as they did if you did not give the monkey if you did not give the monkey rehabilitation. So I think again this is and these these studies and monkeys are unequivocally showing that that stroke changes uh cells at the level of the motor cortex and that if you give rehabilitation that you can preserve the the representation of the motor cortex. Um And you know there's been uh there's been decades of research that has happened since then and I would argue that much of what we do since then with very fancy tools like F. M. R. I. And identity, E. G. And TMS etcetera are basically a proxy for that in that we're trying to map changes that are happening in the brain through different signals including bold and E. G. Et cetera. But really what we're what we're really interested in at a at a plasticity level are changes in the map of the motor cortex and other and other parts of the brain. And so uh this is FMR I research this slide summarizes a lot of fr FmRI research that has happened in the past in the past two decades. And basically one of the themes that people find when when doing FmRI in patients with stroke is that when a when a healthy control attempts or moves um they're right upper extremity, the left, primary motor cortex lights up. But when a stroke patient who has had who has had a stroke of the left side tries to move the right upper extremity, the activity shifts more toward the toward the absolute regional arm, meaning the arm that's on the same day, the hemisphere that's on the same side as the arm moving and the timing of this shift shift. The the reason for this shift and the role for this shift is debated whether it's actually adaptive or maladaptive for recovery is something that's hotly debated. Um And uh and and but but what what people have shown is that the amount of shift that happens from absolute national to con traditional hemisphere is directly related to how much of the cortical spinal tract, which I'll talk about later sides is damaged. So meaning the more severe the stroke is affecting the arm, the more that activity shifts towards the other hemisphere. So I'm not going to talk about these concepts of biomarker and outcome and recovery and how I think about them. And I think it's very important in our field of stroke recovery to be very clear about what we're talking about here. So um so an outcome measure is sort of how a person is doing at a cross sectional poison at that point in time recovery, if we're talking about really measures of delta in my mind. So it's measuring a change between how someone is doing early on and when we measure the outcome and then a biomarker, which I'll talk about again in a subsequent slide really is. Um It has actually a strict definition by the NIH, but really is uh trying to map on a biological process to a recovery process. And so the NIH convened a workgroup in 1998 and basically they defined a biomarker as a characteristic that is objectively measured and evaluated as an indicator of a biological process. Um so there there are different types of biomarkers. There's everything from a monitoring biomarker to a diagnostic biomarker, a response biomarker to a predictive biomarker. And I'll talk in subsequent slides about in recovery after stroke. What's unique is that many of these biomarkers were defined in terms of cancer, because in cancer, what often happens is you're trying to follow a patient as the cancer progresses. And so you measure something either serially or at the start and you're looking for something that tracks with disease progression. What's different in thinking about spontaneous recovery after stroke? Many recovery in the first three months, which we'll get to in a second, is that instead of monitoring decline, meaning that the that the disease is progressing, we're actually monitoring recovery and that's that's sort of unique. And if you look at the literature in biomarkers after stroke, many of the terms that we use are actually quite conflated. So I just want to point out that there is a strict that that the NIH has defined what a biomarker is applying it to the field of stroke recovery has actually not been done that rigorously yet. Um And so uh in terms of thinking about biomarkers of our motor recovery after stroke, a lot of what has been done so far has focused on the cortical spinal tract. Both in terms of thinking about predictive and prognostic biomarkers and response biomarkers. And I define a predictive prognostic biomarker really thinking about spontaneous biological recovery which we're gonna talk about effect a second. And a response biomarker is thinking about what happens to a patient after you give a treatment. But regardless of either of those two themes, a lot of what's been done so far has really focused on the cortical spinal track and and really structure and function of the cortical spinal tract. And so people have done so far is they've used is um they've used um transcranial transcranial magnetic stimulation to to map whether a patient is M. E. P. Positive or M. E. P. Negative. And basically what you do is you do a single pulse over the affected hemisphere to to map out whether someone has a motor evoked potential which you measure in the affected arm in the FBI muscle. And basically there are studies that show that if your MPP positive you're much more likely to recover than if you're mm P. Negative. Um Similarly people have mapped out the cortical spinal track using high resolution imaging by looking at the fibers uh and comparing the fibers of the cortical spinal tract from the tips from the affected arm um um affected side to the non affected side. And they generally look at a ratio of the number of of cortical spinal tract fibers that exist in the fiber bundle. I will say that both of these overall techniques well very exciting. Uh Take a significant time and effort. Take dedicated machines like a TMS machine which is not necessarily in the U. S. Available clinically or for M. R. I. Take extra scanning time and dedicated sequences which as as you as everybody knows for stroke patients is actually very difficult to do at scale. So um but so so our first study, I'm not going to talk about the first study that came out of our database that we have created. Which which which I talked about sort of how we created that in in the first few slides but the first study that came out of our database really was looking at structure of the cortical spinal track and whether we could relate structure of the cortical spinal track to recovery after stroke. Um So this was six strokes that I saw as a as a resident all causing arm weakness and as as you can see strokes affect different parts of the brain from the more cortical parts, the deeper parts. Um So the primary hypothesis for our first study was basically that we wanted to kind of take these themes that had been percolating in the research fields and basically asked the question, could cst injury from acute stroke images, which which most all patients get when they come to MGH um could that predict upper extremity recovery in the first three months after stroke, meaning could we use it as a biomarker for upper extremity motor recovery in the first, in in the first three months. And then I'll talk in subsequent slides about sort of how we did that. So, um from our database, in a prospective way, we took 65 patients with upwards so many weekends after ischemic stroke. Um we, we uh we all of them again had acute M. R. Images, um and we did an initial feudal meyer within the first week of stroke. And then we did a repeat Hugo Meyer at 90 days. And um there were a few patients that dropped out or who passed away in in between the time and we included 48 participants in our final analysis. And so these are upper extremity feudal Myer scores. So the feudal Meyer score is a 0 to 66 point scale, 66 being perfect, zero being plastic that measures motor synergies of the upper extremity. We just use the motor component. Um and so this basically just shows what the feudal meyer was on hospital admission and what the feudal meyer was an any day follow up. And um uh you can see here that there are sort of two patterns of patients that emerged. There are patients that start out sort of mild to moderate and recover very well. There are patients that start out severe and actually recover very well and there are patients that are severe that hardly recover at all. And so another way to plot this, um the same data is to plot it the following way in which you're looking at the, what the X axis here is the potential for recovery, that's just um 66 minus the initial score. So it's basically initial stroke, it initial few Ghimire and the Y axis is the actual change in Fuel Meyer, which is how much they actually recovered between zero and 90 days. So this is actually the same data uh that's shown in a different way from this from this graph here And um what crack with uh so what Jon Krakauer and others had shown previously, starting with the paper, I think in 2006 um was basically that if you do a cluster analysis, on on your data, looking at um patient at, at this group of patients, there are essentially two clusters that come out really and and and the way that they found this was basically, if you do a linear regression, looking at many different factors that predict change in feudal meyer. One of the most robust things that comes out is initial feudal Meyer score. And so this gave rise to this kind of rule uh which is called the proportional recovery rule. Which which which I'm not seeing everybody's faces on this call, but but which has come under a lot of heat recently, which I'll talk about a second. But um basically what the proportional recovery rule says that it is on average, um people tend to recover proportionately to their initial deficit and they tend to recover. And if you do that linear regression, the the The regression factor is .7, so meaning that people tend to recover 70 of their available recovery. And so this gave rise to many different papers in the field that basically started talking about sensory recovery and neglect recovery and gate recovery. And people started to extrapolate from this, that potentially there are fixed mechanisms for which people recover after stroke. And basically it doesn't really matter what you do in the first three months, people either recover proportionally or they don't. Now. This has been the topic of many debates in in at stroke recovery conferences and there are all sorts of statistical sort of problems with um modeling a change score by an initial score. Uh that that that I'm not gonna get into right now. Um and but but just to say that the story is much, is likely much more complicated than proportional recovery. It's not just that there's this binary rule and that patients recover on inpatient don't and I don't think from talking to john cracker that he actually he actually intended it for for it to be a prediction role. But in relatively small data sets like n equals 48. It provides a good way to start to to model the change because with with if you really wanted to to do this in a big way you need a lot of fuego Myer scores on the span of hundreds and so in relatively small datasets. Um I still do think that that that there are that there is something to thinking about people that achieve proportional recovery versus not. So that's what we did in this study to start. So um basically we modeled people that proportionally recovered versus those that didn't And there were 31 people that proportion recovered, 17 that didn't. And as expected in this initial table the patients that proportional recovered um Had lower NIH stroke scales uh than the limited recovers. They had higher fuel Myers than the limited recovers. Uh and um and their outcomes at 90 days were better than people that didn't that that that that did not achieve proportional recovery. And so basically what we did to start was we asked the question, well can we take lesion masks that you know since we get MRI's on everybody who um who has acute stroke, can we take lesion masks uh extrapolate the cortical spinal tract injury and then use it to predict who's going to recover and who's not. So we took our MRI's we transform them into standard M. And I. Space. And then when you do that you can um put them into that lab and ask and ask all sorts of questions about different parts of the brain that the lesions overlap. Um And so so this is showing, so the first thing this is showing two different cortical spinal tracts, the blue and the and the green. The green track is actually the johns Hopkins cst tracks. I don't actually know the origin of that but it's available in FSL which is an imaging program. And the blue track was provided by Steve Kramer, one of my collaborators and mentors um uh in which they scan 17 healthy participants in A. D. T. I. Scanner and created their own version of the cortical spinal tract. And as you can see here people can't even agree on where the cortical spinal tract in the brain is on a template. So that's that's one problem. Um But regardless if you if you look at these two people, one person had a big stroke. One person had a little stroke. This person had a big stroke that spared the cortical spinal tract largely. This person had a relatively little stroke but hit the cortical spinal tract pretty much straight on and the person that hit the cortical spinal track straight on had had had almost no recovery. The person that hardly hit the cortical contract at all had very good recovery. And so there turns out also, in addition to not agreeing on methods for calculating for for where the cortical spinal tract is, it turns out that people also can't agree on how to calculate injury to the cortical spinal tract without getting too far into it. Um When you have a stroke, a stroke lesion in three dimensions and you have a cortical spinal tract, also in three dimensions, simply overlap the boxes on top of each other. You're going to double count boxes because if you if you believe that the cortical spinal tract is a bundle of axons that goes from primary motor cortex with one cell body down to the spinal cord. Um then then if you overlap it multiple times, that's not gonna work. So there are different techniques for actually when you have your stroke lesion mask and you have your cortical spinal track template. Actually calculating how much injury there has been done to the cortical spinal track. One method is just taking the Z slice and overlapping it. Another method was developed by Gottfried and others in boston, in which they came up with something called the Ron waited cortical spinal tract lesion load. And another method was developed by steve in his papers, in which basically you split the cortical spinal tract into 16 which is an arbitrarily picked number Of of of axons that are descending. And you ask the question how many of these 16 axons were cut by a specific amount. And if you and if you've injured that that spaghetti strand by a specific amount then you just take out that spaghetti strand completely. And so when when this started we just said well let's just implement all of them and see which one wins and see if there's a better one. So it really started as a methods paper. And so that's what we did. It turns out that the answer is that it doesn't really matter what you do. So it turns out that for proportional recovery versus limited recovery those people that recover proportionally versus those that don't. Yeah. Um They if you do a logistic regression controlling for lesion volume uh that the that it doesn't matter what. So so these are the rows of the different methods and uh this was the amount of injury and then this was the p value in between proportional and limited recoveries. And so it turns out that all of the methods could distinguish proportional from limited recovery. Um When you ask the question, well how good of a classifier is it using R. O. C. Analysis. You get the answer that it's pretty good but not great. So you get a you see values ranging from .7 to .8 which which is good. Probably not good enough for clinical primetime yet but okay. Um And then if you divide people based on quartile of cst injury going up. You see that you have less proportional recovers and more limited recovers. Again, it's not perfect. So it's very good at distinguishing people that it's it's very good at distinguishing people who have little cst injury. They're gonna probably proportional recover except for one exception versus but for those that have severe cst injury, it's actually not that great. There there are three outliers out of nine. So then the real question is at an individual level, how good is this right for for an individual, if you do a regression controlling for um initial injury, um then how good is the cortical spinal tract of telling you what the delta is going to be? Um It turns out that it's not great. It the r squared is about 20% of variants. So basically to take home from this initial study was that regardless of tract or method used, we could explain using acute stroke MRI's and template overlap with cst about 20 to 25% of variants in the in the change in feudal meyer score between zero and 90 days using these methods. Um And so there again, you know, there there there are all sorts of caveats about using acute stroke neuroimaging. Does it actually represent the final infarct burden etcetera etcetera. But the great thing about this is that it can be done at scale, like it can be done in retrospective data sets. It can be done on clinical trials that have already been done um etcetera and it provides a way to start to stratify people into different buckets based on their recovery potential. Okay, so that's my biomarker spiel, I'm gonna talk about outcomes now. So um As as I mentioned before, we collect multidimensional outcomes for recovery after stroke. Um so this is one way that we started to visualize this. This is a polar plot. And um all of these scales are normalized from 0 to 1 on the outside level, essentially based on the ceiling of what these scales are. And I've color coded here, we've color coded here the different axes of the I. C. F. This is impairment activity and participation and this is just a show and and the different sort of shapes show the different changes over time what in the first week, six weeks, three months and six months. And this is just to show that recovery is multidimensional and just because one person is recovering in the impairment domain does not mean that they're necessarily recovering in the participation domain and vice versa. And so this is just for patients and and and and this very clearly shows that if you just think about this shape, it's not the same for everybody. And and so it shows that, you know, if we're targeting rehabilitation strategies we need, we we potentially need to think about them more holistically. Um So we've thought a lot and haven't gotten to quite closure on how to quantify this, but 11 or two methods of starting to do this or the following. So one kind of story that we're putting together now is that we're starting to look at the concordance of different outcome measures and so this is a story about outcome versus recovery and sort of, what are the different outcome measures that we're using uh in stroke recovery trials. And so again for acute stroke people, the modified Rankin scale is sort of gold And the modified ranking skills are 0-6 scale with six being death and zero being no um being nothing at all and 1-5 being everything in between. And you'll see here that you map on the outcomes of Hugo meyer and boxing blocks and Barthel index at 90 days over the, over the outcomes of, of ranking in 90 days. You see that there's a significant distribution of people who have a lot of, a lot of change, a lot of variants in their feudal meyer outcomes, especially the two and three range. And this is just saying that the ranking scale is not capturing uh in a sensitive way, upper extremity motor recovery, meaning that people can recover significantly on their feudal meyer and completely be missed if you're just classifying them as a ranking of three. Um and another way of thinking about this is, is thinking about the change in between zero and 90 days. So in, in the previously I'm just talking about outcomes here, I'm talking about actually the change between zero and 90 days. Did people achieve concordant change. And so the way we're thinking about that is there's a concept known as the minimally clinically important difference which is the smallest amount of change in outcome scale that a patient perceives as clinically meaningful. And this is established in different ways and all sorts of caveats etcetera. But it's the the M. C. I. D. Is typically used to assess progress through the continuum of care, established benchmark and and to assess the effectiveness um of of of different rehabilitation therapies. And so what we did here was for the feudal meyer, the blocks and blocks text the Barthel index and the ranking, we've listed out the proportion of people that achieve M. C. I. D. And then we looked at those that that achieved M. C. I. D. On one scale versus those that did not achieve em CCD on another scale. So if this if all the skills were perfectly concordant then all of these would be zero. But as you can see there's there's a good there's quite a number of people who are achieving M. C. I. D. Chain on the rank and scale but not achieving M. C. I. D. Change on the feudal meyer as well as vice versa. Um Finally just to talk about outcome measures. Another thing that we started to explore our different trajectories of recovery. And so one way we started to do this is looking at the M. C. I. D. And so a graduate student who works closely with me uh basically took about 40 of our patients and split them up into patients that recovered M. C. I. D. In the first six weeks versus those that did not. Um I'm sorry the red are patients who who who recover the M. C. I. D. In the first six weeks. The green are patients who recover M. C. I. D. At a later time point after six weeks. And the blue are people who never achieve M. C. I. D. And basically she looked at the number of therapy hours between these three groups. Fast improvers, extended approvers. The green and limited recovers the blue and she asked the question what does the number of therapy hours differ between these groups? And it turns out that there's a difference. Uh there's a difference between the the fast and limited group but there there really isn't a difference between the extended uh and the limited group meaning that the patients who have an opportunity for recovery are potentially not getting therapy in a way that they need. Okay so I'm last in our outcome measures um part going to be talking about a story that we're developing that's focused on choosing the right outcome measures after acute stroke. And so here we're looking at so here we're looking at performance on the box and blocks tests versus the grip strength test. And uh this shows the affected or contra seasonal upper extremity. And this shows the absolute national upper extremity, the traditionally non affected upper extremity. And so if you look at the contra legionella proximity, you'll see that as it's the Y axis here is normalized for age matched gender matched controls. And so therefore everything has been converted to a Z score on the y axis. And what you'll see here is that uh is that as expected for the khan traditional upper extremity, there's recovery over time. You start out at very severe Z scores minus 70 scores and there's recovery between one week and six weeks and six weeks and three months. And if you do a repeated measures in nova between these these time points are highly significant, meaning that the affected arm recovers over time for both blocks and blocks and grip strength for the solution alarm. What's notable is that, is that it's actually, and and there is literature a significant amount of literature to to suggest this, the, the absolution alarm is actually not normal after stroke. Um and they're actually also is recovery of the absolution alarm in the first three months. But if you if you compare grip strength two blocks and blocks. Now, if you look at the rose and the differences between the second row and the first row. You'll see that the grip strength scores even though normalized for age and gender matched norms are significant are not as impaired as the blocks and blocks scores at all time points. And so our question really was why is this, why is it that a task in which patients squeeze that involves muscles of the arm and hand? Why is it that that is not as impaired as a, as a task in which patients move a set of blocks? And our our initial hypothesis is that this has to do with attention and cognition because the boxing blocks test takes one minute of sustained attention to transfer a set of blocks from one to another. While the dynamometer is a is a is a, is a task in which you just ask the participant to squeeze and it's a task that happens abruptly onset and offset. And so, uh so so so so our first hypothesis was by by by by varying the nature of the attentional demands of these two tasks. We could pull out that patients actually after acute stroke are not as impaired on the grip strength task of the absolute evil and contribution alarm as they are on the boxing blocks task. Look at that a little bit further. We did a linear regression where we're asking the question, well, does cognitive impairment explain any of this? We extrapolated cognitive impairment from the NIH stroke scale, which there's, there's a sub scale of the NIH stroke scale called the NHS is called four. It's not very good, but it's what we had. And so basically we ask the question while controlling for lesion volume, uh, does cognitive impairment explain any of the variants in motor performance of the contra seasonal and regional arms of boxing blocks and grip strength. And it turns out that for for boxing blocks at T. one, Especially for its original extremity, it does. So the beta is significant here and the model explained 33% of variants for the IP solution, alarm at T1 and 20 to prevent Percent of variants for the contradiction, alarm at T1, meaning that it did not do so for grip strength, meaning that cognitive impairment is explaining variance in motor performance early after stroke for the test that's requiring high cognitive demand. We also, we finally uh, mapped are lesions again, to ask the question, well, how does the anatomy of injury relate to performance of these tests? And so what we did was a Vauxhall lesion symptom mapping technique, where we basically asked the and that's basically a technique where you can on a Vauxhall by Vauxhall basis? Ask the question Well, for, for patients that have lesions at that Vauxhall versus patients that don't. What is the difference in performance of the group that has a lesion at that Vauxhall versus a group that doesn't at the end of the day. If you do that on every Vauxhall in a standardized space that basically gives you a map that associates anatomical injury to behavior. And so if you do that, we did that for contra regional and absolute regional blocks and blocks and grip strength. And so if you do that, basically what you see is that the box and blocks injury that's associated with performance on boxing blocks, it's associated with brain structures that are much farther away from the motor cortex and the cortical spinal tracts than we would expect. So it's really involving significant injury to the insula is particularly the anterior dorsal aspect of it while for grip strength injuries, you know, although it's it's it's still somewhat non specific, it's much more restricted to the primary motor cortex and cortical smile track. We can actually quantify the number of boxes that are output by RBl sm that are non sensory motor versus sensory motor and so for boxing blocks, there is a much higher percentage of box ALS, both for contra legion and absolution upper extremities that that are associated with performance. And so so basically, the, what we're just preparing this for publication. But basically what we're seeing here is that cognition specifically impaired attention is our, what was our initial hypothesis can actually confound motor performance starting early after stroke. And the relevance of that is twofold. First, we know that cognition and motor performance are uh important for functional outcome. But one way that that cognition might mediate uh the impairment and functional outcome is that actually might be directly mediating impaired performance of motor of arm. That's the first thing. And the second thing is that in thinking about clinical trials and the tests that we want to pick for for being able to quantify recovery after stroke starting during early in the acute phase, especially when when cognitive impairment is very prevalent. We want to pick if you want to really quantify the motor system, you wanna pick tasks that are more selective to the motor system. Okay finally in the last five minutes I'm going to talk about sort of how we're thinking about new treatments for recovery after a stroke. So uh as this crowd knows there's a lot of different treatments that are in the clinical translational pipeline uh spanning pre clinical development ready for broad clinical application. These involved biologics to you know robotics which there's a robust program at Hopkins for to constraint induced movement therapy et cetera. And the question is why haven't many of these kind of reached large scale at adaptation. Um And so I was a postdoc with lee Hochberg in the brain gate lab which and for for those that don't know the brain get lab is um it is a brain computer interface trial in which uh intracranial electrodes are implanted into primary motor cortex to allow people to control robotic arms to type on screen. Um And to control their own arms and limbs versus electricals stimulation. And um this is done by decoding activity in the primary arm hand area of motor cortex. But more generally what a brain computer interfaces is basically associating a brain signal with whether that be E. G or F. M. R. I. Or single unit recordings with an output what oh with an output device and that output device can be anything, it could be a robotic arm, it could be an armorer Thanasis etcetera. And so there have been small studies in which in uh in which people have looked at using A. B. C. I. Arm or Thanasis for upper extremity rehabilitation. So in a different way from uh from actually um a patient using the device constantly basically what what what the premise of this is is that if a patient can't move their affected arm the uh if they have an eEG attached to them and they think to move you can actually drive movement of anorthosis so similar to a rehabilitation session. What what what you can do is have the participant drive his own arm and hand movements. And people have shown in some relatively small clinical trials that if you do that you can actually improve the feudal meyer score above controls. And that the FmRI activity has shown here shifts towards the more adaptive side that the side that's opposite the lesion. Again these are early stage trials that have been done mainly in Germany and other places and so we started at MGH to build the components of an upper extremity rehabilitation device. We received a grant for the in motion arm um which was previously the M. I. T. Manus. We've started to collect um upgrade from the cinematics after acute stroke on our award and we're also doing high density EEG recordings for these same patients at the same time. Uh This is just a picture of me and in our prototype. Um and we started in addition to to to building this, we started to look at using high density E. G. We started to look at how different brain regions are connected to each other because with the high density E. G you can look at how the signals oscillate in relation to each other. And this is very early work but basically we're finding that there are there are robust associations between changes in coherence promoter and motor cortex with changes in Fuel Meyer. So um but I think you know, just going back to the initial slide that launched this section, I think there's a long way to go before we can actually translate both neurotechnology uh for upper extremity motor recovery because we don't fully understand yet how to associate neural signals with recovery in the best way possible. We haven't yet solidified the best biomarkers to stratify patients in the clinical trials and and we we we we need to come up with a better approach to capturing outcomes after after recovery after stroke. And so our vision really uh you know with our Natural history study that we've launched in conjunction with the neuro technology that we're developing at MGH is really to think critically about outcomes that improve quality of life to to develop biomarkers that can that can inform where patients go by being able to predict how they do in a clinically relevant way. And then finally to develop new neural technologies that are informed by neural structure and function. Everything. We're learning about spontaneous um everything we're learning about spontaneous recovery after stroke. Um so I'll just stop there. I've been very fortunate in my early career to receive uh um a significant, I've been very fortunate in my early career to receive um support from both the MGH neurology department as well as external sources including the A. N. Um and the V. A. Um and this research as I mentioned, it takes a village of people that really believe in this mission. Uh And so this is a picture of our smart team at one of our retreats. Um and really this is a collaboration between occupational physical therapy neurology slp as well as the MGH Institute of Health professions. Um Again, these are these are some of the people that have helped me with this work. Um but there are many more to thank. Um and so I'll take any questions. Okay, mm hmm. Hi. Um this is pretty Raghavan um David, thank you very much. I think this was really excellent. You know, as you know, we are trying to get our stroke institute started and I think this is a inspiration for the kinds of information that can come out of systematically looking at recovery. So, kudos to you for putting this all together and really dissecting what it takes to recover from a stroke. It's it's definitely not straightforward, it's complicated. And I think what you presented really brings that out. So thank you. Thanks pretty looking forward to working together in the future as well as as as we've been doing over the past few months, so okay, uh dr lin, I I just had a question. I I think I was uh I'm a physic um here here in the department. Um I think and and you know, actually currently I'm attending on the inpatient rehab rehab unit. I think one of the themes that you really, I think was really highlighted in your talk was again, this relationship between motor recovery and cognition. Um I think what I struggle with those cognition is such a big, broad topic. And so when you're talking about somebody being cognitively impaired, that means a lot of different things for a lot of different patients. And so I I just I just wanted to know, I'm sure this is something that you guys have talked about, but if maybe you could provide some more thoughts on that because I think it's you know, I think it's something that as a clinician I definitely see, but but I but maybe maybe to have some more details about how you guys are thinking about that, I think would be helpful. I mean that's a great question totally and and you know, I'll be the first to admit that in my presentation, I kind of hand waved a lot, just said the word cognition a lot. Uh but I totally agree that the cognition is abroad it's not, you know, it's not a 11 size fits all turn, there's visual spatial abilities, there's attention, there's executive function, there's language etcetera. And they probably all have some sort of synergistic interaction with motor recovery. Um And I think we're in, you know, to be honest, I've looked into this a lot as we've been preparing this paper, we're in very early stages of parsing that out. I mean I think broadly as a field we know that cognition impacts functional outcomes. People have shown that if you collect a mocha early uh and then you measure functional outcomes later that the mocha correlates with functional outcomes. We also know from neuropsychological literature that um cognition impacts motor performance and healthy people. For example, in that literature, mainly from geriatric people in which if you're looking at motor performance upgrades should be armed dexterity as as uh as attention decreases with age. Uh there are there are decreases in motor performance and sorry, but I think parsing that out in stroke. I think stroke introduces a neuroscience opportunity to look at that which has actually been looked at very little so far and I think it's a great opportunity for the field. Um but but we're in very early days of parsing now, the data I presented was really highlighting the role of attention because the difference is we saw it in the boxing block versus grip strength task was really an attentional one. So someone has to sustain a minute of of of a minute of attention to be able to complete boxing block while they don't for um for grip strength and the area of the brain that we found in our study, the anterior insula has been implicated in inattention. Um so that's the story that we, you know, we didn't have a very detailed measure of of attention or cognition in our study applied at the time in the 15 patients in our dataset, we've now implemented the click, it the cognitive linguistic quick battery uh as well as the mOCA as well as a few other cognitive tests. Um I think there's a tradeoff there between the political reality of doing detailed cognitive testing on patients and being able to physically do it. So, um I totally agree with you that parsing the different features of attention is very important and and I think it has very direct relevance for rehabilitation. Um if we think about our motor rehabilitation, you know, we can make big progress in the field by being much more specific about the tests were having patients do based on their attentional status. Thank you. And I guess maybe just to follow that up, maybe just one more question. Just again, you know, I'm working in the inpatient rehab setting and and just um a lot of these studies that you're presenting our our patients, you know, sort of 90 days post stroke. And and I'm seeing a lot of these times these patients that are, you know, maybe one week or two weeks after stroke. And I just wanted to know, it's it sounds like, you know, we we perhaps there's sort of these two groups of patients one that are following this sort of trajectory. You know, you you kind of gave that equation about the 70%, you know, that's being debated. But then there's this other group that's sort of falling off. And is there is there any clues you're on in that first couple of weeks? Um that that could maybe indicate uh you know, somebody kind of maybe going into that that they're not falling sort of in that expected trajectory. Just just wanted to know if you had any thoughts or or you know, can point us in that direction in terms of that earlier recovery. Yeah. That's also, you know, I think that that, you know separately from cognition for a second. I think that, you know, one of the themes in neuro rehabilitation which which which hasn't actually been pointed out that much in the literature, but we but, you know, we we do a lot of fancy techniques for example, the best thing right now probably for predicting record upper arm motor recovery is pretty and others on this call. No is the prep algorithm in which they do transcranial magnetic stimulation to do tms positive or negative plus a combination of m ri plus a clinical score. But I think you know more robust than that is probably just serial assessments. Like if you did a serial assessment in some way whether that's the future meyer or or or or a test that doesn't take that long uh you know in the first week or two after stroke and you saw that there are patients that are starting to just show some risk corsa flexion or just starting to go from please jah to you know moving in upper extremity synergy I think that that slope of recovery in the first two weeks is probably going to tell you more about the trajectory of our motor recovery in the first three months than probably any neuro technological test. That's my view or even better would be if you did that in parallel parallel with serial neuro technological tests, how cognition factors into that. So the data that I presented was actually the boxing box score and grip strength data that I presented from our study was actually collected in the first week after stroke and then at six weeks and three months and what and what we showed there is what we're showing there is that cognition impacts motor performance most in the first week after stroke. And really the point is there is that that if you're asking people to move blocks for a minute, their actual score might not just be a motor system score, but it's actually other attentional networks that are brought in and reflected on that score while if you're measuring grip strength that that that score maybe a more pure motor score. Um so so I think those are a little bit separate things but but I think for for predicting recovery, measuring trajectories of recovery, doing interval assessments is probably the highest yield and then how cognition inter plays into that. I think we don't know yet like how cognition affects whether somebody develops a change in there first few weeks of recovery. I think it's a really interesting question that I haven't seen any answer to in the literature. Thank you. Maybe I can. David thinks that's thank you for the dog. And it's very, very nice. And just to amplify what David just mentioned is that, so there are a couple of, I think very important studies that people have been looking into in the north Korea field. One is from Winston B blow group that they do these multiple potentials and this idea of the proportional recovers and so on. Although the proportion of recovery is very flow unless you look at the people who do not recover regardless of that. What they found very clearly is a nice relationship between those who are changing, admire those who have the presence of enemy P. In the in the hand early on. So that's kind of a big group that proposed this this this idea of the MVPs and the prep algorithm that David was mentioning. And then on the other hand, you have this very nice studies from gynecological that follow, you know, a couple of 100 people with stroke and and showing that the extension of the fingers of the wrist in the first week it has high level of predictability of people who are going to be recovering the M. E. P. In his studies show that the M. E. P. S on the A. D. M. Maybe adding a little bit more of the variance of those who are only moving the hand up. So with just those two things you have a huge explanation of who are actually going to recover and some ah So like David pointed out maybe we are going around all the heavy intensity heavy high tech stuff and then we're going back to a kind of earlier of the basics of that, that was describing a lot of the variance of who recovers who doesn't. But what is it as as a government? And again they would think. But sometimes I just want to ask you if you don't mind one of you show a lot of the data and results. But I was curious about that that you showed that the number of hours of therapy, whether you're recovering the first week in the first three months or ordinary recover and showing that those who don't recover actually have more therapy hours versus those who record fast. So because I didn't have a lot of time to present the questions are are you just mentioned the total number of hours of therapy or because it makes sense if you're not going to recover, you're gonna be in rehab for a longer period of time. So you're gonna have more hours of therapy in the patient. Ah If you're going to require fast like you were discharged from rehab you're gonna have less hours of therapy. So if you can explain to me more about that that that those results I'm curious to know. And and the other question with the hours actual time to task, not just the therapy has been a number of sessions. So yeah, I brushed over that. So we so when patients come back to our research that we have this form that we fill out with them that asked them how much therapy they've gotten that's done by a combination of chart review, like we look at Epic. If they get their rehab at partners MGH Brigham or or if they don't then then the best thing we do is we basically ask them and ask their caregiver, how many hours have they received and it's you know, I'll be the first to admit it's a it's a gross estimate and you're probably right. The people that don't recover are probably just those ones that are in rehab and so they're in therapy for longer and so they're telling us that they're getting three hours of therapy a day now. I mean I defer to you dr selma because you you know, you have much more experience in the field that I do. But you know, there are a lot of studies that are trying to estimate the quantity and the quality of rehabilitation including you know stuff from wash u with Catherine langeais group using accelerometers there and and stuff with Heidi chandra who I think was at Hopkins at some point. But at N. Y. U. Who's doing all sorts of body sensors. You know, I haven't seen a good solution to that. I mean when we write about this, we're like, well we just asked them and we kind of say, well that's the best that we're doing right now. I mean I think better or next version would like to have them journal like actually journal an app or on paper like what they're doing. But the the engineering like of estimating upper extremity movement from sensors, that whole field. I haven't seen that yet. Be a solution. I don't know what your thoughts are on that. Yeah, I'm I'm I agree. So there is not a consistent, I think some different groups are working on this and I think that there is a need to have a better assessment project, the assessment of what is actually happening in the therapy and the training exercises. Um so either E. M. U. S. Or some video form of capturing activity or so people have, like you said, people have been playing with different approaches to this and I think that there's no consistency and excluding those things have not rolled out in my opinion. Like you, I believe to to the clinical practice. So Something therapies are yes, writing you have a session, you can have some therapist to write down every 15 minutes what they were doing. So some people have done that, but it's very onerous and the person, so it's not, I don't think today we have a systematic way of objective eyes and what activities are happening within a therapy session. Yeah, I think it's a it's a bright area. Yeah, I agree. I'd be, I'd be curious about you if there, if there are just some units of of trying to utilize some sort of fidelity check or or or is there a way to check to see if there is a measurement based physical therapy care that's being utilized um and um something like that as opposed to hours that I agree with that. Um 11 question I had for you dr Lindsay by the way. Great presentation. That was really enjoyable. Um So um when you were going over the objectives, the first objective was just sort of characterizing some of the challenges that uh some of these systems have these these rehabilitation recovery systems have um uh in hospital settings. And then um and then you follow up with biomarkers and and talking about different sort of systems level neuroscience ways of addressing the needs of folks who are recovering from stroke. And I wonder if you could speak a little bit more to um how some of these systems neuroscience level techniques could address some of those um more systems level health services issues that you describe like taxes or insurance or or even quality of care if you could speak a little bit more to that. Yeah, that's a great question. I mean there's a that's a great question and there's a fundamental disconnect I think, which we found in trying to launch this study. This is very hard to do in the US as as as as pretty, you know, knows and uh you know, the studies in in Netherlands Fromberg crackle, you know, in which he's followed 500 people and done serial feudal Myers by driving a truck to their house. Uh you know, it's very hard to replicate that, that the United States where patients are transitioning across care settings and you know, to the cynical everybody's like, well what is it? You know, it seems almost insurmountable in some ways the health system's problems um You know I think the first thing you know I am excited about uh you know in in new Zealand with the prep algorithm. Cathy Cathy steiner Winston did blow have been using this algorithm called prep which uh which uses T. M. S. And M. R. I. And the shoulder abduction finger extension score. To be able to predict people that recover on the air at the Action Armed Research test versus those that don't and puts them into bins. And they've used that they actually used that. And given that that prediction to therapists in their rehab settings. And basically they showed that if you give a therapist in an interview rehab a prediction of how a person is going to do that. Um The length of stay of rehab actually shortens a bit. So that's that's one way. You know I think in the U. S. Could could we see a future where we do a set of of behavioral and neural technological assessments within a few days after a stroke. And then we inform where people go in the health system or what they need. I would love to think so I mean I think I think that's the natural extension of this work of all the work that a lot of people in the country are trying to do. So I think that's the first step is informing discharge destination. And the second thing which you know you guys know well in this department is I do think that being able to somewhat predict stroke and come up with relevant biomarkers that we can scale is a prerequisite to being successful in clinical trials because the way that we're running clinical trials right now at the first past were not even um stratify patients based on behavioral level. Like we're just including all patients with stroke. We're not even saying, well let's stratify the feudal Mayer in an intelligent way. I think if we add in, you know, cortical spinal tract injury or TMS into these clinical trials, then we can start to say, well these people will respond to these people won't or these people respond this much and these people won't. I think that clinical trials in stroke recovery need to head that direction. So I think that those are the two ways that I see. I see kind of the systems, neuroscience informing both clinical trials and and the care system. Because the other the other thing is that you know the fare like uh you know, I'm speaking to a group of visitors who knows this well. But you know if you talk to OT and PT and S. LPS their clinical practices built much like our neurology exam is right, it's a it's a clinical thing and there's not there's not an objective quantifiable approach that standardized and so saying that we're gonna we're gonna standardize therapy in some way we're we're really far away from that. Um great talk great talk dr lim um Just wanted to highlight the importance of something you mentioned about the relevance of choosing the right outcome measure after stroke And from where I sit. Um you know, I've been doing vocational rehabilitation for the last 22 years and one of the great macro outcome measures. The measure that emerges long after long after the continuum of care is near completion is is whether people return to work and I think what's brilliant about your research and your work and and our group here at Hopkins is that um it's marvelous to know that these outcome measures um these objective measures all add up to some wonderful large outcome later on whether it's person returning to work or person returning to their role as a caretaker or a homemaker. So, thank you very much. I was wondering if you had any specific thoughts about macro macro outcome measures? Yeah, I think you hit on it you know, precisely. I think the measuring outcome measures is key. And picking the right ones. I think, you know, thinking about what are the outcome measures that are clinically relevant is key? You know, are we gonna come up with a new outcome measure for recovery? Like the, you know, like the wolf motor function test? I'm not sure if the answer is that maybe the answer is doing a panel of outcome measures that's that that people agree on that span the I. C. F and that collects outcomes and you know across different domains and then integrate them in a in a in a mathematically and functionally relevant way is the way to go. Uh You know I was inspired by the recent Queen's Square nick ward paper in which they you know they did this intensive therapy session and they gave people many hours of therapy in the chronic stroke setting and and they measured outcomes across many different domains. And I think that's what made that that study so powerful. But I agree that understanding how outcome measures map onto each other and understanding which one and be very clear about the ones that we're measuring for you know for the purpose of systems neuroscience as well as the purpose of understanding function is really important. And I think that uh in neurology we tend to lean more towards just you know measuring the feudal meyer because that seems to map onto systems neuroscience. But I think to do that alone is a little bit shortsighted. Thank you. Okay well thank you everyone. Um This has been very interesting thank you David for uh very exciting talk and an exciting discussion that followed um you know just for everybody's information. We are now collecting standardized outcomes across the I. C. F. Domain um at regular intervals just clinically. Um as part of the stroke institute and um I think we would be able to have a lot of synergies with some of the work that David has already done. Um So again, thank you very much everybody and um enjoy the rest of your day. Thanks. Pretty thanks everyone. Created by
Facebook
Twitter
YouTube
LinkedIn
Instagram
E-News
BROADCASTMED