 Thank you very much, Beverly. And thank you, Julia, for the invitation. I was actually wondering about what to present at this translational psychiatry meeting, and I thought perhaps trying to convert knowledge base into criteria for application in the field, both for research and for clinical practice may be one good example of that. And in that I probably am informed by my participation in the neurocognitive disorders workgroup of DSM-5 as well as VASCOG in trying to get a new set of criteria. And perhaps it also kind of exemplifies the kind of processes that go on when committees get together to try to come up with sets of criteria. Vascular cognitive disorder is probably a good example in that one would imagine that what you're trying to see is you have patients where you have demonstrable cognitive problems. And what you really want to establish is whether this person has subrovascular disease that can account for those cognitive problems. And because currently we do have the technology to look at subrovascular disease, one would think that this is probably an easy process. It's much easier process than trying to define some of the psychiatric disorders we've been hearing about earlier today. But from what we can see in the literature, it hasn't been an easy process. In fact, going back to DSM-1, we of course talked about chronic brain syndromes in those days. And we still had the concepts from the past, from the 19th century, in fact, that a lot of cognitive problems were because of arteriosclerosis. And in fact, arteriosclerotic dementia was supposed to be the predominant form of dementia in those days. But then early work, which is not listed here, of course in the 60s from Martin Roth's group, et cetera, showed that yes, certain amounts of infarcted brain could lead to a dementing process. And they actually said that you'd probably need about 50 mil of brain tissue loss to produce a picture like that. And the landmark study by Hachinsky in the 70s, which was actually with neuropathological verification, actually set the scene for perhaps two decades in terms of how vascular dementia was being looked at. And that was that you were talking about multi-infarct dementia. So you basically had people who had multiple strokes, and they'd added on cognitive problems to the point that you had a picture of dementia. So you had a situation where vascular dementia became synonymous with multi-infarct dementia. And that's what we were taught in medical school, when things have changed to some extent. And that change has again been brought about by technology to some extent. That with the introduction of MRI, we realized that vascular disease was much more prevalent and had more varied manifestations in the brain and not just large strokes. And people started looking at revision of these criteria. And there were some attempts in the early 90s to come up with a new set of criteria. The California group, which was led by Helena Chu, came up with their set of criteria, which are often referred to as the ADTTC or the California criteria. And then there was an influential workshop in the 91, which was published in 93, which was the NIN's Aaron workshop. And this proposal actually became the standard for many years in terms of how neurocognitive disorder, due to vascular pathology, was being defined. And then there have been a number of attempts following that. So let me try to actually define the issues here, and what are some of the core issues that one is trying to grapple with in coming up with a definitive set in the end. The first concept, of course, is of dementia and how is that defined? And here there has been a problem. And that problem has been that dementia, of course, most people accept that you're talking about cognitive disturbance in a number of cognitive domains, as you need two or more domains to be affected to make a diagnosis of dementia. And most definitions of dementia mandate that memory impairment is one of those domains. And that has been one of the difficulties in defining criteria for vascular cognitive problems. The other problem has been that just as any other neurocognitive disorder, that there are, of course, grades of disturbance of cognition and not everyone with silver vascular disease meets the criteria for dementia, per se, because there are sub-syndromo patients, so to speak. So they're milder end of the spectrum. And how does one deal with those has been the other issue. Now, this problem of memory being a mandated disturbance has been a kind of an Alzheimerization of the definition of dementia. In fact, when you talk to people from the Alzheimer's disease field that they think that they kind of own dementia and don't touch this and don't touch the definition. Whereas the vascular group have said that, look, we see a lot of people with neurocognitive problems who do not have significant memory impairment. And what do you do with those people? Do you really call them dementia or not? Now, so part of the process in DSM-5 has been to actually see, okay, what do you do with this problem of memory impairment as being a necessary feature? And the other problem has been that what kind of evidence do you need to establish that there is significant cyber vascular disease? And do you always need laboratory evidence or is clinical evidence sufficient or do you always need neuroimaging? And what do you do when you do not have neuroimaging available and as is common in many settings? And the third problem has been, which I'm not highlighted here, is most people do not have access to neuropsychological testing. So what kind of thresholds do you actually apply to make a diagnosis of dementia? Now, just put this up to say that ICD-10 similarly defines dementia as having memory and other domain decline, like DSM-4, so it has the same difficulty. And in fact, when DSM-4 put forward the criteria, they included neuroimaging as possible evidence, but ICD, because it deals with an international audience, said that, okay, we should not include neuroimaging in this. And they actually said that vocal drain damage due to clinical evidence was what was needed to make a diagnosis. Okay, so we can start with the next slide. Let's skip this. Now, how did NIN's Aaron deal with this issue? A few things came up. The first, of course, was the issue of dementia. Should they actually tackle that discrepancy between the presentation, vascular dementia, and the definition of dementia? They left it. They actually did not address that issue of dementia. They said dementia was diagnosed as is usual with DSM-4 and ICD-10, so memory remained a necessary criterion. Then they, in trying to establish evidence for Sibyavaskar disease, they said, okay, you could have, of course, clinical evidence, you could have neuroimaging evidence, and they said, okay, if you had one or the other, you would probably suffice to make a diagnosis, but to make a more definitive diagnosis, and that was the level of probable vascular dementia as opposed to possible vascular dementia, you did need neuroimaging. But there were a few more concessions given in terms of neuroimaging, and the concessions were that, okay, multiple infarcts were usually necessary, but sometimes you can get vascular dementia with a single strategic infarct. Say, if it's a large temporal parietal infarct, for example, or thalamic infarct, you may actually have single infarcts, and you get a picture resembling dementia. Now, this was actually not including patients who were aphasic because of a single stroke. They were not considered to be demented. The other concession given was that you could have multiple lacuna infarcts and not necessarily large strokes affecting the cortex for this diagnosis, and finally that you could have extensive white matter disease. And I think this went back to the old concept of Binzwanger's disease when dementia was diagnosed in these couple of patients that Binzwanger had published in the early 20th century who had cognitive impairment and only white matter disease. And with the introduction of MRI, it became apparent that there were many patients with extensive white matter disease who presented with cognitive impairment, and that concession was given by these criteria that you could have just that and yet make a diagnosis of vascular dementia. The problem, of course, was how much of white matter lesions were sufficient to make this diagnosis. And I'll come to that quandary as to how does one deal with that issue. Now, of course, the third point is that you have cognitive impairment on one hand, you have evidence of sebavascular disease on the other hand, and how are the two linked to each other. And the typical story in the past has been that, okay, you have stroke and you have disturbance of cognitive impairment soon after the stroke. So it's a kind of a stepwise progression or an abrupt onset of deterioration. And the fact that extensive white matter disease was considered to be sufficient in many cases suggested that often this may not happen in that stepwise fashion or in an abrupt onset that one would expect. And that actually presented a problem for the classification in that if you included this third point to establish that relationship between the two then you were left with this difficult issue of extensive white matter disease which presented with a slowly progressive course. Now, of course, when criteria are published people try to look at their reliability and, of course, correspondence in terms of validating concordant validity. And with these criteria it was shown that the concordance between different criteria was actually, in fact, quite low in the majority of cases. So, I mean, that's something that one has to live with to some extent. I think in trying to establish the truth, really, whether you can say, okay, the progression is towards the truth or in towards the reality as we really see it. And maybe that doesn't tell you what has been in the past. But one problem was highlighted by this paper by Ingmar Skug and colleagues is that depending upon what you included as your evidence for vascular dementia, the proportion of patients of vascular dementia in an epidemiological sample varied considerably. So if it was just stroke history, it was only 35 percent. It was focal symptoms and CT in factors only 13 percent. Whereas if you included white matter lesions on MRI scan or stroke history or in facts, the rates went up to as high as 85 percent of that of your cases of dementia. So, where is the consensus? The consensus is, okay, we have a neurocognitive disorder, however we carve it, what thresholds we use for defining it. Evidence, this and an association. So people accept that these are the basic principles on which you're going to define your criteria. But then things, of course, consensus breaks down after that. And in terms of looking at cognitive domains, after this various DSM-5 sessions and the VASCOG meeting and all, generally there is acceptance that if you're looking at some of the core features of cognitive disturbance in these cases with the cerebrovascular disease, the disturbance is in the speed of information processing and frontal executive function. And memory function is important, but the nature of this memory disturbance may not be the same as an Alzheimer's disease. And sometimes memory disturbance may not be present. And of course because of the very nature of cerebrovascular disease, there is great heterogeneity in the nature of the cognitive deficits that you see. And therefore, what DSM-5 has done is that it has retained all the cognitive domains. And it hasn't kind of mandated or isn't going to mandate any particular domain as being necessarily present, but it's not going to either exclude any cognitive domain either. And I think that's really what the VASCOG consensus seems to be. So essentially the diagnosis of dementia or mild cognitive disorders could occur with any one or two neurocognitive domains and memory will not necessarily be mandated. I think that's really where people have settled really. There's not much more in terms of defining that. Now the other issue is of course can neuropathology inform us much more about defining vascular neurocognitive disorders. The main points are that it's customary to distinguish hemorrhagic and ischemic lesions. And in fact, many studies of vascular dementia only take ischemic lesions into consideration or only follow up patients with ischemic strokes. And partly that's because it's much more difficult to actually follow up hemorrhagic patients and they've been excluded as a consequence. They're not pure in terms of the nature of the lesions that you see on the brain. And you actually look at the neuropathology that's very varied. And one of the interesting findings that is emerging from neuropathology is of course that the micro-infox, small infox which may be 0.5 millimeters or so in size may be very important and that they are often ignored even by neuropathologists and they're underreported. And when you actually look at of course the range of neuropathological lesions it's very wide in any series of patients. But when you try to relate these neuropathological lesions to vascular dementia you find that some trends seem to emerge. One is that cystic infarcts or large infarcts which have kind of hollowed out in the middle are not that common in these patients with vascular dementia. And that microinfarcts and lacuna infarcts are really very important. And when you combine that with non-infarct changes in the brain they seem to be very important in the clinical, as clinical determinants of these cases. And winter is concluded from the neuropathological data that ischemic vascular disease appears to correlate with widespread, ischemic vascular dementia appears to correlate with widespread small ischemic lesions distributed through the CNS. So I think that's the message that is emerging from neuropathological data and increasingly from MRI data as well. If you look at another study which I think is quite illustrative as well and these are neuropathological data from kind of a community cohort so to speak. So these are not select patients. But when they try to look at indicators of dementia in terms of neuropathology found that we look at plaque counts, they really were not significant indicators in this epidemiological sample. The BRAC stages which is based on tangles, you find that the stage five and six increases the odds of being demented by about six fold. And having microinfarcts, more than two microinfarcts were in fact the most significant determinant of having dementia prior to death. Similarly, Lewy bodies but they were not that common and cystic infarcts were of course having two or more were not significant but I mean they had odds ratio about two but it was a non-significant finding. So again, this is one area which is receiving increasing attention, the microinfarcts in the brain and something that neuropathologists have not paid a lot of attention to. And also neuroimaging may not be able to easily distinguish and many of these microinfarcts may actually lie within the white matter lesions that we see on MRI scan. Maybe I can skip this. As far as infarct volume and dementia is concerned, this is one of the other areas that I mean have stuck around since Roth's findings and then of course Hachinsky's paper in the 70s that you needed a certain amount of infarcted brain to produce dementia. So if you contrast individuals with less than 15 mil of infarcted volume with more than 15 mil infarcted volume and look at the ones with dementia, of course you expect that these people will have more Alzheimer's typology. But what really is important is that these people with less infarction they are more likely to have white matter disease and microinfarcts. So the picture again painted from here is that you have these patients with extensive white matter lesions, microinfarcts and not necessarily a large amount of overall infarcted brain tissue to produce the picture of dementia. And of course the non-study which most of you familiar with showed that it was often a combination of Alzheimer's typology and infarction that was important in producing a picture of dementia and that is the other quandary that faces taxonomous in that when you come to older people often it's multiple pathologies that really produce end result rather than one disease per se and that I think is another issue that classifications have got to address. So are we looking at subclasses of vascular cognitive impairment and one could say that okay there are these perhaps these subtypes of large infarcts or cerebral infarcts, multiple small or microinfarcts, strategic infarcts, cerebral hyperperfusion which is non-infarction related change, hemorrhages and cerebrovascular disease with Alzheimer's disease pathology which has been sometimes called as mixed dementia. Okay now where does neuroimaging stand and can we have neuroimaging criteria for this and of course when it comes to classification such as CDSM-5 or ICD-11 we have to accept the fact that neuroimaging will not always be available to the clinician and therefore you cannot impose those as necessary criteria. They have to be optional. The NIN's neuroimaging criteria for vascular dementia was evidence of relevance of vascular disease by brain imaging including multiple large infarcts or single strategic infarct as well as multiple basal ganglia and white metal lacunes or extensive ventricular white metal lesions or combinations thereof. But I think it also is necessary to point out that neuroimaging has another role to play and that is in excluding vascular cognitive disorders or vascular dementia in that absence of vascular disease on MRI and CT is strong evidence against the diagnosis and I think that's how clinicians often use neuroimaging. Okay I can skip this. One of the things that if you look at the NIN's error criteria and they address the issue of how much white metal lesions were really necessary and they came up with this figure of 25% that if 25% or more of white matter was showing hyper-indensity or abnormality on MRI then that was a suggestion that this may be sufficient to produce dementia. Now not everyone actually agrees with that but that 25% came from these two studies from a group which actually looked at people with less than 25% or greater than 25% or mild and moderate and in fact showed that people with dementia who had less extensive white metal lesions had a picture which resembled more like Alzheimer's disease whereas people with extensive, more extensive white metal lesion had a picture which was more like subcortical dementia but this has not been replicated by other studies so in fact there is no consensus on how much white matter should be, should show abnormality to make a diagnosis and in fact that's what the VASCOG consensus is that we have to leave it open and leave it for judgment and perhaps just say that large confluent lesions are probably what are necessary for the diagnosis. So I think what's happening with neuroimaging is that as newer techniques are developing especially say with diffusion tensor imaging you find that there is often abnormality in what looks normal on a T2 weighted, your traditional T2 weighted image. So the question is how far does one chase this kind of abnormality and there is no definite answer to that issue. So when I actually asked the California group, okay now what would you change from your criteria that you've suggested in the past, they said okay we will include executive disturbance as probably an important criteria. We say that two lagoons, one or two lagoons may not be enough, you probably need more than two lagoons to make a diagnosis and that severe white matter disease may be sufficient which wasn't something in the California criteria per se and that also that hippocampal volume in fact seems to suggest that hippocampal volume may not be a good indicator and of necessarily Alzheimer type pathology and their neuropathological studies seems to suggest that people with pure vascular dementia often have smaller hippocampal. So that's something that again needs replication but it's again something that goes against the conventional system. Okay so and when I asked Gustavo Roman about the NINZERAN criteria he seemed to suggest that we have to look at several blood flow and increasingly we have measures of several blood flow as to introduce them into the criteria. So where do we stand now then with proposal for DSM-5 and in terms of consensus on the criteria for vascular cognitive disorders. The first point is that there is acceptance that these disorders lie on a continuum. So you need to have mild disorder kind of diagnosis in addition to a dementia diagnosis and that how you define that mild neurocognitive disorder there's no consensus on using precise cutoffs but DSM-5 suggests between one to two standard deviation for age adjusted means for this diagnosis and only mild impairment in daily functions whereas the question of the major neurocognitive disorder or dementia is greater than two standard deviation. Now again I haven't gone into this issue of the term dementia should that continue or should that be dropped and be replaced by major cognitive or neurocognitive disorder I think for DSM-5 will probably be retained in parentheses. In fact the vascular community vascular cognitive community is quite happy for dementia to be dropped they were never happy with dementia per se but Alzheimer's disease community is strongly opposed to dropping dementia and generally neurologists and psychiatrists feel that dementia is a very useful term to be retained in the classificatory system. So the proposed criteria would have of course the diagnosis of that syndrome and supportive clinical features that suggest that this may be possible vascular etiology including if it could be a stepwise course or it could be a gradual course in which case certain cognitive dysfunction such as speed information processing and frontal executive functioning would gain prominence. And then neuroimaging evidence which basically is a sum of what I'll be discussing and could have either neuroimaging evidence or neurological evidence to suggest that there is significance of vascular disease and some supporting features such as gear disturbance or falls, urinary problems, suitable for palsy, et cetera. And exclusion criteria suggest history suggestive of another disorder or neuroimaging suggesting lack of survival disease or other medical conditions being present. And then level of certainty of the diagnosis from definitive which was histopathologically proven or having a genetic basis or probable or possible. I think I've covered most aspects and of course recognition that there could be multiple etiologies and could be so diagnosed in the classification.