So the next speaker is Clouse, he's actually my boss and I have a microphone, has it feels exactly oddly powerful position to be in, but I promise not to abuse it and change originally in Bonn. And then his Takakura training was in Aberdeen and his first academic post was in Edinburgh. But he's been with us in Oxford for about nine years now. And I think he's going to talk about the research that we're involved in the Y2K. Thank you. OK, now for something completely different.
Well, not completely different because it's about people getting older. One of the things I started when I arrived here was linking up with Whitehall two study. And this is not really a study of civil servants as such, but see them as a group of people initially about, oh, there's the food bank is missing. And yes, I used to get so initially they were 100 people who worked in any of the offices in London associated with government.
And the reason for this study was that people wanted to find out what the risk factors for heart attacks were. This was an 85 or so. And, you know, both of you with a medical background, you'd be familiar with the Framingham study, you know, where our whole community was followed up and people looked at various potential risk factors to see what predicts mortality and particular heart attacks and similar cardiovascular incidents.
So in a way, this was trying to replicate this. And that always been this assumption that if you were in middle management, that's the worst place you can be because you're kicked from above, but also from below. So work stress is at its absolute maximum and Whitehall did really for the first time.
And that's why it's so important to make sure that the lower down the pecking order you are, the higher is the the various risk factors and the higher the likelihood that you are going to suffer from cardiovascular disease and risk factors to identify where those we all know about them. Now, you hear about them in the news all the time. But this was identified in Whitehall that was continued into Whitehall, too.
And the study was extended not only, you know, beyond heart disease to cancers, chronic lung disease, other types of diseases, but also depression, suicide, sickness, absence, back pain and general feelings of ill health. And to the general purpose of the study is to take a very detailed inventory of people's habits about the general state of health, about their work situations. You know, are there where are there in the hierarchy?
What are their their social networks? What quality does their workplace have in terms of potential stress causation? In fact, initially it was called the Stress and Health Study. So it's all about identifying what's generally described as stress and the effect on on ill health. And from an epidemiologist epidemiologists point of view, really, the idea was to to build models, mathematical models predicting health outcome by all sorts of observations that are taken throughout the lifetime.
Now, the way the study was organised was that it started in 85 with 10000 odd participants, and they were then followed up every few years. We're now at phase 12. But, you know, this is just the first 11 phase of it. And you can see that the average age obviously, you know, goes up and in between each of the large inventories, as it were, where the person they examined are about five years.
And in order to to cover all important factors, they would, you know, collect the standard descriptive values, then general habits, you know, like alcohol, sleep, diet, smoking, et cetera, some health markers, anything from you know, from from body size, weight, et cetera, to some biochemical markers, inflammatory markers, blood pressure, cetera, then the social environment, the amount of support at work environment, um,
details about employment in particular when they retire and what the activity level is after retirement. Lots of health outcomes, which partially would have been linked to other databases, like, for instance, if someone. By way, trying to track down the death certificate and find the cause of death and then a detailed physical workup with a number of examinations, cognitive tests, memory tests, etc., physical function tests that people have to walk, have to breathe and to diagnostic apparatus.
And so a more detailed psychiatric assessment if they are entered into the oxygen component of the study. So what we did was we got people's agreement to be contacted and said the way we did this was just select them randomly from this group. Obviously, if you pick people who volunteer, you get a certain type of people, maybe they're likely to be better educated, healthier, et cetera.
So we selected them randomly and then they were contacted at phase 11 and asked, would you be prepared to go to Oxford and take part in this? So we have a list to work for. We for them up. We screen them. If people have metal bits that conferences have an MRI scanner. So we have to exclude a certain number of people. And then we arrange a visit to Oxford.
They come, you know, having filled in detailed questionnaires and we have administrative stuff to do, like consent to investigations and so on, that have a detailed psychological test, actually looking at the various components of mental functioning memory, concentration, executive function, et cetera, that have a systematic psychiatric interview at that age to on average, about 70.
Now we're particularly interested in cognitive and dementia, but also in depression, because that's even more common than the cognitive impairment in that age group. And then to have a detailed MRI scan, which may give you details of what we actually looking for in a minute, plus various tests to have another walking test. We test we take blood and saliva for people who agree to this test, all sorts of other things, including, you know, the bacterial makeup of the jested trap, et cetera.
And then when they finish, they get a little stressful to a lot of experience, which hopefully wasn't too unpleasant. Um, right. So what we looked at, for example, with the MRI is we looked at the structure of the brain. That's a typical MRI scan, looking from the front, from the side and from the top. And so because we are going to process about 800 of these imaging modality, you can't do these things by hand.
So it's done automatically. In Oxford, we have Premarin, which is the centre for ephemerally, which is mainly used by experimental psychology and psychology to some extent of neurology. And they have expertise in analysis software that deals with this. So what we can do, for example, is we can extract the brain, discard all the other tissues like bone and skin and so on and so forth.
And then within the brain, we can break it up into grey matter and white matter and cerebrospinal fluid, and then we can go further down and identify certain structures, for example, basal ganglia and all that can be done more or less automatic, systematically, quickly and in an objective fashion. And then we find one important structure in the brain is the hippocampus, which is the structure which is at the inner side of the temporal lobes, and that's mainly associated with memory.
And that's, for example, the structure that this tends to shrink in Alzheimer's disease. And if you look at the first world country, that is probably 500 or so and you can see that there are normally distributed. So there's a range within people and then at the left and towards smaller statistical competence, a little bit of a tail.
So as a subgroup of people who already have shrinkage in the hippocampus, they are selected in the sense that they manage to get to Oxford, managed to find a within the realm of that geriatric hospital.
So most of the people who arrive with us are not demented, but many of them already have memory impairment, even the ones who've driven to Oxford and quite a few of them have a reduced size of hippocampus, which, you know, we obviously would be interested in and try to relate to their positive performance as the same on the other side. Then we do some diagnostic scans. Typically, as you get older, you develop sort of changes in the structure of white matter.
You can see it here. There's quite a bit sometimes called white matter, hyper intensities. And so they are supposedly related to vascular changes, small vessels getting blocked, you know, micro strokes if you want to, so you don't have to have a stroke. But most people who are in this age group already have these changes. And the degree to which they have these changes may be related to cognitive function, but they're also linked to, for instance, to depression.
And more commonly, people become depressed. So we do a number of scans like that because the difference is that people start scanning. You can see something called the microbeads micro stroke. If you want to move a blood collection which can be picked up from this, then we are looking at so-called diffusion tensor imaging. That's really quite clever application. That's clever or glamorise clever in the way that it's clever because it gives a lot of information and you know that process again,
to get into normal space. And then we're able to extract a number of bits of information with top row here. What you can see the fibres that connect different parts of the brain with each other. So these are actually the weight of the fibres and the stronger the signal. This is the teacher that is called fractional isotropic. So that is the measure that tells you to what extent water diffusion in the brain.
This is confined to one direction. If you think about about a white matter that consists of the cabling of axons if you want to, and they're well insulated from each other like any electrical cabling. And if you if insulation breaks down, it doesn't function properly. But that also means that the water, if it moves in the system, moves along the direction of the axons, not across it. And the water's constricted to this one direction, the better the quality of the insulation if you want to.
So the FAA gives you a measure of how well the white matter is organised structurally to perform its function. And then you can look, this is another example of a modality. This just gives you the total amount of. If you're 17, you know, it's great in the ventricles, so the water can move in all sorts of directions and less so in areas where you have repetition. And so similarly to what they become politicised. I mentioned that the changes are quite common.
If you look at this measure of white man's integrity to go want to you can see it's, again, the moment to curve and there is a tail on the left side. There's always a subgroup of people who already have abnormalities in this particular measurement. And last, at least 12 people in this country for ten minutes not doing anything in particular to keep their eyes open, but they do what you do if you're lying somewhere, not having anything particular to do.
So they think about things that remember things. And intriguingly, what you find is that if you compare the correlation of different bits of the brain, what we actually measure is blood flow to be more accurate, the amount of oxygenation of blood that gives you an idea of which bits of the brain are active. And if you compare different different parts of the brain in terms of how they fluctuate over those 10 minutes, some of them will be correlated and others won't.
If they're correlated, you can assume that there's someone communicating with each other, OK, if they're not correlated, they don't. And if you look at the whole dataset, you know, the whole brain box, money box, live picture unit, that picture unit, and you follow that up over ten minutes and the scans are quiet all the time, you can then identify the networks that are connected with each other and those that are separate.
So, for example, here. You have a network that mainly contains this reading, the central bit of the occipital cortex. Can you read the writing up here? If you can't read it, what do you think it tells? Much, yeah, right, right out the back, that's where your primary visual cortex sits, and although they don't watch any film or do anything more active, that's part of the brain already.
It changes in parallel. And that may be I mean, they usually just look at someone at the scene or at the surface. That doesn't change. But clearly in the imagination what's happened in the same area. And for some reason, these bits of the brain already are connected because there's nothing special to look at. That would be the actual visual cortex that is there to process visual information and to be able to look at directions of movements,
look at shapes, etc. and colours. Um, this would be the system associated with hearing, but with movement. When you see pictures of this this homunculus, you know, this projection of the human body upside down, on top, at the bottom, in the middle of the brain, that's reflected in those two areas here. So this is the sensory motor cortex, which already although the lying there still with nothing is touch.
And then, you know, in spite of this absence of activation, these bits of the brain already correlate activity. And similarly, you have areas which are of particular interest to us, which are which tend to be activated when nothing else is happening. And they tend to move to the so-called default mode networks that will act as if you're depressed. So this is rumination. This may have something to do with activating those, but they are the areas that are more active.
If something is not engaged in the task as opposed to doing a particular task and focussing on doing that task. And then there are others which are more frontally which have something to do with executive function and planning, etc. So you can break down the various networks and the brain can see which networks are more active and less active, depending on, you know, risk factors or mental performance. OK, so let me just focus on a few questions that one could ask.
Looking at this data set which are relevant. The first three, we have the information which goes back to 95, which had the risk factors predict brain changes, you know, up to 30 years after the start of the study. How far back can we go to predict changes? Does it matter if, for instance, you know, we have a certain type of lifestyle or certain risk factors when we were in our 40s? That's a matter of what these risk factors were in relation to how our brain looks when we had women 70 years old.
And how quickly do these changes manifest? Now, this shows you the kind of data we're looking at, so this site and imaging data, which is done in Oxford at the end of the study, if you want to, and then we've got data, which I showed you before, five years, 10 years, 15 years, 20 years back from the scan. So essentially, we can look at those data and see whether we can make sense in terms of, you know, which ones are those important in how the brain looks when people are in the 70s.
And I want to focus on one particular type of risk. It's called the Framingham Stroke Index. Um, you may have heard of Framingham. This is the the town I think it's in, uh, Massachusetts. It's in New England anyway. And there's the whole whole town took part in the study and based on follow up and who developed heart attacks and strokes, they managed to put together as risk factors and predict whether you're going to develop a stroke. Now, this is something your GP has in front of him.
You know, when they tried to calculate your risk and see whether you should be on a statin, for example. And that takes into account all sorts of factors. Age, sex, men are at greater risk because we would expect compared to premenopausal women and then other things like blood pressure, like being overweight, like blood lipids, you know, all those risk factors that we know about.
So I thought, but let's look at this this risk index and see how long the effect lasts and how important it is for what the brain looks like at the moment. Now, if you think about in theory, what should happen is if you look at the risk that is closest to the scan, if there's a correlation, that's likely to be the biggest. But then if you go back in time five, 10, 15, 20 years, of course, things will have happened between between those measurements if you measure the risk at phase three.
Twenty years ago, lots of events happened in person. I must have put on some weight or increased blood pressure or, you know, the risk profile would have changed. So if you look at the correlation, it should be the highest closest to the scanner and that should probably go down in a cold, dark time.
Does that make sense? And then if you want to, this, of course, could mean that changes happen in all sorts of people and, you know, that correlation may be as high and lower than that because a subset of people have changed their body. But that reduction could be due to other people changing. So you're not quite sure whether this is actually a change within people or just between different subgroups.
So what you can do is you can look, this is the uncorrectable correlation at Face 11 and you can check of how much predictive value remains once you correct, for example, for the first before you do a partial correlation or regression because face the risk of the first nine and 11 are very similar. Once you've accounted for everything that's predictive at phase nine, there's very little left about the intervening time can add to the risk.
So that's correct. The correlation is quite low. And then the further back in time you get, the more the similar the correlations become. And then you know about phase three, you would have had the contribution of quite a substantial contribution that is made between phase three and phase 11. And that will be phase one percent. Make sense of it. But if you think about those correlations being similar to each other, if you do a partial correlation, you remove whatever variability is there.
You basically are left with very little and then it goes the opposite way the further back you go. And then if you look at the correlation over time, sometimes you may find something like this. So the further back you go, the less it's, you know, the risk is correlated with how the brain looks. But then there's a sort of falls off when it goes down a little bit more closely to the to the actual scan. Any idea what that.
Exactly, yeah, so these are risk factors and the risk number is essentially the percentage risk of having a stroke in the next 10 years, you can actually, you know, calculated down to that to that specific meaning. And of course, they don't they don't have to have a stroke. They have other changes in the brain which may be more sensitive, be picked up by the scan, but it takes a while for the risk actually to manifest itself.
So, you know, you you would expect that the closer you get there may have been you know, the risk is not yet represented in the pattern you perceive. OK, that's the theory. That's the practise. So if you look at those those pictures up here, they're essentially show your areas in the brain where the grey matter is correlated with the risk at phase three five seven nine 11. I'm just looking at them in particular in the small form, look very similar.
But they are in fact, if you take the average correlation across the whole body and it's sort of more or less at the same level, if anything, those two are a little bit lower than those two sets of sort of trend for it to get higher. But by and large, the correlation stays within the same range, a significant correlation.
So the higher the risk of developing the stroke of any of those stages, the greater the atrophy and brain matter at age seven, been at the time of the Oxford scan, or if you put this in more drastic terms, the risk you have for the Framingham risk you have when you're 50 years old already significantly predicts the degree of brain atrophy you have when you're 70, OK, and that stays the same. But of course, this is across the whole cohort.
You don't know who's contributing to this. So we do this exercise with partial correlation anyway. And we actually find that if you correlate if you correct the phase eleven risk again for phase nine, it goes down. And if you correct the full phase seven, it's higher again. So you get this this negative slope, I suggest it will be part of the time dilution effect. And there are a number of interesting things to note. One is that the risk at Phase 11 and phase nine are sufficiently similar.
Once you correct the Phase 11 correlation for the first nine, you've got nothing left. OK, and then on the other hand, if you then go back already from phase seven, along with those other green areas from phase seven onwards, the the risk that accumulates from then onwards would be reflected in atrophy of those bits of the brain. OK, and the same is true for the earlier ones. So let's look at that in more detail. These are the same two lives of the top.
One is the the correlation of the risk 20 years ago with the brain scan now and 15 and five years ago and immediately beforehand. And this is the risk from the most recent phase.
Correct. So there's nothing there. But then 10 years ago, between 10 years and the presence of the additional risk, the risk that accumulates this particular found in these areas now these areas here, any suggestion that it's those two areas of the inside into the temporal lobes, if a complex find, so that changes the more recent changes occurring in the campus.
But then if you look at the top part of the brain in the frontal cortex, you can see that there are correlations all throughout the places. And in fact, even if you correct for the phase three risk, you remove the face of evidence. In other words, the risk of having atrophy in those parts of the brain is already mainly determined 20 years before the scan. So you can start teasing apart the bits of the brain that are responding to the risk and what the dynamic is in terms of time.
This is a similar plot looking at bitmap so you can essentially reduce the white matter to a so-called white matter skeletons. The skeletons essentially are just that, a representation of the main tracts and. Roll again, you have the risk of face three five seven nine 11 correlated with fractionalised sexual predators, the quality of life, not a chance. In other words, the greater the risk, the poorer the quality of things.
Women, I'm the tracks in, you know, from those faces. And interestingly, the risk at phase three doesn't seem to have any effect at all. Things have happened in between. So the actual risk putting of phase is not correlated with the brain. 20 years afterwards, about 15 years, it starts having an impact and so on.
And if you go the other way around, if you go back from phase 11 and correct the phase nine, rather than like in the previous case where you have nothing left actually between phase nine and 11 in those five years, they are reacting to changes already contributes to what marked a deterioration in those areas here. And then if you go back, it becomes more and more. So the short summary of that is we looked at three different types of tissue, if you want to.
There was white matter and there were two areas of grey matter, grey matter. This would have been the upper part of the brain. This had been the campus. And it seems that the campus, the important areas are the most recent ones from phase nine onwards, for the grey matter.
You know, in other parts, if you want to add neocortical areas, the important phases from the beginning to phase nine and not so much dead and white matter the importance effect start, but phase five and then go through if you want to. This reflects the plasticity of these various tissues we know about. White matter is pretty malleable.
You know, the Black Panthers and dependent on so-called glial cells, which are not those cells that can grow and regrow, that can form scars if you want to, but they can also adjust the wiring of the brain. That's part of the plasticity that happens in the brain. You know, bits of the brain still functioning, but the connexion becomes switch over and kind of go to a different area and the bank are compensating. So the white matter is pretty variable and plastic.
So the you know, it starts what happened 20 years ago is not recognising, again, that the other extreme is grey matter, where the cells seem to be fairly static. So the risk here is recognisable as the term is determining the pattern of brain matter.
And follow up an extreme example for this most recent study, which came from the millennium cohort, where people were able to follow up a group of mouse 70 year olds who have the intelligence test when they were 11, OK, and they found that the IQ distribution at age 11 was cold at a certain amount of structures at age 17. So that's the extreme of how, um, how stable dramatic patterns can be.
So it allows us to identify areas of different risk and put the difference between hippocampus and neocortex. Any suggestion of why that should be the case in New Haven should be much more plastic than. This is one part of the paper we know about continuously, new brain cells that generated generate the plasticity becomes actually one of the highest in the brain.
If you feel like you should enter a London taxi driver, of course, and have an MRI at the beginning of the end, at the end, it becomes a much bigger event. That's over a few months. So this is this in a way reflects the fact that it encompasses one bit of the brain that's most plastic. OK, go to. It's good. OK. And then you can go one step further and you can look at how the performance in various tasks is related to brain structures. This is how it looks before it's all tidied up.
These are all the various risks and phase three, five, seven nine, and they all correlate with each other. This is the hippocampus because I'm interested in the campus because that's changed in dementia to some extent in depression. And this is one of the tests that is thought to depend on income for function. This is the this is a verbal learning test. And essentially it consists of 12 words which are sort of fairly random, but not connected in any sort of logical sense.
So the idea, at least theoretically, theoretically, is that you have to learn those 12 words. So they read to you. You have to repeat them. They read to you again, you have to repeat them again and for a third time. So in a way, it's a very simple learning test about just giving this word was to learn. And generally people get better as as they go along. So three times 12, the maximum score would be thirty six. If you score lower than 18, you're likely to have a problem with your short term.
So that's the test and that tends to be related to the campus. And if you tie this up a little bit, we find that hippocampal size is related to the risk at phase three. But then, as you saw, there's most more recent contribution as well. And the first time it becomes significant because then phase 11. So there's an initial relation to the risk and then the later one. And then there's also a direct effect of risk on performance of the learning task.
So you can actually model the causal relationship between how in terms of how the vascular risk that people have is reflected in brain structures and to what extent the change in brain structures affects the performance of particular tests. And this are presumably has something to do with how this risks this risk affects other bits of the brain, but also for the test. So just one example, example of how you can tease this further apart. Right. The next important question is how?
But are there any characteristics that predict good mental performance and do these protect from the effects of brain changes? If you think of what happens during the lifetime, the risk factors on the minus side. If you want to add that protective factors, for example, higher education or IQ, certain drugs, particular diets, supposedly moderate alcohol consumption, but then also mental and physical activities, which social networks, the business, as it were.
So that about factors and the pluses and the negative side. So let's focus on education. This is education in terms of highest and qualifications, all levels of A-levels and so on. And this is the year of full time education, as you would expect, the higher achievement amongst we've gone to school, but also it's also correlated with the actual IQ of this case, a particular measure of IQ that's unlikely to change with age.
And then you go back to our learning test, you perform that task work, learning tasks, the better the higher your initial IQ is. But also there's this correlation with hippocampus. So the bigger the volume of the hippocampus the brain, the better you are performing these tasks. And if you look at the correlation between IQ and the the campus and the learning task, you can see there have some correlations.
And when you remove the other relatively other factor, it actually increases from zero to the partial correlation. Or in other words, if you just look at the contribution of those two predictors, hippocampal size and IQ, if you just add them up, you get 10 percent. If you actually look how they interact with each other, it's almost 13 percent.
In other words, the best thing is to be interaction between those two factors in the sense that if you have a small IQ, you sort of you have a small hippocampus, a greater IQ is going to make a relatively bigger contribution to your performance. So IQ can actually be shown in medical terms to be a factor that protects you from poor performance. And then finally, that action in particular changes in characteristics in the brain that pertain to IQ.
Of course, it's very abstract. We don't know what it means. And one you recognise and one approach which I have thought about, is that you and what you have to do is you have to look at people with a small hippocampus and then you compare the two groups, the one group that performs very well in this case on the verbal memory task, and the other is the group who don't perform well. So they both have the same degree of hippocampal atrophy.
So they both have the same degree of organic impairment, if you want to. But then there's a well performing group and a poorly performing group. And if you compare the white man for the cabling on somewhere between two groups, you find that the Musladin group here has a much higher quality of life. The Connexions, in other words, a good wife, a good a good connectivity, and the brain compensates to some extent for it will come to actually be interested in what actually protects you.
And that would be certainly one factor. And, you know, she thought off this index and that is the bigger the verbal memory score and the smaller the hippocampal volume, the greater resilience. So given the particular hippocampal volume, the guys with a higher memory score are more resilient and vice versa.
So after this couple of months, if you look at this, this resilience index, I think what you find is that the areas where people have high points of the green, it's just this character and those it's just a show where changes can occur and the orange and red are the areas where are significant relationships.
So in other words, where the orange colours are, there's a significant positive relationship with resilience or say, again, if you have high quality white kinds of connexions in those areas, you're more likely to perform well and given a certain size on campus. And, you know, we always be worried in case this comes about by confounders socially excluded, things like total grey matter density, stroke risk, alcohol, and you still have this relationship.
So the areas that have better quality by Connexions and the guy is going to have a little. So that's nice. And then if you add the last one that we had shown before is related to performance, once you remove the effect, I actually that. So this is not the final proof, but it looks as if maybe the connectivity, the connectivity of the brain is partially the reason why people have a higher IQ measure.
And both of them together obviously will increase their chances to not present with memory problems in spite of already some developed brain changes. And then the very last thing is that as you get older, you're the front of your brain becomes more active. And that's because you already start getting to work harder to achieve the same thing. Just put the next two minutes. Do you know the the, um, that there's a game called Timboon Family Game?
It's a bit like Sharath, but it's not that you have to act out and talk and you describe something, but you have to avoid certain words. Okay, so if you, if you let's say you have to get people to get out of the room, you must not mention hopping Australia and some other words that would obviously be connected. And if you play in a family context, any book, anybody or anybody over 30 is going to do much worse than the younger members of the family.
And it's really quite entertaining and it's very gratifying for children. But that may be one sign that as you get older and your frontal lobes have to work harder. And that's actually not bad, because if you look at particular tasks, you know, they've got to look somewhat active as you get older.
But if you look at specific diseases like Alzheimer's disease, and then you tend to find people who are equally impaired, equally scoring in terms of their dementia severity score if you want to, and you can divide them into people who have high levels of education, the ones with intermediate and the lowest education, you find that in order to present with the same degree of dementia, you have to have a much poorer brain function in the areas that are normally dementia.
If you already have a high IQ and then as you you know, the sort of low, low education groups need much less of a brain change that could be changed to have the same degree of dementia. And what you also see is that particularly in this group, but also in the frontal lobe, seem to be more active in those, you know, after two groups. So it's almost as if the activity of the frontal lobe can compensate to some degree for the established brain damage that comes about the dimension.
And that's maybe one of the reasons why you should activate your brain if you get your frontal lobes to work to some extent, up to a certain point, you can compensate. OK, and of course, it's not my my work and many other using other people's efforts. Mr. Confectioner's.