 Thank you, Scott, and thank you, Julio, for the invitation. I must say at the almost the end of the day, or just the precursor to Huio's presentation, you know, discussion of translation. I love that expression Huio has in his editorial about being a cesspool. I think I have really descended into the cesspool of psychiatric research in recent years. I've fallen in a very bad company, along with Pat McGory and colleagues, in really trying to focus on areas of risk to onset and illness progression in the course of mood disorders. I don't want to discuss at some length. But I was really impressed at the start of the day with the presentation by Helen Christensen and Kathy Griffiths about the translation that we don't often talk about, which is actually the wider public effects of introducing, in fact, mental health programs at the population health level. And when people despair about translation, I think we should remember some of the successes, and I think in Australia and elsewhere, we've had tremendous successes at the population health level in terms of people's understanding, in terms of access to care, and I think it's the work I've been associated with Wayne Hall and Phil Mitchell and others in reduction in suicide rates associated with the provision of care and the expansion of health services. I think the cesspool we all refer to is actually the application at the individual patient level and advancing our understanding of really clinical care on an ongoing basis. When I was the CEO of Beyond Blue, I unfortunately spent a lot of time with Pat McGory in Melbourne, who persuaded me that really a lot of the focus of the traditional work that I'd been involved in the mood disorder area and elsewhere focused too late in the course of the illness and constantly trying to find unique and homogeneous groups amongst people with longstanding and chronic illnesses was probably not only unlikely to reveal real truths, from the point of view of the individual people you're looking after, it often missed the boat. He was also of the view at that stage that even for the psychotic disorders that he had made his reputation in, that really there was not that much that was specific about many of those disorders. And certainly when you saw those young people early in the course of their illness, it was very hard to say where they might go. And around that time, the Victorian burden of disease study had been published and particularly the incident rates, rather than the prevalence rates, are an interesting picture. That Everest of onset in the post-puberty period through to the early 20s, depending on how you look at it, either impresses you with each of the disorders that arises or if you forget the colour schemes in there, what strikes me and others is the commonality, what must be going on that is common during those periods that actually translates into all those particular disorders or shades of grey or shades of orange in this case that make up our diagnostic and treatment manuals. Adopting the language of both Pat and of Tom Insaw in more recent times in the NIMH, I think the focus of many of the much work in this area is much in the affective disorders now as in the psychotic disorders is really what is going on during that onset period. And if translation is really going to mean anything, it's going to mean the identification of people during that period and the provision of interventions that have the greatest chance of decreasing the likelihood that they will go on to ruin lives. And in some way, we will, out of that mess that is the current system, be able to individualise our selections of treatments and make serious recommendations. It is, in a sense, I think, a realistic step. Phil was just talking about the difficulties of actually looking at at-risk populations where people have not yet presented for care and the really great challenge of actually being able to identify anyone before they actually have an expressed illness. This kind of version, I think, gets around that and cuts to the chase. These are people who are presenting in various ways, illness phenotypes, even if that is unclear, and are already, as I've shown in the data, we've been collecting on a pathway to considerable disability. When you look at these particular areas, you look at it in the course from the point of view of the traditional clinical psychiatrists and the groups that we have, one tends to be preoccupied, as we've heard throughout the day, with schizophrenia, all with bipolar disorder, all with unipolar depression, all with anxiety disorders, all with 101 other specific categories that we've made up on a phenotypic basis. I think if we all agree about anything, I'm not sure there is anything we all agree about, is that that system certainly hasn't tracked on to pathophysiology very well in any meaningful way and certainly hasn't given us much that is in common, as in the discussion and Phil's presentation just took place, the issue of comparing one disease group with controls has become less the issue than comparing certain disease groups and whether they have anything different or not. I think it's Phil Elegantly showed from the work of Naomi Ray and others when you look at the general genetic and family risk factors, much more seems to be shared. That has become the focus of our particular work, is trying to look for processes and pathophysiologies that may actually be shared. And of course, what we're looking for are reliable targets that may reflect in some biologically meaningful way the underlying pathophysiology that may themselves become targets for meaningful early intervention strategies. So the studies that I'd like to talk about are two, and I've been very grateful to the process of the Australian Fellowship to be able to sink good money into Nick Martin's endeavours in Queensland. I think we have things in Australia that really are of great value, and the twin registry is one, and Nick's work in developing longitudinal twin samples is particularly important. I've been very fortunate of working with Nick over some considerable period of time in association with the Brisbane adolescent longitudinal twin study, which looks at genetically informative populations across that key onset period. And increasingly, to introduce into those studies a range of technologies that might help us to understand what is going on. There are other large population studies we're engaged with, and one is the drive study through the George Institute in Sydney, looking at effects of sleep cycles with regards to maintenance of affective disorders. The other big set of work that I've been involved in in recent years, and I was just saying to Colin Masters, who's done good money and unlike my temperament, I don't think we're going to answer out of this for a good 10 to 15 years of seeing whether it's been a good bet or not, has been to say, really, we don't really have the health systems to investigate the questions in which we are most interested. If we're serious about investigating these questions during the period of onset, then we need health systems that engage with young people when they get unwell and study them in terms of their clinical progression, the longitudinal samples. I have a comment earlier in the day. We're constantly comparing people in cross-sectional studies from different stages of the illness. It doesn't really answer the question of what happens to the same people longitudinally, whether they were different to start with or whether there are questions that actually change the consequence of their illness. That's basically what we've been about through our Headspace Clinics nationally, which have not only been about improving services for young people, but really are creating a series of integrated clinical platforms, much as one sees in the cancer trials areas with larger samples over the course of illness, and trying to study within those particular groups aspects related to the technologies that we currently recognise. Obviously MRI in terms of brain structure, spectroscopy in relation to the chemistry of some brain structures, my own particular favourite interest in circadian systems, and then issues related to cognition and neuropsychological function. Just to talk a little bit about the twin studies, Nick Martin educated me a long time ago about genes. We're not genes at any particular point. They were points across the whole life cycle that different genes come on and off, of course, at different points, and different genes may be relevant in the early phases of life as in later phases of life. And that not only genes may be relevant, but environments may be quite different at different stages of life. And the continuous process of simply lumping people together all the time, independent of the stage of life they're at, let alone the stage of illness that they had has the tendency to make it less likely to find factors. In that work we were associated with some years ago, what was really interesting was the extent to which you had common factors that were operative early in the genes in the environment, but new sets of factors in mid-life and later life that were affecting age of onsets. It's very interesting in the constant debate about types of depression. I think the one for which there's the greatest validity is simply early onset versus late onset, but it's probably the least discussed in most of clinical psychiatry around depression, particularly in this country. And that the genes associated with depression have their maximum effect quite early on. Actually genes related to vascular disease that become increasingly likely to explain your risk to depression as you age. I think Jin made the comment earlier in the day he hoped in his lifetime to see one disorder that was explained in terms of pathophysiology. Jin, you don't have to continue to live much longer. I think there is one. It's called late onset depression, where actually I think we have a good idea about the genes, the pathophysiology and the presentations during that particular issue. Kind of interesting that we don't necessarily think about it as being a psychiatric disorder in that sense, even though it presents with depression. Almost sometimes when we find the underlying pathophysiology we have a tendency to give it away to some other medical group as we've done with epilepsy and I think in the cognitive disorders we may see that happening. But I think late onset depression is clearly within those sets of disorders that Paminda was discussing earlier in the day in terms of understanding the pathology and I think increasing the genetics and increasing the chance for actually prevention of late onset disorders. At a population level the early onset ones I think adding to that work in a considerable depth has been the work of Ken Kenwood to say look, adolescence isn't one simple period either. If you just look at the number of genetic transitions that are taking place during the genetic period using Swedish twin data, Scandinavian twin data, he's modeled a number of different genetic factors that are operative and those that are necessarily relevant at 8, 9 and not necessarily those that are having the largest effect by the time you're 20. And there are new genetic elements throughout that particular period. If we attempt to sort of lump adolescence together it probably doesn't stand up at a genetic or obviously at an environmental level in terms of the factors that are taking place. So what Nick Martin started doing back in 1992 was initiating an association with his MOLS study studying melanoma in Queensland was actually using the opportunity to start to look at cognition and psychopathology in adolescence. And so it's a set of studies now that starts at 12 and runs right through in those aged approximately 26 at this point in terms of this long set of questionnaires that are now being done both online and then followed by detailed interviews. But the important factor is the numbers that have actually been followed and you can't quite see the small numbers here I'm sure but already 2,900 individuals have gone from age 12 throughout the whole study and then by the time the study by 2015 there'll actually be 3,500 3,800 who come in from age 12 onwards. We've been increasingly introducing MRI, circadian measures and cognitive measures throughout these sets of studies so that we start to look at the different brain pictures as well as the onset of different forms of psychopathology. And if you actually start to look at the bits I've been most interested in for a long period of time we now have extensive data on looking at various patterns of anxiety, depression and somatic distress. One of my own particular preoccupations the somatic bit particularly being tied up in looking at sleep disturbance and fatigue states one of my own interests outside of the adolescent area in the general depression area and the extent to which that expression of distress is often underrated as distinct from the psychological expressions of depression. This is really driven also by the work of Jules Angst and the extent to which having those expressions in adolescence particularly of fatigue states may well be associated with lifetime risk of bipolar disorder as you age but the phenotype might be quite different during the adolescent period. So that's a data set now of almost 2,500 adolescents where we've been collecting data throughout that particular period looking at the genetic and environmental nature associated with that. And the classical way looking at the extent to which MZ and DZ differences reflect genetic differences. Quite interestingly in the extent to which you get what you would expect for these sorts of situations have strong genetic effect of the way in which the MZ and DZ rates are approximately twice those of the DZ rates. But more importantly than simply looking at the models what we've really been trying to do is look at male female differences and look at the extent to which these periods and these patterns change over the course of the adolescent period by frequent sampling during those key years. Just to make things really complicated of course when you put all the samples together there are big effects not only of age but of sex and increasingly we've had to treat the female and male samples differently from the point of view of both the brain imaging and other sets of factors. The numbers of size and scale become quite different but again in a lot of our studies and a lot of our studies in many of these areas we simply have a tendency to lump together the males and females as if they were the same. Really work led by Margie Wright in association with Paul Thompson and her other US colleagues we've now been able to increasingly look at in fact the MRI findings that have started to emerge out of this study so you see the numbers have started to get quite large which is really very helpful in terms of understanding factors and Margie's been able to produce these sets of heritability charts now for DTI for white matter tracks. The interesting thing being what this is meant to show is the genetic influence actually being greater in younger adolescents than in adults. They've gone on to actually look at the interesting kind of sets of factors when you get to larger data sets like this you can start to look at the differential effects of sex and particularly in these particular groups socioeconomic indexes as well as age and it's really the socioeconomic when I was to dwell on for a moment what Margie's demonstrated quite eloquently in this data set is the extent to which of course as you enhance environments in this case as you get wealthier then actually genes explain more of the variants. When the environment is adverse genes explain less of the variants and in terms of starting to unpick the variety of factors that we actually need to be dealing with in understanding what is happening throughout these periods this starts to become an issue. Additionally for those sets of issues like we've come very interested in and we've heard already about today in terms of brain imaging studies as part of the fMRI data out of the same sets of data sets in looking at what is actually happening in the anterior singular and elsewhere you can be able to look at issues like the heritability in normal adolescents as they move through these periods. We've also been able to look at the data increasingly in terms of patterns of depression and my personal favourite is the atypical depression as they emerge and the extent to which they may be characteristic of a different class throughout these sets of periods. The sort of data that we get here is very similar to the data that Kathleen Merrick-Hangus has been getting out of the national co-morbidity studies of adolescents in the United States and the extent to which an atypical picture in adolescence is more likely to be associated with bipolar disorder in the long run but also with more severe depression as you age and that the change in phenotype over time is quite important what is actually linked. There's been a series of other sub-analyses already done out of these studies and I won't dwell on this except to say that in these sets of studies the chance of using cannabis in association developing a later psychosis is strongly predicted by psychoticism scales at age 12 suggesting that a temperamental factor is actually driving the cannabis exposure which in this particular study is associated with psychotic symptoms on an ongoing basis and it's a good example of just simply being able to start to look at through these longitudinal steps considerable more detail about interacting factors than one would from traditional cross-sectional studies. To move from the Brisbane adolescent twin study to other sets of studies, this is a work really coordinated by my colleague Nick Glosier in looking at short sleep duration. One of those big social factors that is out there at the moment is changing sleep patterns in young people in association with the lives we lead and the technologies that we now have access to. In association with the drive study which was set up by the George Institute in New South Wales in Sydney took at risk to motor vehicle accidents Nick had the chance to look at length of sleep versus likelihood of developing psychiatric disorder in this case psychological distress where shorter hours of sleep was associated with higher rates of distress but most importantly in this particular situation that lower hours of sleep was associated with increased odds ratios of persistence of depression. I think it's one of these other big characteristics that we're really interested in not just what gives rise to risk of onset but what other factors are running that might give rise to in fact persistence in this particular group of 17 year olds and found strong evidence of the extent to which you were actually sleeping less during those periods being associated with 12 month risk of persistent psychological distress. So we have a number of sets of studies in normals in twins and then the population which further inform our work actually at the clinical end. So at the clinical end in dealing with the individual patients we have the problem of simply having numbers but also really wanting to have a better sense of what might be those common underlying pathophysiologies and hence as I discuss with many of my colleagues every day it's kind of difficult when you go to publish the stuff because we're not too preoccupied with which diagnostic system the people it's not that we don't record those particular factors but we have a fairly open door policy and in a sense to answer Beverly's question from earlier on everyone are controls all those groups are in and we don't pre-select to leave out childhood risk factors as well I mean clearly those who work in child psychiatry are well aware of the extent to which people who have childhood onset disorders go on to have further problems and often an elaboration of their disorder in adolescence as well as those who arrive in adolescence don't know though so we don't pre-select it or out for the presence of earlier childhood disorder or developmental disorders I think comes and I'll show some data and answer partly to the question Jim was raising earlier on one needs to know for a lot of imaging studies and other studies what is the pattern of cognition coming in or what is the pattern of brain change before the onset of a number of these disorders and are people really equivalent or have some people been different before they develop the illness of particular interest obviously where we're really headed and I think the key issue in translation that is achievable at the moment in psychiatry are those or potentially earlier interventions and that is of course the most controversial everywhere else in medicine has been alluded to a number of times today people no longer wait for the first heart attack or the disseminated cancer they are obviously trying to find those earlier points whether the management of obvious risk factors or earlier phenotypes that increase survival or reduce disability in all of those other areas many of the debates that we have about the relative merits of intervention also arise I'd suggest however that we've been relatively slow to move out of our late onset clinical environments to actually taking that view that actually assisting people earlier on although it may not be great for simply studying longitudinal samples is really where the translational challenge lies so we've been associated with the development of the headspace services around the country 30 new sites around the country and depending on what happens in next May's budget it may go to 60 or it may go to 90 sites around the country starting to create new service platforms that are distinctly designed to engage young people in healthcare for appropriate mental health problems at an early stage and then in association with our more specialised sets of clinics at the BMRI we've added sets of technologies particularly to those studies in an attempt to better characterise what is happening at a neurophysiological or neurobiological level we've also recognised the need to try new treatments and when new treatments come along that have a particular kind of basis whether they're old treatments to find out how they work testing or new medications that come on the market with different means of actions then we need to particularly be concerned with the chance to investigate those in younger people particularly where the side effect profile may well be favourable at the same time we are constantly in search of markers not just simply of risk but of active pathophysiology we're really interested in anything that can mark a critical period or mark a response to treatment or a cycle of really testing and innovation may be considerably speeded up in our particular area and obviously we have really suffered from the fact that we haven't had not any good reliable markers that we can use in that way in our own work we tend to be preoccupied with structural changes in depression the extent to which there's atrophy and the extent to which that's correlated with years of illness in a circadian area we tend to be preoccupied with the extent that we can actually map and have measures that we can now use in effective ways and the extent to which we have novel interventions that we can actually potentially use to test these systems our general kind of strategy is the extent within individuals that we can use multiple markers over time as shown in this sort of schematic and we don't have a simple belief system that any one particular marker may necessarily solve this riddle but we are kind of interested in the extent to which we may have markers at different periods of time to help individuals particular profiles that suggest to us they are in a particular phase of illness a critical part of this whole thinking has been the collaboration with Pat McGorrie which I think Michael Burke referred to earlier today of accepting basically that simple clinical notion of a staging model so popular in cancer and other areas of clinical medicine and it is so obviously the case of course in most major psychiatric disorders that that which we are most comfortable in the later stages where we are confident about the diagnosis did not come from nowhere that many of the earlier phenotypes were able to be recognised and have been tracked for long periods of time in the particular model that we are preoccupied with it's particularly the transition from ultra-high risk now in this case not meaning just ultra-high risk to psychosis but ultra-high risk to transition from a disorder that does not seem to persist or does not seem to be headed down an deteriorating course but from disorders that seem to persist and then recur so we've done away essentially with the distinction necessarily between what they may necessarily be as to simply a notion of our severity and the extent to which at a brain level they may be associated with persistence or self-perpetuating course much in the way that Michael Burke showed this morning around recurrence of bipolar disorder or recurrence of uni-polar disorder or you see in the psychotic literature I think you see in all of the literature how to run a spontaneous course even if there was an earlier period where they may have remitted more spontaneously so for us we are looking for opportunities, markers, descriptors of that transition from 1B to 2 and then the extent to which any interventions at that period may prevent that progression and I think that does represent as Pat has shown I think in a quite sophisticated and the simplicity but the beauty of what he and his colleagues have done in an intervention at that point may actually in the psychosis literature at least or in psychotic states delay or perhaps even prevent the further progression of illness I think that's a really important insight for all of us in these areas and whether that holds up across a wider range of phenotypes so we've actually gone about building specific sets of services and started to publish this and most of this work has been done by my partner Elizabeth Scott who takes care of most of these people unlike me who mainly talks about what we do and we work with her colleagues of actually providing clinical environments where we actually see more people we run a more multimodal sets of services involving neuropsychology and general psychology as well as general practitioners it is in its nature more multi-disciplinary than traditional psychiatric clinics some of the key findings out of that particular area however are really worth dwelling on in the affective disorder area there is a constant debate about whether we are intervening too early in people who do not have a problem because they are below some arbitrary threshold for illness what's really been obvious to us is the extent to which the young people we are seeing through these clinics already have well established disability even though they don't necessarily have very definitive symptom profiles at the time that they present so one could do this in classical diagnostic ways and look at the number of days out of roll that they've had in the last month but simply if you cut down here to the issues you see that quite large percentages of these young people even at what you would consider mild symptom levels have already spent long periods in the last month out of functioning correlated with actually not being functional at school or in education or in the older group in employment I think one of the myths that we have to get over at least in our healthcare setting as it currently is that a lot of people who come along with low symptoms don't have much need I think the ones we are seeing and we are taking in the notion of a well developed disorder at a social level they are already displaying a high level of disability we've taken the staging model and also started to use it in those with anxiety disorders the really interesting thing with anxiety disorders the extent to which they have disability when they are already at these areas you see 33% of these people with anxiety disorders being unemployed and in our staging models they develop depression and substance abuse they move into more permanent types of states we've been trying to look at this staging framework then in following people up in the clinical samples that we've been looking at and split them up into various groups this is data from the first 209 of these sets of young people what you see is of course those at earlier stages tend to have earlier ages of onset we tend to pick them up earlier but you can see we are picking them up right across this particular set of age ranges and the extent to which as they move across ranges they are across stages their actual functioning deteriorates is rated by surface type scales for the whole sample you see the interesting sets of neuropsychological impairments particularly with regards to memory as you move across now the key thing is that we've been following these people longitudinally and as you remember earlier on I was talking about for us the key we think, we hypothesize that the key transition is from stage 1B to stage 2 then over a two year and a half year period that we are following up this particular cohort 20 out of this particular 100 actually transition from 1B to 2 to about 20% transition rate most of it within the first 12 months of seeing those young people so you can get these cohorts of young people and actually follow them and without deciding a prior that they are bipolar disorder or they are psychosis or they are just depression you can actually follow them together and see whether they meet criteria for a set of transitions and they have been the focus of additional study for us most importantly it's been clear to us those who are already in stage to what their situation actually is and then we've been particularly interested in those could we see those who would be likely to transition on the basis of their earlier neuropsychological profile so this is of the 20 who actually did actually transition and while they were earlier on had better neuropsychological profiles to start with that group that transitioned actually already had a degree of memory deficit we've tried to use a number of other ways to carve up the data in terms of what might actually be happening so we try to do this by clinical clustering studies to see whether there are actually suggestions of different trajectories quite independent of the diagnosis that people might actually have and having generated those clusters empirically based on the clinical data the extent to which they may be associated with other validating characteristics basically when you do that if you just cluster on the basis of symptoms and socio-occupational function you tend to end up with three particular groups that don't really respect much with regards to formal diagnosis you end up with basically a functionally impaired group with moderate symptoms and a functionally impaired group who are highly symptomatic these are particularly interesting group, the first group who are not that symptomatic but they're actually quite functionally impaired when you see them there's a high correlation in that group actually with being boys and with earlier neurocognitive difficulties and the suggestion that they may well have quite different neurodevelopmental trajectories from the others you have a middle group who actually demonstrate only limited functional impairment and seem to have more of a genuine adolescent onset and not be terribly symptomatic so there's differences in the particular groups that you see that group that's more well if you like tends to have a higher rate of being female or somewhat higher rate of being female on an ongoing basis there's different cognitive groups then we start to look at those particular groups in terms of their cognitive function their important differences there's also groups that actually either premorbidity or relation to illness actually have quite significant memory difficulties can be quite easily differentiated from those who do not appear to have those difficulties so you end up with two groups that look very similar although one group appears to be highly symptomatic the other group not so symptomatic but quite neurocognitively impaired from the other group which is symptomatic but not as impaired neuropsychologically a different way of doing that is simply to throw away all the clinical data and cluster just on the basis of the neuropsychological data in these people who present with particular symptoms and he would suggest that if you do that you end up with three groups one characterised by poor memory one by poor mental flexibility and a third with impaired attention and memory so there's a different way of clustering just on the neuropsychological data alone again what you end up with interestingly are groups that are quite impaired with their particular sets of memory function not dissimilar to what Chris Pantelis has shown with groups who are at risk of psychosis on an ongoing basis there are groups who enter those populations who already have well-established neuropsychological difficulties before the onset of their illness who are themselves intrinsically different from those who do not have those changes prior to onset now another way of looking at data just for those who prefer more traditional pros to effective disorder we've been interested in the extent to which young people with unipolar versus bipolar depression are cognitively different or not I'm quite impressed with the literature that suggests that a significant proportion of people with bipolar disorder have neurodevelopmental difficulties so we would have guessed that the unipolar group would be less likely to have those same sorts of memory disturbances in these sorts of studies in fact what we found was the contrary that the unipolar and bipolar groups at these young ages had very similar degrees of cognitive impairment in similar sorts of memory profiles at least in the groups that we are seeing and it's important to say within those depression groups that we are seeing many of those young people run very poor paths and it may well be that a significant proportion of those also end up developing psychosis in the long run one of the things about depression in this age group is that it indicates severity but it doesn't necessarily indicate where you're going to go in the classical diagnostic sense so you see depression as a precursor to psychosis you see it as a complication of anxiety you see it obviously in the bipolar people before you see necessarily the classical bipolar picture later on but it's interesting in the groups that we are seeing in this age that actually the cognitive sets of characteristics don't help us to differentiate unipolar versus bipolar sets of young people Daniel Hermans has been really leading our work in examining mismatch negativity in these particular sets of groups of patients as well and the extent to which in classical first onset patients as you might expect there is reduction in their MMM but interestingly and just accepted for publication last week what's really interesting in Daniel and his PhD students Reena's work has been, these are actually those with schizophrenia and those with affective psychosis that those with schizophrenia and affective psychosis in terms of their reduced MMMs are almost identical compared with controls very little difference in fact in that particular signal which has been so associated with schizophrenia in a lot of the recent neurophysiological literature that in fact in our younger people in these sets of situations shows a very similar pattern of impairment in those with affective psychosis Daniel's taken this work further in terms of the MRS work with Jim Legopoulos in terms of looking at good and motorgic potential mechanisms and just recently have been looking at the correlations between the degree of reduced mismatch negativity and the extent to which they find glutamate signals I don't know how well this actually projects but in terms of there's a strong correlation between glutamate expression through MRS and MMMs and mismatch negativity and the important thing about this particular correlation this is in the early stages of illness it's the reverse of what you see and has been reported in chronic schizophrenia so the actual correlation runs in a completely reverse direction in later stages of illness now this is kind of a, I suggest that they might like to stop here it's not often that a correlation in biological psychiatry lines up, I'm sure we can click a lot more subjects, something will happen to the correlation but it's an intriguing idea that a relationship between something like glutamate in this particular issue as an potentially cytotoxic element and a change in mismatch negativity which we can closely link to that the relationship should be quite different early in the course of illness as distinct from later in the course of illness suggesting that the pathophysiology may well change there may be a period or a strategy that might be quite different as distinct from later in the illness in terms of altering pathophysiologies that leads off in a whole other direction which I want to elaborate on here but others have elaborated on about whether in fact there may be quite different approaches to treating patients with mood disorders based on those good hematurgic ideas of greater relevance to the work I've been directly involved with myself is really related to melatonin based therapies and this is a review we have come out and launched shortly because there are new therapies new melatonin based therapies is a chance to look at those systems in greater details in people with young people with depression so we have the chance to look at circadian cycles not just in terms of when you sleep but when you wake but what's the relationship between melatonin and cortisol and body temperature and the extent to which that has classically changed advanced or delayed or becomes asynchronous actually becomes totally disrupted in patients with depression we utilise continuously in these sets of issues not just neurocognitive measures and self-report measures but dim light, melatonin on set getting young people into the labs to look at where their sleep cycles actually are and where their circadian cycle is as measured by their melatonin as well as sleep studies in a smaller proportion Actigraphy is really interesting it's a way of recording sleep wake cycles on an activity over longer periods this is what we've been doing again in relation to stages of illness and looking at the changes particularly important things here are the extent to which in fact with later stages of illness you can see that these younger people are going to bed later and rising later continuously and sleeping for longer in later stages of illness interestingly those who are doing that actually have better cognitive function it looks like a compensatory mechanism that in fact part of the problem as a shift is you actually need to sleep longer in fact to have better cognitive function and quite consistent with those who are working up earlier and have shorter sleep have poorer cognitive function the interesting thing in our particular work though is the extent to which that delayed sleep and the shift which is really phase delay in their circadian cycle is associated with later stages of illness so unlike in the community where we see shorter periods of illness in our adolescence those who are still functional have to get up go to school, go to work, sleeping less and having psychological distress by the time you get into our clinics it looks as if and you get more severely unwell and later and sleeping longer in terms of the dim light melatonin assays this is just to show you the degree of variability one gets in those assays and I think the intrinsic heterogeneity of the samples that we are looking at which is not much help by applying diagnostic labels to that but what we have been doing is splitting these into those who are basically phase delayed or not by actigraphy which is in the top graph or by dim light melatonin onset which is in the bottom graph we are looking into things like early onset versus late onset and further examination of their circadian cycle as well as collecting sleep data which I will skip over except to say the sleep data itself shows nice correlations with neurocognitive function in the young people that we have been studying you see many of the issues that have been discussed here earlier in the day but just in terms of transition what has been really obvious to us is the extent to which circadian disturbance has been associated with transition to bipolar disorder at 11 out of the 18 that we have seen transition from what you would say was unipolar to bipolar they have had significant circadian disturbance during the adolescent period just to say our approach to these things is then to try and plan studies which we are doing now in terms of novel interventions behavioural novel interventions which include behaviour therapies or novel agents like aga melatonin or medaffinil by actually splitting groups a priori according to one of these particular markers in this case phase delay a lot of our clinical trials need to change in the future to actually make a best guess about what sets of stratifications may be meaningful in terms of underlying biological systems for us the circadian system and the shift in phase delay may well be one of those just to say I showed you data from the original 300 sample we now have over 1200 people in our new health services cohort and importantly unlike most other health systems we have been able to recruit half the sample as young men to study through this particular period which takes real effort in fact to get young men into these particular groups and they are as disabled as the groups we've been looking at previously so the nature of our work has really been to use longitudinal twin studies to give us a much better idea of the multiple transitions that are actually occurring during the adolescent onset period and the extent to which they provide a basis for our predictions in clinical samples in our clinical samples to actually look at these cohorts of young persons presenting early to find track their underlying changes without any prior decision as to which diagnostic group or classical psychiatric diagnostic group they may belong to and it's our belief that all these courses and trajectories tend towards those common factors and that they in the end of the day may be much more useful than the traditional diagnostic factors it is important and we see increasingly the role of premorbid cognitive changes starting to characterize at least one of those groups that is at high risk and seems to behave differently to those that are genuinely of adolescent onset and the extent to which we can see at least in early phases of this sort of data the extent to which there is active physiological changes in adolescents may be marked by circadian changes by good and meturgic changes and importantly that both of those systems may be amenable to specific treatments thanks