 This webinar will be presented by Dr. Mohamed Nazma Saqib, and we will also be hosting his supervisor, Dr. Peter Hall. Dr. Saqib is a postdoctoral research fellow in the School of Public Health Sciences at the University of Waterloo. He's also a foreign-trained physician from Bangladesh. His research interests involve bi-directional relationships between adiposity and brain health, as well as neuromodulation technologies. And Dr. Hall is a professor in the School of Public Health Sciences at Waterloo with a research emphasis on brain health, translational neuroscience, and behavioral aspects of disease prevention. So now I will pass it along to our presenters and specifically Dr. Saqib to start. Okay, thank you. Thank you everyone for joining in today and good afternoon. Today I'm going to present my research on the bi-directional associations between adiposity and cognitive function. At first I'd like to thank for this opportunity to present my research on the CLSA webinar series. So as I mentioned, I'll show you some of our recent findings from the CLSA and ABCD data analysis. So there are a lot of topics to cover. So without further ado, let's get into it. So this is the agenda for today. So at first I will give you some background information from the literature. Then I will discuss study one and two where we examined bi-directional associations between adiposity and cognitive function using CLSA data sets. So study one is the cross-sectional analysis of the CLSA baseline data set, and study two is the prospective analysis of the CLSA baseline and first follow-up data set. In the general discussion section, I'll show you some major findings from study one and two, and as well as I'll discuss some strength and limitations. And finally, at the end, I'll show you some of our recent findings from the ABCD data analysis where we examined bi-directional association among adolescent population. Okay, let's move on to background discussion. So let's talk about obesity and its connection with cognitive function for a minute. So the term adiposity indicates the degree of modified accumulation, whereas obesity indicates excessive pet accumulation that presents a risk to health. Obesity is a public health concern because it increases the risk of other chronic diseases as well as death. In literature, it was reported that obesity increases the risk of some other chronic diseases. For example, metabolic syndrome, tattoo diabetes malattice, cardiovascular disease, even it can cause premature death. Besides contributing to the development of these chronic diseases, obesity can adversely affect cognitive function. It was reported that obese individuals tend to have lower executive function, poor memory, they tend to perform poor on neurocognitive tasks, and long-standing obesity can also increase the risk of dementia and Alzheimer's disease in life. So how can we measure adiposity? So there are a number of different accepted measures of adiposity, and each measure has its own advantages and limitations. Body mass index, it's the measure of weight adjusted for height, and it is calculated as weight in kilogram divided by height in the meter square. One limitation of body mass index is that it does not account for muscle mass in an individual. Therefore, it tends to overestimate adiposity in individuals who have higher muscle mass, for example, athletes. Also, it is not a very good measure of central adiposity. So with circumference, it's a good measure of central adiposity, but one problem with this measure is that there are a number of different measurement sites reported in the literature. I think eight different measurement sites are reported, and which could be problematic to compare central adiposity across studies. Next, waist to hip ratio, this is also a proxy measure of central adiposity. And again, as because it's a ratio of waist circumference and circumference, so this also can be influenced by different measurement sites of waist circumference. And dual energy extracts of geometry or the exchange short, it is often considered as a gold standard to measure fat mass. One problem is that it's it requires expensive instrument and time consuming. Therefore, it may not be suitable to use routinely in research and clinical settings. Okay, now I'd like to discuss two important concepts related to the bi-directionality. One is called brain is outcome perspective, and the other one is brain is predictor perspective. So let's look at the brain as outcome perspective first. So in brain as outcome perspective, brain is considered as outcome or end result or consequence of adversarial condition. For example, if an individual have adversarial condition like uncontrolled diabetes or hypertension, it could lead to cognitive impairment in future. So this is an example of brain as outcome perspective. So in the context of today's discussion, we can think brain as outcome perspective as an idea that baseline obesity predicts future cognitive impairments. A number of longitudinal and cross sectional study actually showed evidence for the existence of this brain as outcome perspective. And in medical literature, this view actually is predominant and brain health is almost always considered as an outcome of adversarial condition. As I mentioned, previous studies showed existence for short evidence for the existence of this brain as outcome perspective. And those studies reported that this association is predominantly observed in the domain of executive function, attention and memory. So executive function has several domain, for example, individual control, working memory, cognitive flexibility, and each of these domain can be affected by the adverse effect of excess adiposity. Previous studies also reported that if individuals have obesity in midlife, it can substantially increase the risk of dementia and Alzheimer's disease in late life. So it actually further strengthen the evidence for the existence of brain as outcome perspective. So as I mentioned, in literature, there are abundant evidence for the existence of this brain as outcome perspective. And meta-analysis of those studies also revealed similar finding. So one meta-analysis by young and colleague, these studies, these meta-analysis reported that always participants showed broad impairment in all domain of executive function, including innovation, cognitive flexibility, and working memory, as well as they showed impairment in some other cognitive domains, for example, verbal fluency, decision-making, and planning. So in that meta-analysis, you can see, obvious individual, they showed lower control in the inhibitory control domain of the executive function compared to their normal weight counterpart. Okay, that's all about brain as outcome perspective. Now let's move on to brain as predictor perspective. So in brain as predictor perspective, brain is considered as a predictor of adverse health condition. For example, if individuals have cognitive impairment, for example, executive function deficit, it could lead to unhealthy lifestyle. For example, sedentary behavior or unhealthy food choices. And in the long run, it could lead to the development of override and obesity. So this is an example of brain as predictor perspective. So in the context of today's discussion, we can see brain as predictor perspective as an idea that baseline cognitive impairments predict future weight gain. This view is less well explored in previous literature. However, a number of small scale studies and experimental studies showed evidence for the existence of this brain as predictor path. It is believed that a region of the brain called prefrontal cortex, which contain major nodes of executive function plays a significant role. In experimental studies, it was shown that if individuals, if dorsolateral prefrontal cortex that contain executive function is suppressed by inhibitory brain stimulation, it can result in increased consumption of calorie dense food. You can see in the figure. So this is the dorsolateral prefrontal cortex that contain major nodes of executive function. So experimentally suppressing this region of the brain using suppressive brain stimulation can result in increased consumption of the calorie dense food. So this basically suggests that if individuals have impact executive function, it could lead to the development of obesity by facilitating unhealthy dietary choices. Epithemological studies also showed similar finding. If individuals have lower executive function at baseline, they tend to present with higher adiposity at follow up. Okay, now I'd like to show you two study from our lab that demonstrated the existence of brain as predictor path. So this study by Lowy and colleagues published from our prevention neuroscience laboratory in 2018 in neuroimage. So in that study, 28 female right handed individual were recruited in the study and they received both stimulatory and inhibitory brains and they received both active and sham CTBS brain stimulation. So for your information, CTBS or continuous theta bus stimulation is a suppressive variant of brain stimulation that can temporarily reduce the cortical excitability. So in that study, it was shown that individuals after active CTBS brain stimulation that temporarily suppress executive function results in increased consumption of appetite food, but there is no changes on the control non-apetitive food. So this study actually suggests that CTBS induced attenuation of the left DLPFC in this apetitive snack food consumption. So this is another study from our lab by Sapathy and Hall. So this study was published in brain stimulation in 2019. So in that study, the joint effects of contextual cues and CTBS on eating while examined. So for that study, 107 participants were recruited and they were randomized into four different study conditions. So conditions are sham stimulation and inhibiting cues, sham stimulation and facilitating cues, active stimulation and inhibiting cues, and active stimulation and facilitating cues. And during session, participants received CTBS brain stimulation and after the brain stimulation, they had the opportunity to consume food in a test test. And this study showed that CTBS resulted in increased food consumption in the presence of facilitating cues, but not in the presence of inhibiting cues. So in brief, this study actually suggested that, suggested that stronger CTBS effects in the presence of facilitating cues. So to summarize, both of these two studies suggested that if individuals have a lower executive function, it may result in uncontrolled eating or consumption of calorie dense food, which could lead to a high-diposity in the long run. This is the last slide on brain aspect perspective. I just want to show you the findings from a meta-analysis here. So in that meta-analysis, it examined whether baseline executive function in children and adolescent can predict follow-up weight status. So this meta-analysis actually showed that if children have high executive function at baseline, they tend to present with lower adibosity at follow-up. Okay, now let's discuss some limitation of the previous research. So there is a lack of research using large-scale population-based datasets. So if you consider brain aspect or perspective, it is actually more true. So most of the brain aspect studies are small-scale experimental studies. So ideally, bi-directionally should be explored using large-scale population-based dataset. So previous studies also had a number of limitations. Most previous studies are unidirectional analysis on different samples. So basically, they examined either brain aspect or path or brain as outcome path on different sample. Therefore, it is not clear whether these two paths can happen simultaneously in an individual. And another limitation is lack of temporality. So a number of previous studies are cross-sectional in nature. So therefore, we cannot actually comment on directionality of the association. Many previous studies have sample size issues. So because of small sample size, those studies were not sufficiently powered to detect the small to medium-sized effects. Previous studies also had lack of measures on focal variables. So previous studies, they only examined very few cognitive domains or very few adibosity measures. Previous studies also had limitation in terms of potential compounders. So it is actually not clear whether the association between adibosity and cognitive function is independent or just because of adibosity-related comorbidities. Previous studies also did not explore mediation path in detail. And there is positive of research on this topic in the Canadian context. So it is actually not clear whether the association reported in earlier studies can also be generalized for Canadian population. Okay, based on the finding from the literature, this is our conceptual diagram and the bidirectional association model that we tried to explore in our series of studies on bidirectional association. So we actually anticipated that measures of cognitive function, for example, executive function, verbal memory and reaction time is bidirectionally associated with the indicators of adibosity such as BMI, Wester conference, Wester to HIP ratio and DXA. And we also anticipated that this association be mediated through health status mediators such as hypertension and diabetes and lifestyle mediators such as diet and exercise. So specifically for the brain as outcome perspective, we anticipated that this association be mediated through health status mediators. So adibosity, it would lead to the development of other health conditions such as hypertension and diabetes, which in turn can cause cognitive dysfunction. And for the brain as breakthrough perspective, so cognitive dysfunction can lead to unhealthy lifestyle behavior such as improper diet, sedentary lifestyle, which could ultimately contribute to the development of adibosity. And finally, we also expect that this association would be moderated by socioeconomic status because individuals from high socioeconomic status, they have more flexibility and resources on their dietary option and lifestyle behavior. Okay, now I'm going to discuss our study one. So this study is the cross-sectional analysis of the baseline CLSA dataset. So this study has been published in psychosomatic medicine. So let's talk about the objective and hypothesis of this study. So in the study, our objective was to examine the association between adibosity and cognitive function and test the potential mediation path. So we had several hypotheses. So our hypothesis is one that better score on test of cognitive function would be associated with lower adibosity. So our hypothesis two was that the bidirectional association, the aforementioned association are mediated through lifestyle behavior and medical condition. And finally, we hypothesize that the aforementioned association would be stronger for those of higher socioeconomic status. Okay, now I'd like to give you some information about the CLSA data collection. I think most of you already know about that. So CLSA is a long-term longitudinal study that started data collection in 2011. And baseline data collection was completed in 2015. And first follow-up data collection was completed in 2018. So CLSA has two cohort, tracking cohort, which consists of over 21,000 participants and CLSA comprehensive cohort, which consists of over 30,000 participants. So at the time of our analysis, we only had baseline and first follow-up datasets were available. And for our analysis, we only used CLSA comprehensive for our analysis. Okay, these are the study variables for study one. So our outcome variables are the adiabosity indicators. So these are body mass index, total fat mass measured by TXA, waste circumference, and waste heat ratio. So our cognitive explanatory variables were our cognitive indicators. So we selected stoop task, choice reaction time task, and animal fluency task to represent our cognitive function. So we adjusted our analysis for major socio-demographic variables, such as age, sex, ethnicity, income, education, residence, physical activity. So for the comorbidity, we created somatic comorbidity index, and neurologic comorbidity index, and use those index to adjust our analysis for comorbidity. So for mediators, we selected health condition like type 2 diabetes malatas and hypertension as mediators. And for lifestyle mediator, we selected physical activity, healthy and unhealthy diet as our mediators. Okay, now I'd like to give you some information about the cognitive measures we used in our analysis. So first is stoop task. So stoop task measure executive function. So in stoop task, there is two block congruent and incongruent blocks. So in the congruent block, participants are provided with color name, where color name and font name matches. For example, red is written in red ink. However, in the incongruent block, there is a mismatch between the color name and font name. For example, red is written in blue ink. So when there is a mismatch like that, the time required to identify color increase compared to the congruent condition. And stoop interference is calculated by taking the difference in completion time between incongruent block and congruent block. And higher stoop interference indicates lower executive function. Next is reaction time task. So this task measured the speed of visual information processing. So in that task, participants sit in front of a computer screen and then it took place a keyboard key in response to a visual stimuli on the screen. And higher reaction time indicates lower processing speed. So next is animal fluency task. In that task, participants are asked to name as many animals as they can in one minute. And individuals toward those who are cognitively sound tend to produce more awards compared to those who are not cognitively sound. And higher animal fluency task score indicates better verbal fluency. Okay, now let's move on to the statistical analysis. So for the study one, at first we conducted hierarchical multivariable linear regression. And we assess two models. So model one was the association between adiposity and the covariates. And model two was the association between adiposity and cognitive function while controlling for the effects of covariate. So for the mediation analysis, we used lifestyle variable, such as physical activity and diet and medical condition variables such as type two diabetes malignments and hypertension as potential mediators. And finally, we conducted moderation analysis by income groups. Okay, let's move on to the results for study one. So our analysis showed that hierarchical function was associated with lower adiposity by most matrices. And we also observed some significant mediation effects. For example, for decision between stroke interference and adiposity health conditions such as type two diabetes malignments and hypertension immersed as a significant mediator. Next, we also observed a significant moderation effects. So individuals from high socioeconomic status, individuals from high socioeconomic status showed stronger effect compared to those from low socioeconomic status. Okay, that's all about study one. So study one showed that cognitive function is associated with adiposity indices at the baseline level of the CLSA data. So our next step was to examine whether this association is bi-directional or not using CLSA baseline and first follow-up data set. So this study has already been published in Journal of Gerontology Medical Sciences. Okay, now let's move on to the objective and hypothesis for study one. So for this study, our objective was to examine the bidirectional association between cognitive function and adiposity and examine the potential mediation path. So we hypothesized that cognitive function was bidirectionally associated with adiposity such that higher baseline cognitive function would be associated with lower follow-up adiposity and vice versa. We also hypothesized that for the brain as outcome perspective, this association would be mediated through health status mediator such as blood pressure and diabetes and for the brain as breakthrough perspective, this association would be mediated through lifestyle mediators such as diet and exercise. We also hypothesized that middle-est individual would have stronger effect compared to those of older adults because older adults people are exposed to longer years of adverse effects of adiposity and other comorbidities. Okay, let's move on to methods. So as adiposity indicators for this study, we used body mass index and waist circumference. So our cognitive indicators were stoop task, choice reaction time task and animal fluency task. So we had same set of covariates, sex, ethnicity, income, education, residence, physical activity, comorbidity, and sleep duration. For mediators, we used type 2 diabetes, blood pressure, physical activity, healthy and unhealthy diets. So in that study, we applied some exclusion criteria. So basically those individuals whose first language was not English or French, so individuals who could not work without assistance or individuals who have cognitive, who have brain disorder that may impact cognitive functions such as dementia, stroke, Parkinson's disease, and multiple viruses were excluded from the baseline analysis, were excluded at baseline. So after applying this exclusion criteria, we had approximately 26,000 participants at baseline. So few participants were lost to follow-up. So at WEF2, we had little over 24,000 participants available for the analysis. Okay, now let's move on to the statistical analysis for study 2. So at first, we conducted multivariate, multivariable regression, and then we conducted cross-legged panel model analysis with latent variable modeling. So this is the figure of cross-legged panel model. So we observed that adiposity indicators, BMI and WEST circumference are strongly correlated. So we used these two as a latent adiposity construct. However, cognitive measures are poorly correlated. So basically, different cognitive measures are representing different cognitive domains. So therefore, we used cognitive measures separately in the analysis instead of making a latent construct. So for meditation analysis, we used lifestyle variable and medical condition variable, and we stratified all our analysis by age group, so two broad age group, middle-est and older adults. Okay, now let's move on to the results. Let's first look at the findings of multivariate, multivariable regression. For the brain-est outcome perspective, in middle-est adults, we observed that higher adiposity was associated with higher stroke interference, that means lower executive function and lower animal fluency task performance, that is lower verbal fluency. So finding is similar for older adults, except that higher adiposity was associated with better verbal fluency. So it actually indicates that in older adults, adiposity provides some protective effects on verbal fluency. For brain-est predictor perspective, we observed that higher stroke interference was associated with higher adiposity, and higher animal fluency task performance was associated with lower adiposity. So our findings are in the expected direction. But in older adults, we did not observe any significant effect for brain-est predictor perspective. Similar finding was emerged with cross-legged panel model analysis. So in that figure, dotted line indicates middle-est, whereas solid line indicates older adults. So for the brain-est predictor perspective, animal fluency was immersed as a significant outcome in older adults for adiposity. But in a direction that suggests adiposity provides some protection on verbal fluency, stroke interference was immersed as a significant outcome in middle-est adults. For brain-est predictor perspective, only stroke interference was immersed as a significant predictor of adiposity, but only in middle-est adults, not in older adults. So we observed bidirectional association here with stroke interference, that measured executive function, and only in middle-est, not in older adults. So next, move on to the mediation analysis. So for the brain-est outcome perspective, systolic and diastolic blood pressure and diabetes mellitus immersed as a significant mediator. And for brain-est predictor perspective, diet and pastries immersed as a significant predictor. Okay, this slide is just to summarize the cross-legged panel model finding. So for stroke interference, okay, in that figure, solid line indicates statistically significant path, and dotted line indicates non-significant path. So path B is the brain-est outcome path, and path C is the brain-est predictor path. So for stroke interference, both path B and both path C immersed as a significant in middle-est adults. So we observed bidirectional association here involving stroke interference and only in middle-est. For animal fluency, in older adults, only the brain-est outcome perspective immersed significant, but in a direction that suggests adiposity provides some protection on verbal fluency. And for mean reaction time, we did not observe any significant effect. Okay, now let's move on to the general discussion. So I'll summarize the major findings from our study one and two. So in study one, we observed that beta cognitive function is associated with lower adiposity by most matrices. For study two, we observed that in middle-est adults, adiposity is bidirectionally associated with executive function that is measured by stroke task. In older adults, only a brain-est outcome perspective immersed significant, such that higher adiposity was associated with better verbal fluency in older adults. For mediation analysis, we observed significant mediation effect for lifestyle variable and medical condition variable. Okay, now I'd like to show you some strength and limitations of analysis. So let's discuss the strength first. So we use large-scale population-based data sets in both of our study one and two, unlike many previous studies. Those are like small-scale studies with a small number of participants and not sufficiently powered to detect small to medium-sized effects. And our study two is a longitudinal analysis. So we can actually comment about directionality here. So what is the effect of baseline cognitive function on follow-up adiposity and vice versa? So in our analysis, we used multiple indices of adiposity and cognitive function. So therefore, we can actually comment that this association is not limited to any specific measures of adiposity or cognitive function. And we have, we had rich information on potential compounders, particularly the information on comorbidity. So we created comorbidity index and adjusted our analysis for those comorbidities. Okay, now let's move on to the limitations. So one major limitation is that prospective analysis were limited to only two waves of CLSA data. So as I mentioned, at the time of our analysis, we only had two waves of CLSA data and two waves of data may not be sufficient to properly quantify the effect size. And another problem was that shorter follow-up interval. So between baseline and first follow-up dataset, there was just three years of gap. And it may not be sufficient to properly quantify the cross-lake effects because we are expecting a gradual accumulative phenomena here. So the effects of adiposity on cognitive function and vice versa should be accumulated over time. So if we have a longer follow-up interval and more follow-up data, then we could actually would see more stronger effects compared to what we get here. And although we have very large sample size, population representative may not be perfect because we used only CLSA comprehensive cohort. And as because CLSA comprehensive cohort participants were recruited from the major urban center, it may not be entirely representative for all Canadians. And finally, cognitive domains examined within CLSA were not exhaustive. So for cognitive function, we use three domains. We assess three domains. And for prospective analysis, we had like two adiposity measures. So future studies should actually examine other cognitive domains using a comprehensive set of adiposity indices. Okay, so our study to actually show that there is bidirectionalization between adiposity and executive function in middle list errors. So our next step was to examine whether this association is also true for adolescent population or not. So we actually conducted this bidirectionalization analysis in a sample of adolescent population using ABCD dataset. So I'll be showing you now some of our recent findings from the ABCD data analysis. So this study is conditionally accepted in JAMA Open and currently in review now. So at first, I'd like to give you some information about the ABCD study. So ABCD is a multi site longitudinal study in the United States. And this study recruited approximately 12,000 children ages 9 to 10 year. And this study began data collection in 2018. And at the time of our data analysis, we have web one, two, three data sets wide available. Okay, now discuss the variable and findings of our ABCD analysis. So cognitive indicators of this study are the cognitive tasks included in the NIH toolbox cognitive battery. So we actually included five cognitive tasks in our analysis. So these are flanker tasks that measure executive function, pattern matching tasks that measure the speed of visual information processing, picture sequence task that measure episodic memory and picture vocabulary task and oral reading task. Adiposity indicators of this study were BMIZ score and WESTA conference. And we adjusted our analysis for major socio demographic variables. So for the results, so we conducted both multivariate, multivariable regression and cross like panel model analysis just like our study too. For brain as outcome perspective, we observed that higher BMIZ score was associated with lower picture sequence task performance that measure episodic memory. And higher WESTA conference was associated with better picture vocabulary task performance. For brain as breakthrough perspective, higher flanker task and picture sequence task performance was associated with lower adiposity. Similar finding was immersed with cross like panel model analysis. So for brain as outcome perspective, only pattern matching task and flanker task emerged as a significant outcome for the adiposity. For brain as breakthrough perspective, except the pattern matching task, rest of the cognitive tasks emerged as a significant predictor for adiposity. So we basically observed bi-directional association here using flanker task that measure executive function. Okay, this slide is just to summarize the cross-lect panel model finding for our ADCD data analysis. So for the flanker task and pattern matching task, brain as outcome perspective emerged as a significant and except pattern matching task, rest of the task emerged as a significant predictor for the brain as breakthrough perspective. So we observed bi-directional association here using flanker task but not with other task. So in the ADCD data sets, we have brain morphology data measured by structural MRI. So we used brain morphology as a mediator in our analysis and we observed few significant mediation effects here. So for the brain as outcome perspective, several brain morphology emerged as a significant mediator, for example, middle frontal gyrus thickness, inferior frontal gyrus thickness and the volume of middle frontal gyrus and the volume of lateral prefrontal cortex. For the brain as predictor perspective, only the thickness of lateral prefrontal cortex emerges as a significant predictor. Okay, so I'd like to wrap up by showing some implications and future directions here. So one of our limitations was that we had only two waves of data and only three years of follow-up interval. So feature studies should employ longer prospective data collection interval to more conclusively characterize the magnitude of cross-lect defects, ideally 10 years or more. So if we have like multiple waves of follow-up data and sufficiently enough follow-up interval, then we would see more stronger effect because we are expecting a accumulative nature here. The next is explore the extent of biorectionality using other cognitive domains. So in our analysis, we were able to include few cognitive domains, but future studies should include other cognitive domains. For example, memory attention and we should see whether those have any biorectionality or not. So for our prospective analysis and also for our abcd data, we were limited to use only DMI and Wester conference. So future studies should use a comprehensive set of adiposity measures, particularly DxA, which is considered as a goal standard for FETMAS assessment. So in our abcd analysis, we used brain morphology as a mediator only. So future studies should also explore the biorectionality using brain morphology features and functional new imaging. So in our both abcd and CLS analysis, we observed a biorectionalization using executive function. So we should actually explore the utility of executive function training in eating and weight loss interaction, particularly for those who have undergone bariatric surgery and explore the novel method of enhancing executive function using TMS. So TMS is a non-invasive brain stimulation technology that can be used to enhance executive function. So future studies should also explore this kind of novel method. Okay, that concludes my presentation. So I would like to take this opportunity to express my gratitude to my supervisor, Dr. Peter Hall for his constant support and guidance throughout my doctoral studies and also supporting me now as a postdoctoral researcher in his lab. I'm very thankful to my collaborators, Dr. Reza Ramazan and Dr. John Best and Dr. Mary Thompson for extending their expertise in statistical analysis. So I'm also thankful to all the participants and researchers associated with the CLS and abcd data sets. So for making these amazing resources for the researcher. And I'm also thankful to all the current and previous member of the prevention neuroscience laboratory. So it's really great to work with such a great group of people. Okay, thank you everyone for your attention today and listen patiently. So we'd be happy to answer any questions you have. Thank you. Okay, well, thank you very much, Dr. Sakeem and Dr. Hall. I don't know if you, Dr. Hall, did you have anything you wanted to add before we move into questions or are you just the silent partner today? Mostly the silent partner, unless there are any specific questions that I can have my perspective. Well, I'll just remind everyone to use the Q&A box down at the bottom of the screen to post any questions. I just want to say I think this is a great example of use of longitudinal data and platform like the CLSA as we get more follow up data. I think, you know, having that 10 year minimum span to be able to look at issues like this and the aging process is, you know, that's just the power of it. So I think that's just a great example of this. So we do have one question from Bazma Ahmed. Great talk. Did you consider looking into depression, either diagnosis or treatment as a co-variable in your analysis, knowing the relation between depression and cognitive impairment, especially in older adults? So in terms of depression, I think we created Comorbidity Index and I think depression was one of the conditions that we included in our Comorbidity Index. So we actually already adjusted our analysis for depression and some, I think many chronic diseases. So in our Comorbidity Index, so we had somebody Comorbidity Index that includes, I think, 22 Comorbidities and Neurologic Comorbidity that includes four Neurologic Comorbidities. So our analysis are already adjusted for some of those Comorbidities. Great. Lots of compliments on your presentation today. So really great job. So Laura Anderson here at McMaster would like to know what did you do with people who are underweight and could there be a U-shaped association? Could you please repeat it again? What did you do with people who are underweight and could there be a U-shaped association? So for our obesity analysis, so now we actually do a sensitivity analysis by removing those who are underweight and we actually did not find any significant difference between these two analysis, like main analysis and analysis after excluding those underweights. So findings are almost similar. So yeah, we'll actually check our analysis with underweight and without underweight. So there is not much difference in the findings. So this is certainly a valid point and something that we have has been on our radar for both sets of analysis. The idea that people who are underweight may have because severe underweight is associated with cognitive performance decrements on some of these tasks in other research, both in older adults and in young adolescents and adolescents. And so in ABCD, as Nazma's pointed out, we did do a sensitivity analysis, removing those who are underweight and did not actually affect the pattern of findings. Honestly, I cannot recall what we did with CLSA specifically. We may have to get back to you. It seems to me we discussed this and tried something to deal with it. I'm not totally sure if that ended up in the reported paper or not, but certainly it's a good point, something that needs to be considered when considering adiposity and cognitive function. Next question is, was there a way to account for acute issues such as sleep deprivation, that they diet stress that may have affected cognitive ability? Also, was BMI a single point or is length of time with BMI considered? It might be easier to read that question in the Q&A. So in terms of sleep deprivation, it's a really good point. So sleep deprivation, it could affect cognitive function as well as it has shown association with adiposity in literature. So in our study one, we did not adjust it for sleep deprivation, but for our study two and our ABCD analysis, we used sleep deprivation as a co-period, and I think we accounted for those in our analysis already. What is the other part of the question? Stress. Stress as a second co-period. I believe we did not. Yeah, we did not actually adjusted our analysis for stress, I believe, yeah. Okay, hopefully that answers the question. There's a question that, just a reminder, to put the question in the Q&A box, but I do see a question from Cynthia in the chat box, so we'll ask that. She's wondering if you looked at any blood-based biomarkers. I don't think I saw that in your presentation, but maybe you, maybe I missed it or maybe you've done other work. So in our dataset, we actually did not have blood biomarkers. So like in future, actually, we'd like to check other mediation path. For example, metabolic syndrome, and we'd like to see the blood biomarker and how it's linked to the adipocytic adipocytic function. So it's something we want to do in future. So we have a follow-up question about the BMI. Was BMI a single point, or was length of time with BMI considered? That was the second part of the question we didn't answer. So I'm not sure if I understood this question properly. So BMI was collected at baseline datasets, I think as a single point collection, and also in the first follow-up dataset, BMI was collected, and we used BMI in our prospective analysis. We used BMI of first follow-up dataset, and BMI of baseline, and BMI of first follow-up dataset. We used this in our analysis as like our focal variable in our adipocytic indicator, as our adipocytic indicator. So I think certainly with more follow-up interval as a greater time lag, total time lag between baseline and final measurement, we'll have an opportunity to more comprehensively address this question, where we can actually approximate how much time someone may have lived with BMI. But we did examine this implicitly and almost explicitly in our discussion of how to interpret the findings of the ABCD study, which involves a nine and 10-year-olds baseline versus CLSA, which is older adults. And we found more evidence of bidirectionality in the older adult sample, sorry, in the middle-aged segment of the CLSA compared to the adolescents who really it's the brain as predictor pathway that's most consistent. And so the way that we discussed those findings was that for adolescents, there may be less evident brain health impacts of adiposity prospectively because they haven't lived with obesity very long. So we may have kids who have lived with it from months to years, not decades or more. Whereas by middle-aged people have potentially been obese for years to decades. And so you'd be more likely to see prospective facts of obesity on the brain when it's been kind of on the scene longer like that. So I think, yeah, we haven't sort of integrated that. I think the data limits our ability to do that in terms of a direct test of that. But in terms of comparing those different age groups, it's certainly how we conceptualize the difference in terms of where you say see the brain health impacts and where you don't in terms of age range. Well, I think that is all the questions. If anybody has any additional follow-up questions that you think of, I'm sure you could, you can definitely email the CLSA and we will follow up with our presenters or I believe the emails of doctors at keyboard and Dr. Hall were included. So I will wrap things up by just thanking you again for your excellent presentation. Again, lots of positive comments about the presentation of your slides and ease of understanding them. So I think you did a great job on that and our participants appreciate that. I'd like to remind everyone that the next data access application deadline is January 18th of 2023. Please visit our CLSA website under Data Access to review available data as well as additional details about the application process. Registration for our January CLSA webinar will be posted on our website at some point early in January. We actually do not have a topic confirmed for that webinar yet, so stay tuned for that. And finally, remember that the CLSA promotes this webinar series using the hashtag CLSAWebinar and we invite you to follow us also on Twitter at atclsa underscore ELCV. And I hope everybody has a restful holiday season, whatever you may be doing over the next few weeks, and we will see you in the new year.