 All right, well, let's get started. It's about 12 o'clock. My name is Jennifer Boyko, and I'm with the Canadian Longitudinal Study on Aging, or the CLSA, as the acronym goes. Thank you for attending the September talk of our 2019 CLSA webinar series. Today, our webinar is entitled, Seeing, Hearing, and Thinking, the Cross-Sexional Relationship between Sensory Status and Cognitive Function in CLSA Participants. Dr. Natalie Phillips and Dr. Paul Mick. Now, I'll get their bios here. Natalie Phillips is a professor in the Department of Psychology at Concordia University. She examines the neuropsychology of healthy aging and Alzheimer disease. Her current research interests focus on sensory cognitive interactions and speech and language processing in older adults, including those who are bilingual. Dr. Phillips is also one of the principal developers of the Montreal Cognitive Assessment, which is a screening instrument for the assessment of mild cognitive impairment. She is one of the lead neuropsychologists for the Compass ND study in the Canadian Consortium of Neurodegeneration and Aging, or CCNA, as some of you may know it by, and is the founding leader of CCNA Team 17, which examines sensory decline and cognitive function in people with dementia. And Dr. Paul Mick is an associate professor in the College of Medicine at the University of Saskatchewan. He is a head and neck surgeon who leads the Saskatchewan Neurodology and Cochlear Implant Program. His research seeks to understand relationships between sensory loss and broader health outcomes, including social isolation, cognitive decline, and dementia. He also leads studies that assess the effects of sensory interventions, such as Cochlear Implants on broader health measures. Finally, he is also a member of Team 17, which Natalie Phillips is also part of, which again focuses on interventions at the sensory cognitive interface. So, I think that is the biographies and I hope you can all still hear me. I will now pass it on to Dr. Phillips, who will start the webinar off. Thank you, everyone. Thank you very much for inviting us to give the webinar today. And I think it's apropos that we all just had the same collective hearing loss. Perhaps that'll give us some insights into the data today. So, we're very pleased to share some of the findings today, and we want to start right off the bat with our acknowledgements. And that is the analyses we're presenting today. We're supported by a CIHR CLSA catalyst grant and a supplement from the Quebec Research Network on Aging. Collectively, our team is supported by an infrastructure and network grant from the CCNA, the Canadian Consortium on Neurodegeneration, as Jennifer mentioned in support of Team 17. And in a broader context, funding for the CLSA, as we know, is provided by CIHR and CFI. So, I've alluded to our team. We're nine members strong across Canada, and we're a group of individuals who are interested in the interface between sensory function, cognitive function, in older adults and in individuals who have dementia. And in this presentation, I particularly want to highlight the contributions of a postdoctoral fellow we've worked with, Annie Hemelain. I also want to acknowledge that there have been previous CLSA webinar series, both from members of our group, Kathy Pakora Fuller and Walter Widdich and, of course, Paul Nick today. And those are available online. And there have been also previous webinars by Holly Tuolco and Ellen Friedman also reviewing issues on cognition and vision. And so these are useful webinars to review as well. So we've known for many decades that there's a relationship between aging and sensory function and aging and cognitive function. And what we see on the graph on the slide here, on the left-hand panel, is an illustration of how changes in visual acuity are associated with age and changes in hearing acuity are associated with age. And in the lower portion of that left-hand panel, we also see that aspects of intellectual functioning decreases across the lifespan as well. Work from the Baltes and Lindeberger group in the 1990s, though, showed that the relationship between age-related cognitive change can actually be shown to be mediated by changes in vision and in hearing. So this literature on sensory loss and cognitive outcomes has really grown over the last number of years. And there's quite a robust literature now showing a relationship between sensory loss and risk for cognitive decline. And we can see in this slide that the risk ratios, if one has a hearing loss, are amplified, the risk ratio for the development of dementia with an average increased risk of dementia of about 30% if one has a moderate hearing loss. There's a similar literature for vision impairment as well. There are a number of possible causal pathways that link sensory loss with cognition. And I'm just going to use the case of hearing as an illustrative example. So one possibility is that we're simply looking at a common cause. Cognition declines with age, hearing declines with age. They're not causally related to each other. They simply reflect the third variable. So that is possibility. A second possibility is that changes in cognition cause or divert attention away from the perceptual task. And this is the cognitive load on perception hypothesis. There's not a lot of strong evidence for this hypothesis. In contrast, a hypothesis, the information degradation hypothesis states that hearing loss causes immediate impaired cognitive performance because you're getting poor quality auditory information into the cognitive information processing system. And a somewhat more elaborate form of this hypothesis is the sensory deprivation hypothesis. And this one states that prolonged hearing loss causes functional and structural changes in the brain over the long time and in a time lag fashion causes cognitive decline. Finally, there's a social mediated hypothesis. And this states that hearing loss has an effect on cognition over the long term through an indirect path. Hearing loss is associated with greater depression, greater social isolation, reduced social participation. And it's these factors that actually relate to poor cognitive outcomes. So at the moment, all of these hypotheses are viable and are awaiting the appropriate longitudinal studies to adjudicate between them. All right. Paul here. So, yeah, so we have taken advantage of this rich data set available through the Canadian longitudinal study on aging. We have in the studies that we'll be describing today are utilizing data from the comprehensive cohort because performance or behavioral based hearing and vision data are available as opposed to the tracking cohort in which only self-report data are there. The comprehensive cohort consists of about 30,000 participants between the ages of 45 and 86 at baseline who live within 25 to 50 kilometers of 11 data collection sites across the country. Now, within the CLSA, individuals who have cognitive impairment were excluded at the time of recruitment. In other words, any person who was unable to understand the consent discussion for the study as determined by the CLSA staff was not invited to participate. Most participants were recruited by telephone, which would have excluded most individuals in the population who are deaf. The data collection has occurred in participants' homes and at these data collection sites, and over 4,000 variables have been recorded, including behavioral measures of hearing and vision. So in the CLSA comprehensive cohort, hearing is determined using semi-automated air conduction audiometry. The machine is shown here. Essentially, participants indicate when they can hear pure tones at different thresholds, and the threshold at which they can only hear 50 percent of the sounds is considered the threshold for that frequency. These different thresholds can be combined to create pure tone averages. In this study, we've used the pure tone average, which is used by the World Health Organization. But we've also looked at whether other averages better reflect self-reported impairment and found some interesting results. You can contact our group if you're interested in knowing more about these analyses. For vision, visual acuity was measured both with habitual correction, in other words, using the contact lenses or glasses that the participant walked into the data collection site with that day. And also with pinhole correction, pinhole correction is performed using an occluder, such as shown here by looking through the hole in the center of the occluder. The light that passes through goes straight through the middle of the ocular lens without being refracted, landing on the focal point of the retina. Thus, the pinhole correction measure estimates visual acuity corrected for refractive error. So for any hearing specialist in the audience, it's somewhat analogous to the bone conduction threshold. So it will be important to have a basic understanding of the scales of the hearing and vision measured when evaluating our regression results later in the talk. In terms of what constitutes an impairment, definitions vary according to different organizations. And of course, the degree of impairment exists on a continuum and is affected by a person's other biological, social and environmental factors. But nevertheless, looking at the World Health Organization definitions, we see that a 20 decibel difference is about equivalent to a difference between mild and moderate or moderate, severe and severe and severe and profound hearing loss. And for vision, a 0.2 log mark difference is equivalent to the difference between a mild and moderate impairment. So we'll be using these intervals later to describe some of our results. We've previously looked at the prevalence of hearing and vision loss in the Canadian population using data extrapolated from the CLSA. Here are the results for hearing. The solid lines indicate the percentage of people at different ages who have a hearing impairment. The top group of lines is mild or worse. The bottom group of lines is moderate or worse. The dashed lines indicate the absolute numbers of Canadians with a hearing impairment at different age groups. We see that hearing loss is more common in males, both proportionally and in terms of absolute numbers, except at the older age groups when there's actually more females with hearing loss because of a survival effect. For vision, in contrast, there are more females than males with vision loss and the increase with age is less exponential than for hearing. People with one sensory loss may rely on the other sense to compensate for their functional impairment. For example, people with hearing loss use vision to lip read and people with vision loss use sound to orient themselves in space. Thus, having a dual sensory loss may compound the impact on daily function in a synergistic fashion. We see that the prevalence of having just a single impairment drops off at ages older than the mid 70s due to an exponential increase in dual sensory loss, which becomes more common than having a single impairment in people in their mid 80s and older. So over to Natalie. Thanks, Paul. Regarding the CLSA cognitive measures, there are a number that the CLSA has included and oh, sorry, just trying to get control of the slides again here. Is that visible? I hope. OK, so we can put them into two general categories of executive function and memory. For the executive function measures, two of the measures are measures of oral fluency. So in this task, you have to generate as many words as you can in 60 seconds, according to some kind of rule, such as words that begin with a particular letter or come from the same semantic category. And this calls on aspects of executive function because it's a relatively unconstrained task that one has to structure oneself and then use structured and flexible search of semantic memory to perform this task. Well, another task is the mental alternation task. And here you have to say out loud numbers and letters in alternating ascending order. So you say one, a, two, B, three, C, and so on. Again, it's caps into executive function because one has to represent those that information and working memory and then flexibly alternate between the stimuli. We also use this strupe measure included in the CLSA. This calls on aspects of executive function. In this task, one has to name the color ink in which words are printed. And in some cases, that is a relatively straightforward task because you have to name the ink color green and the word is green and there is no conflict. But on the more complicated trial, one has to name the color of the ink and say red. However, the word is written in a conflicting, the word that is written conflicts and one has to inhibit the competition associated with that. For the memory test, it is an adapted version of the Ray Auditory verbal learning task. And here a participant is read a list of 15 words. They then have to recall those words immediately and again after a five minute delay. And we use two measures. One was the delayed five minute recall and the other was a ratio of the delayed performance divided by the immediate performance. And this is a measure of how many words you retain in memory. We subjected these measures to principal components analyses to reduce the number of variables. And this in fact generated two principal components, one for executive function and one for memory as we predicted. We then use multiple variable linear regression where dependent variables are the principal components for executive function and memory. Our independent variables of interest were the pure tone threshold averages for hearing, visual acuity for vision and a host of control variables that I'll mention in a minute. So first of all, let's do a reality check. If we look at the left panel, we see the individual test scores. We see that these decline over the age span. So older participants perform somewhat more poorly than younger participants. And on the right side of the panel, we see that that is true for the principal component scores as well. So very much as one would expect age related change in cognition. What else relates to better cognitive performance? Well, in the case of executive function, it's being younger, being female, having higher education, higher income, fewer multi-morbid health conditions, being tested in English, and being multilingual and tested in one's native language, if that's English or French. And for memory, it's quite similar, being younger, female, higher education, higher income and being multilingual and tested in one's native language. There are also social factors that relate to cognition as well, and Paul will be telling us about those in a minute. But turning to our key question, is there a relationship between sensory function and cognition? In our analyses, we control for crucial variables that we know are going to be related to cognitive performance, including age, sex, education, household income, cultural background, language status, whether there are multi-morbid health conditions and for the women, their menopausal status, or whether they were taking hormone replacement therapy. And importantly, we control for the status of the other sensory modality. And by that I mean, when we look at the relationship between hearing and cognition, we're controlling for an individual's vision status and vice versa. So what we found is that poor sensory function is associated with lower cognitive performance. We see the results for executive function on the top panel and for memory on the bottom panel. What we see is that hearing is a significant predictor of executive function and for memory performance. So with increasing pure tone thresholds, we see decreasing cognitive performance. And for vision, we see a similar relationship with executive function. I note that there isn't a relationship between vision and memory performance. These relationships between sensory status and cognitive performance are over and above age and the other factors that I mentioned to you. And there were no sex differences in our analyses. One simple thing we needed to check was whether these relationships were simply a confound of sensory input. So let's consider for a second, what's the primary sensory modality of any one of these tasks? In the case of executive function, we can say that the stroke indeed relies on visual perception. The other aspects of executive function that we measured do not. And for memory indeed, they rely on adequate hearing as well. But we do see that nevertheless, executive function is predicted by vision and hearing. So we have cross modal predictions, not simply due to sensory input. And indeed, we also use ratio scores that control for baseline performance that is influenced by sensory input. So we don't think this is a viable explanation. We also asked whether the relationship between sensory status and cognitive function is moderated by other variables. And in the case of hearing and cognition for the domain of executive function, we see that the negative effect of hearing loss is magnified by older age, lower income and more multimorbid health conditions. However, the relationship is partially mitigated or diminished if one has partial retirement. Similarly for memory, the negative effect of hearing loss is magnified by older age. So here are some hypothetical results or illustrative results on the left side of the screen. And so here we see that indeed there is a negative relationship between cognitive performance on the y-axis and hearing status. And that's true for younger and older participants, but that relationship is exaggerated in the older participants. For vision and cognition, there were no moderating variables to explain this negative association. And I also note that hearing and vision did not interact with each other. So what does any of this really mean? Well, to try to get at that question, what we did is we looked at the beta coefficients from our multivariable linear regressions. And these are essentially estimates of effect size. So here we're taking a change in the dependent variable, either executive function or memory scores, as a one unit change in the independent variable, assuming all the other independent variables are being held constant. And what we did is we expressed the sensory effect on cognitive performance as if it was the equivalent of a cognitive aging effect or the equivalent of the effect that age would have on cognitive performance. And so in the case of executive function, we can say that the 20 dB difference in hearing is equivalent to 4.6 years of cognitive aging and a difference of 0.2 log Mara visual acuity is equivalent to 4.2 years of cognitive aging. And we can make similar statements about memory. So if we take the hypothetical case of Sally and Joan, who are the same age, they're the same sex, they have the same degree of education and health status. However, Joan has a moderate hearing loss. Joan will perform as if she is five years older on a test of executive function compared to Sally. And she'll perform as if she's almost nine years older, if she also has a mild vision loss. All right, so transfer over to me now. And so, yeah, we wanted to, we confirmed that there were associations between hearing and vision loss and executive function and memory. And we wanted to perform an exploratory analysis to see to what degree social factors might mediate these associations. So we see the basic premise of our hypothesis in this figure from the recent Lancet Commission's paper in which a meta-analysis was performed to identify determinants of dementia. Hearing loss was identified as a mid-life risk factor while social isolation was identified as a late-life risk factor. Interestingly, hearing loss was actually calculated to be the potentially modifiable risk factor with the largest influence on dementia incidents at the population level. The population attributable fraction was 9%, meaning that if you could theoretically eliminate hearing loss from the population, the incidence of dementia would drop by 9%. So here is a directional acyclic grass illustrated the proposed causal pathway. Of course, without data at three separate time points, we are quite limited in terms of our ability to draw causal conclusions from the data. We do plan such longitudinal analyses in the future, but with the cross-sectional data such as what we have for the current study, reverse causation is possible. For example, associations between cognitive and social variables could easily be the result of the effective cognitive problems on social interaction rather than the other way around. So thus, just keep in mind, we considered our current analysis to be exploratory in nature rather than confirmatory. We've previously examined links between self-reported sensory measures and various metrics of social structure and function. Hearing loss was associated with lower scores on measures of the quality of social interaction. So for example, loneliness and social support, possibly due to its effects on interpersonal communication. And I'm referring to the table on the right here. In contrast, vision loss was also associated with smaller social networks and lower participation in social activities as well as low social support and loneliness. The additional associations between vision loss and low social participation in smaller social networks may be because vision loss also affects mobility and transportation to a greater degree than hearing loss. The CLSA has a wealth of measures of social structure and function, some of which are shown here on the left. These are the factors which we considered in our current study. So we used the so-called causal steps approach to mediation analysis. So in this approach, you first perform multivariable linear regression models to determine associations between sensory loss and A, executive function and B memory, such as already described by Natalie. In these models, you include all relevant independent variables except for the social variables. And next, we re-ran the same models, but this time including all of the social variables. And we compared differences in effect estimates for hearing and vision between these two models before and after adding the social variables. And the percent difference between the beta coefficients is indicative of the proportion of the sensory cognitive association potentially mediated by social factors. So here are the results for the executive function model. On the left is the model output. On the right is a graph which compares effect estimates for the independent variables in the model. The bar on the very left is the effect estimate for a five-year change in age. And so you can use that sort of to put the other effect estimates in context. We see here as expected that worse vision and worse hearing were fairly strongly associated with worse executive function. And in addition, some of the social, well, many of the social factors were also associated with executive function. Being lonely, having a smaller life space and having less social support were associated with worse executive function. Being able to drive or being partially retired, living alone and wanting more social participation were associated with better executive function. Keep in mind that a number of social variables were put in these models and collinearity might have decreased their observed effect estimates. Future studies might use a composite variable such as the one described by Andrew Wister and his colleagues to determine overall effect of social factors. So here is the results for memory. I'm looking at the graph on the right here. The effect estimate for older age was smaller than before. We see that worse hearing is associated with worse memory, but there wasn't a significant association for vision as Natalie's described already. Having smaller life space and less social support were both associated with worse memory while being able to drive and partial retirement was associated with better memory. So this is sort of the meat of the study here. This figure describes the difference in beta coefficients in executive function models that did not include social variables. In other words, which is shown here in blue, and those models that did include the social variables shown here in orange. For hearing, the difference was negative 6.8%, meaning that about 7% of the overall association between hearing and executive function could be mediated through these social factors. For vision, the difference was negative 8.3%, meaning that about 8% of the overall association between vision and executive function could be mediated through social factors. In the memory models, we saw a smaller difference for hearing, there was only about a 4% difference in beta coefficients, meaning proportionally less of the overall association could be mediated through social factors. And again, vision loss was not associated with poor memory. So finally, we did perform stratified analyses according to sex and according to age groups. The age groups we used were 45 to 64 and 65 to 85. We found that the mediating effect of social factors, in other words, these percent changes in beta coefficients appeared to be slightly higher among females and the older half of the cohort. Okay, so in conclusion, let's put all these results together. We saw that lower levels of sensory functioning is associated with lower cognitive performance. As a reminder, in the case of hearing, it's associated with lower performance on executive function measures and on our measure of episodic memory. And in the case of vision, it's associated with executive function. I think it's very interesting that these relationships are observed in the CLSA dataset. As a reminder, these are middle-aged to older-aged adult Canadians. And so it's notable that we see these relationships even in individuals who are middle-aged. Also, these are participants who have relatively good health and clinically normal sensory abilities. So we were able to see these effects between sensory functioning and cognition over and above the major, the effects of major demographic variables that we know do influence cognition, including age and education and so on. So we saw that there were small but statistically significant associations between executive function and most social variables. For memory, there were associations with smaller life space and less social support. These strong associations between executive function and hearing, executive function and vision and memory and hearing remained even after controlling for social variables. Thus, there was fairly weak evidence for a mediating effect of social variables on the sensory cognitive associations. Okay, some of the limitations of the work to date is one could consider the assessment of hearing and vision somewhat simple in nature in the sense that they are largely oriented towards threshold detection. It of course would be desirable to have other more complex measures, including measures of speech and noise, visual contrast sensitivity, visual field function and so on. And we're happy to note that our colleagues, Kathy PCORA-Fuller and her collaborators have developed a speech and noise task that is going to be available in subsequent waves. Of course, as Paul mentioned earlier, the sample and the recruitment is biased towards generally healthy individuals. And he also had mentioned earlier that individuals who have very severe hearing loss or would be considered deaf would not have been enrolled in the study. So the sensory impairments that we saw at baseline were very mild if there were any at all. And we know that the CLSA sample, although a very remarkable sample is not 100% fully representative of the general population of Canada. And of course, it is notable that we were only able to evaluate cross-sectional relationships at this point in time. So in terms of our future directions, we of course want to re-examine these sensory social and cognitive associations using longitudinal data. And the results of such studies would be more relevant for causal inference. We'd like to identify groups who are particularly at risk for sensory loss associated, cognitive decline. For example, older age groups, lower income groups, people with certain health conditions because we do see that there is a large amount of inter-individual variability in the risk of that sensory loss in the risk of sensory loss on cognitive decline. We'd like to evaluate other factors. For example, genetic factors such as ApoE haplotypes that may be common factors that could explain sensory cognitive associations. In summary, a better understanding of the causes of these sensory cognitive associations will help us determine if sensory interventions should be investigated as strategies to improve cognitive health, help identify vulnerable groups who might be prioritized for further research and or public health interventions and explain this inter-individual variability in risk. Our research here fits into a wider program that aims to integrate sensory health into broader approaches to healthy aging. So lastly, we'd like to acknowledge our collaborators, all the members of CCNA team 17, but in particular, Walter Wittich and Kathy Pakora Fuller who are co-conspirators in this research. Although Paul and I were speaking today, this represents work from all of us and crucially the contributions of Annie Himalane, who is the postdoctoral fellow who worked with us on these analyses and of course the trainees and research staff who supported us. And on the note of trainees, we are looking for new trainees to join our research group and if you know of trainees who are interested in these topics, please encourage them to contact us and we would be happy to grow the team and with that, I'll just say thank you on behalf of Paul and myself to the audience for listening and to the CLSA organizers for inviting us. Well, great, thank you very, very much. I think that was, I know I learned a lot. I don't come from this area of research background myself and I think the context you provided and overview of your methods and the results as well as how it links of course to CLSA is extremely valuable. There's a couple of questions that did come in. First, I think I'll just address one quickly myself because I can and that was whether the CLSA has any other sensory measures such as smell. I can tell you that the answer is to, in terms of smell is no, but it is something that people have asked about and moving forward to CLSA is hoping to integrate. So hopefully perhaps within the next wave we may be able to integrate. I think it's like a scratch and sniff type test and so we'll have that as part of the CLSA moving forward. And then the other quick question I thought maybe Natalie or Paul could just answer quickly before we get into some of the bigger ones is to just explain exactly what life space is because I was wondering about that term myself and somebody else also asked that. Sure. Yeah, so the life space index is a measure of I guess one's footprint on the earth relative to their house. So with the life space index it takes into account five different levels of mobility. The innermost level is your bedroom. The second level is your house. The third level is your neighborhood. The fourth level is your city and then the fifth level is outside of your city. And so the measure calculates your frequency of travel to these different levels and whether or not you needed assistance and the score ranges between zero and 120. So if you have a score of 120 it means that you're traveling outside of your city without assistance every day of the week. If you have a score of zero it means you're bound to your bedroom and you require assistance even to be there all the time. I hope that makes sense. Right, hopefully I think it's clearer for me. So hopefully it's clearer for everyone. The other, the next question, again maybe a bit of a point of clarification and a bigger question but it was a question about corrections for vision loss through the use of aids. And the question is, are we talking about loss in general or loss that cannot be corrected for by aids? And the example is I definitely need glasses to read but if I can always see with eyeglasses and always have access to eyeglasses are these relationships not mitigated and then the same question for hearing aids? So if you can maybe touch base on that that would be great. Yeah, so there were different measures of visual acuity that I touched upon earlier. And so we had a choice of which ones to use. The habitually corrected visual acuity is your visual acuity wearing your glasses or contact lenses. And in contrast to the visual acuity that might be measured at a clinic or at an optometrist this isn't optimal correction. It's just what you have. And so if your glasses aren't optimally corrected that's the score you get. It's not corrected for. The pinhole corrected measure is an estimate of the optimal visual acuity. And this is what we used in this particular study. Other studies in other studies if it might be more interesting to use the habitual correction but for us we focused on somebody's optimal visual acuity. And then commenting on the issue of hearing loss and hearing aid status. The inclusion of hearing aid as a variable in the analysis does not change the effects. And of course, one of the challenges that we face is that individuals who do have a hearing loss frequently do not access hearing aid rehabilitation or even if they have a hearing aid, they don't always use it. So I hope that clarifies. Yeah, we basically didn't find a protective effect of wearing the hearing aid in this study. And that's very similar to other studies that have looked at that as well. So in studies that look at associations between hearing loss and cognitive decline the science isn't settled but many studies have not shown that wearing a hearing aid protects against that decline. But as Natalie said, the questions are fairly simple. They don't stratify people by the amount of time they've worn a hearing aid for, how regular they wear it. And so the question of whether hearing interventions will protect you from cognitive decline remains unanswered. Now, sorry, I think I'm here now. So now we'll, if you can touch base on the question from Mark Oremis at University of Waterloo, did you consider analyzing the data separately for each cognitive test rather than in combination through PCA? Yes, and we did do that. And so you had a little snapshot of that. We know that age is associated with lower performance on all of the individual cognitive tests. And that is true for the sensory variables as well. So the PCAs actually capture quite well what we learned from looking at the individual cognitive tests. So we stayed with the PCAs to have more straightforward models and to reduce the number of variables that we were ultimately looking at when we started adding in all the other covariates and so on. Okay, and I just a quick reminder I know some people often leave to start leaving around now. Hopefully you'll stay for a few more minutes, but if you can take the time to complete the poll questions before you go, that would be great. I'm just gonna go down to the next question now, which I can't actually see chat. Oh, there it is. Sorry about that. So great work. Have you had a chance to see if the sensory loss effects are mediated through lifestyle and behavior changes like regular physical activity, retirement, poor sleep and et cetera? I think that's a really great question. We haven't had the chance to do that. We do see though that there's a large variability in how sensory loss, like the risk associated with sensory loss and cognitive decline. So for two individuals who have the same level of hearing loss, for example, one might have a fairly steep cognitive decline and another one won't. And we don't quite know why there's this variation between people and in their risk. I mean, on average, sensory loss does put you at risk for cognitive decline as we've seen, but there's a lot of individual variability. And so maybe some of these factors, like you've mentioned, play a role in why one person seems to have a greater, there's a greater effect in one person versus another. And of course, we're most interested in these factors that are potentially modifiable like physical activity because that suggests that we can do something about it. And we did see interactions with the variables that we had mentioned earlier. So some of the variables that amplified the relationship between hearing loss and executive function, for example, would be age. So that effect is exaggerated in individuals who are older, who have lower levels of education, lower levels of income and more multi-morbid health conditions. So whether the converse is true, whether you could intervene with those variables that we can't answer, but we do know it interacts with these other factors. Yeah, we did see that partial retirement was possibly protective. And so that suggests that maybe there's something about working part-time in older age that it's protective maybe just in terms of access to social interaction or other things. I'm not sure we quite understand that yet, but... Great, I actually just related to that. I actually did have a question and if this question doesn't make sense, I again declare I have no background in this, but I was wondering whether, you know, I know the CLSA has a lot of social factors that we collect, that we measure and collect information about, but do you have a sense that there's any other social factors that might mediate the effect that CLSA isn't, that you weren't able to look at, like from the, you know, do we know from the literature or other evidence and that you weren't actually able to look at within the context of your cross-sectional research? Ooh, that's a good question. Depends on how broadly you want to define social, but we, in our analysis, we haven't looked at personality variables, and I'm embarrassed to say, does the CLSA include measures of personality? Because those could be quite interesting to look at. Yeah, I don't think in terms of personality, I know there's no measures. I don't think that would relate to personality. All right, we have another question, and probably the last question, unless we get one more through, a question with a follow-up question from the same person. Can you mention some of the co-morbidity variables that interacted negatively with the risk of sensory cognitive associations? And Roberta's specifically interested in mental health co-morbidities. Okay, thanks for asking that question. There were 27 that we looked at, we didn't look at them individually, we looked at them collectively, and really just did account. And these are really medical health co-morbidities rather than psychological co-morbidities. So this would be diabetes, hypertension, peripheral vascular disease, pulmonary disease, and so on. Okay, well, I see that, I don't think there's any other questions, and we are just about at time. So maybe I'll just do the closing spiel now. If anybody does have a question to post, please feel free to still post it. We do have a couple more minutes, and we could always stay on a couple of minutes longer if needed to answer it. So just to, again, say thank you again for the excellent presentation. We really appreciate your participation in the CLSA webinar series. I know you've presented before, I know at least Paul has. So thank you for coming back. I'd like to remind everyone that the CLSA data access request applications are ongoing if you're interested in applying for CLSA data. The next deadline for application is actually September 25th. So you can visit the CLSA website under the data access tab to review available data that we have, and further information, as well as details about the application process. I'd also like to remind everyone to complete their survey located under the polling option. If you have any questions or concerns that we can help with, please write to us in the chat box, and we can help out right now if you need. And then finally, if you could keep in mind that the CLSA promotes this webinar series using the hashtag CLSAWebinar, and we always invite you to follow us on Twitter at CLSA underscore ELCD. Just a note about our next webinar, which is going to be on October 29th. We'll be welcoming Assistant Professor from Brock University, Dr. Arn Stinchcom, as well as Dr. Kimberly Wilson from University of Guelph. I've realized that I don't have the topic or title of their presentation, but I'm sure it is going to be wonderful. So if you're interested in looking into that, please go to the CLSA website to register. And again, thank you very much for attending. And if we didn't get to your question, we'll make sure that we do follow up with you. All right, well, thank you very much. And this brings us to the end.