Periodic Reporting for period 4 - CRISP (Cognitive Aging: From Educational Opportunities to Individual Risk Profiles)
Reporting period: 2023-07-01 to 2024-07-31
The five-year project CRISP (2019-2023) investigates cognitive ageing and dementia from a life course and social perspective, with a particular focus on inequalities related to education, socioeconomic status, and gender. By understanding better how to create environments that enable the development of cognitive reserve over the life course, we will be able to make recommendations to policymakers in the domains of education and work. By improving current knowledge on who is at particular risk of cognitive decline, and who will benefit from lifestyle interventions, we will be able to identify vulnerable individuals and ways to support the delaying of cognitive decline. Ideally, we will be able to make recommendations for behavior changes in order to decrease risk of cognitive decline. This project receives funding from the European Research Council (grant agreement no. 803239).
Inequalities by education and socio-economic background, gender, and cognitive functioning later in life
Do societal conditions determine to which extent individuals are able to build up cognitive reserve? Since there is no medical cure available to delay cognitive ageing, we need to understand how to create the best possible environments to build up cognitive reserve. We investigate the different opportunities of men and women in terms of education, work and pay, and how they relate to cognitive performance in later life. We also investigate how inequalities in educational opportunities – schooling systems that favor children from higher socio-economic backgrounds – play out their influence on cognitive functioning over the life course.
Improving long-term dementia risk prediction and lifestyle interventions with new methods
We have some understanding about the high risk groups to develop dementia, and the Lancet Commission on dementia estimates that up to 45% of all dementia cases could be prevented by eliminating modifiable risk factors. There is also evidence on benefits of multidomain lifestyle interventions to delay cognitive decline. However, we have very limited generalized knowledge of what intervention works for whom and when, and over longer time periods. That is why we need to understand more clearly the long-term effects of risk factors, and the potential and limits of lifestyle changes. How do we do this? We use new causal inference frameworks to analyse observational data in order to identify target groups and promising components of lifestyle interventions. Additionally, we implement recently developed machine learning methods to improve estimation accuracy.
Inequalities by education and socio-economic background, gender, and cognitive functioning later in life
Do societal conditions determine to which extent individuals are able to build up cognitive reserve? Since there is no medical cure available to delay cognitive ageing, we need to understand how to create the best possible environments to build up cognitive reserve. We investigate the different opportunities of men and women in terms of education, work and pay, and how they relate to cognitive performance in later life. We also investigate how inequalities in educational opportunities – schooling systems that favor children from higher socio-economic backgrounds – play out their influence on cognitive functioning over the life course.
Improving long-term dementia risk prediction and lifestyle interventions with new methods
We have some understanding about the high risk groups to develop dementia, and the Lancet Commission on dementia estimates that up to 45% of all dementia cases could be prevented by eliminating modifiable risk factors. There is also evidence on benefits of multidomain lifestyle interventions to delay cognitive decline. However, we have very limited generalized knowledge of what intervention works for whom and when, and over longer time periods. That is why we need to understand more clearly the long-term effects of risk factors, and the potential and limits of lifestyle changes. How do we do this? We use new causal inference frameworks to analyse observational data in order to identify target groups and promising components of lifestyle interventions. Additionally, we implement recently developed machine learning methods to improve estimation accuracy.
The ERC CRISP project has given important insights into how contextual determinants and inequalities influence population-level cognitive functioning and -ageing, such as inequality of educational opportunity at time of schooling (doi: 10.1016/j.ssmph.2021.100837). We showed that sex/gender differences in prevalence of cognitive impairment and dementia were due to differences in risk factor burden, but not to risk factor-outcome relationships, across several studies with data from Europe, South Korea, and Brazil. However, we demonstrate an important role of gender norms for later-life brain health in the European and Latin American context. One study on gender-role attitudes and their associations with later-life cognitive functioning, considering also the role of work-family biographies (doi: 10.1007/s10433-023-00751-4) has received two prizes, the 2024 Health and Medical Sociology Excellence Prize by the European Society of Health and Medical Sociology (ESHMS), and the 2024 Vontobel Prize for Ageing Research (Vontobel Preis für Alternsforschung, University of Zurich, Switzerland), both awarded to the first author Ariane Bertogg.
We further provided a comprehensive review of theoretical mapping and empirical applications of machine learning in our fields of research (doi: 10.1126/sciadv.abk1942).
Among other important outputs, the work on cognitive impairment and dementia in Latin America and the Caribbean with regard to gender and socioeconomic inequalities was an important contribution on the role of contextual-level inequalities for population-level cognitive functioning and ageing, as for the first time we have a comprehensive discussion of the differences in brain health by sex/gender and educational level, and secular trends in brain health that suggest worsening population brain health, in fact quite different from those in high-income countries. These papers (among others: doi: 10.1186/s12877-021-02542-x; doi: 10.1016/S2666-7568(23)00052-1) help to significantly advance our understanding on sex/gender as determinant of older-age brain health, and question the universal applicability of findings in high-income countries on secular improvements in brain health for other world regions. We did this work to understand more deeply the differences by sex/gender, age and education, as well as the temporal trends in mild cognitive impairment and dementia in Latin America and the Caribbean. We find that the associations and secular trends of improved brain health observed in high-income countries (Europe and the United States) don't necessarily hold in this world region.
We further provided a comprehensive review of theoretical mapping and empirical applications of machine learning in our fields of research (doi: 10.1126/sciadv.abk1942).
Among other important outputs, the work on cognitive impairment and dementia in Latin America and the Caribbean with regard to gender and socioeconomic inequalities was an important contribution on the role of contextual-level inequalities for population-level cognitive functioning and ageing, as for the first time we have a comprehensive discussion of the differences in brain health by sex/gender and educational level, and secular trends in brain health that suggest worsening population brain health, in fact quite different from those in high-income countries. These papers (among others: doi: 10.1186/s12877-021-02542-x; doi: 10.1016/S2666-7568(23)00052-1) help to significantly advance our understanding on sex/gender as determinant of older-age brain health, and question the universal applicability of findings in high-income countries on secular improvements in brain health for other world regions. We did this work to understand more deeply the differences by sex/gender, age and education, as well as the temporal trends in mild cognitive impairment and dementia in Latin America and the Caribbean. We find that the associations and secular trends of improved brain health observed in high-income countries (Europe and the United States) don't necessarily hold in this world region.
(1) The paper on inequality of educational opportunity (SSM–Population Health) is the first to provide an extensive linkage between sociological theories of mobility and epidemiological and neuropsychological theories of cognitive development. It is also the first to investigate the role of unequal schooling systems (inequality of educational opportunity) on cognitive decline and identify those individual profiles with highest risk to be harmed by unequal schooling conditions.
(2) We improve the understanding between the social and health sciences and machine learning experts by providing an extensive mapping of possible machine learning approaches to research questions of description, prediction, and causal inference in the social and health sciences. This paper is the first of its kind to help researchers from different disciplines understand each other, and advance the uptake of new and improved methods in the social and health sciences:
Leist, A. K., Klee, M., Kim, J. H., Rehkopf, D. H., Bordas, S., Muniz-Terrera, G., & Wade, S. (2021). Machine learning in the social and health sciences. arXiv preprint arXiv:2106.10716.
(3) Continuing the ongoing research in the CRISP project, we expect to make significant contributions to improve our understanding in the following fields:
In the field of dementia prevention, to answer the question, what works for whom and when?, and determine the value of new protective and risk factors in cognitive ageing such as
- Continuing to work up to advanced ages
- Depressive symptom trajectories
- Area- and individual-level socioeconomic status and dementia
- Value of behavior changes to reduce risk of dementia in individuals with depression
- Effects of gender inequalities on later-life gender differences in brain health
(2) We improve the understanding between the social and health sciences and machine learning experts by providing an extensive mapping of possible machine learning approaches to research questions of description, prediction, and causal inference in the social and health sciences. This paper is the first of its kind to help researchers from different disciplines understand each other, and advance the uptake of new and improved methods in the social and health sciences:
Leist, A. K., Klee, M., Kim, J. H., Rehkopf, D. H., Bordas, S., Muniz-Terrera, G., & Wade, S. (2021). Machine learning in the social and health sciences. arXiv preprint arXiv:2106.10716.
(3) Continuing the ongoing research in the CRISP project, we expect to make significant contributions to improve our understanding in the following fields:
In the field of dementia prevention, to answer the question, what works for whom and when?, and determine the value of new protective and risk factors in cognitive ageing such as
- Continuing to work up to advanced ages
- Depressive symptom trajectories
- Area- and individual-level socioeconomic status and dementia
- Value of behavior changes to reduce risk of dementia in individuals with depression
- Effects of gender inequalities on later-life gender differences in brain health