Longitudinal studies of cognitive aging: Multivariate and fMRI outcomes with non-ignorable dropout
Life expectancy from birth is increasing in many countries, and the number of individuals with age-related cognitive impairments is expected to at least double in the next 50 years. Cognitive decline due to aging may seriously affect an individuals´ well-being, and also represents an increasing welfare cost for the society in terms of, e.g., dementia care.
As we are facing longer lives and thus larger cohorts in older age groups, it is of increased interest to understand which factors affect the process of cognitive aging in general, and maintaining cognitive functions in particular.
The Swedish Research Council, 2013-2016: SEK 4,800,000 kr
We have a unique individual longitudinal data infrastructure with several separate studies covering up to 25 years follow-up, where cognitive aging is studied using memory tests, brain imaging (e.g., fMRI), and other information such as socio-economic, life-style, health and genetic data. The general purpose of this project is to develop statistical theory and methods to study cognitive aging. A major problem to be addressed in such follow up studies is non-ignorable dropout and thus we aim to: (1) Develop a framework to study the effect of causal agents on an outcome (cognitive measure) observed longitudinally, where dropout is not ignorable; (2) Develop methods for non-ignorable dropout when the longitudinal outcome is multivariate; (3) Develop methods for longitudinal fMRI outcomes in general, and in particular for situations when there is non-ignorable dropout.