Dementia constitutes a major burden on society, both in monetary costs and the suffering of patients and their relatives. Alzheimer’s disease (AD), the most common form of dementia, is one of the most devastating healthcare problems faced by western society. The prevalence increases with the average lifetime year by year, and in most countries, the population is on average getting older.

Different types of dementia require different treatments. Cardiovascular treatments may prevent or slow vascular dementia (VaD), which is the second most common type of dementia. There are currently no disease-modifying treatments for AD, despite numerous promising drugs in development. To achieve effective treatment it is crucial to identify people at risk, who already have signs of pathology associated with one of the various types of dementia. Moreover, sensitive methods for assessing the efficacy of a treatment are needed, as during the early phases of dementia, the changes will still be quite small.

Another source of difficulty in terms of effective treatment is that many of the dementias have clinically similar presentations and may co-exist (mixed AD and VaD is the most frequent mixed condition, and as many as 50% of AD patients may have VaD as well), making clinical diagnosis challenging. However, signs of dementia-related pathology are measurable using biomarkers far before clinical presentation, and the various types of dementia have quite different pathological bases. Thus, methods to stage, differentiate, and identify at-risk sub-populations based on biomarkers are critical to many areas of dementia research.

Modelling the dynamics of dementia biomarkers over the entire disease course is still a very new research area. In 2010, Jack Jr et al. hypothesised a qualitative development of AD in terms of ordering which biomarkers attained  abnormal  values earliest during the disease process. Computational approaches have since then been proposed that infer the biomarker trajectories and ordering in a fully data-driven manner. In this project, deep-lerning based imaging biomarkers of dementia will be developed that, in conjuction with existing biomarkers, will serve as basis for a disease progression model in the spirit of Jack Jr et al. that models AD and VaD jointly (see Objectives for further details).