Computational models over the lifetime: discovering the roots of dementia
The causes and earliest signs of neurodegenerative diseases, such as Alzheimer’s disease (AD), remain poorly understood. However, such understanding is essential to develop disease-modifying treatments, which currently remain elusive. Over the last decade, data sets from large groups of dementia patients have become available that provide fundamental new insight into disease biology once it takes hold. For example, the Alzheimer’s Disease Neuroimaging Initiative (ADNI: http://adni.loni.usc.edu/) maintain an open data set (including image data, cognitive scores, genetics, protein measurements, and demographic data) from thousands of AD patients and cognitively normal controls. However, at these relatively late stages, very likely it is already too late to slow or reverse the disease substantially, since pathology has already taken hold. Recent data gathering exercises, such as Insight46 http://www.alzheimersresearchuk.org/research-projects/using-brain-scans-investigate-changes-brain-time/, combine imaging data sets acquired in old age, with a wide range of measurements acquired over the whole life course. It studies a cohort of subjects all born in 1946 and now reaching ages where, statistically, many of them will be in the earliest stages AD and other dementias. This provides a unique opportunity to use machine learning and modelling to find links between early and mid-life events, choices, etc. and late-life diseases, such as AD.
Aims and Objectives:
This project explores rich whole-lifecourse data sets, such as Insight46 and the wider 1946 population study run by the MRC National Survey of Health and Development (NSHD: http://www.nshd.mrc.ac.uk/lha), for early signs and causes of neurodegenerative diseases. It will build on recent advances in computational modelling and machine learning within the POND group at UCL (http://pond.cs.ucl.ac.uk) to elicit new understanding of disease aetiology from such data.