Deep-learning based prediction of Alzheimer's disease using genomics and neuroimaging
Approved Research ID: 56236
Approval date: August 4th 2020
Alzheimer's disease (AD) is a neurodegenerative disorder, characterized by progressive cognitive decline and is the most common form of dementia. AD is an important reason for concern since it is affecting more and more people over time and no treatment is available yet. Brain cells are damaged during the process of developing AD and since this damage is irreversible, a huge focus is on predicting the onset of the disease. The goal is to find a reliable way to identify individuals who will probably develop AD as well as the time to conversion for those individuals and after achieving this goal, it will become possible to find and apply a treatment to prevent or slow down the brain tissue damage. Alzheimer's disease is a complex disease and depends on many factors. Genetic variants contribute to the risk of developing AD. Each of these genetic risk factors contributes to the disease risk in a different way. In some cases, the contribution to the disease risk will be small, while in other cases having a specific variant can have a large impact on the disease risk. In this project we will design a deep neural network (DNN) to predict an individuals' time to conversion to AD using combination genetic data, structural measurements extracted from image modalities such as MRI and clinical information gathered from different sources including UK Biobank. Our hypothesis for this project is that genetic markers will increase the accuracy of identifying those that convert to DAT. We hope to develop a clinical software in the next following years to predict the accurate time to conversion to AD that can facilitate the process of searching and applying the right cure at the right time.