Machine learning tools for personalised diagnosis in dementia.

Predicting future development of dementia is one of the top priorities for research with major ramifications for clinical care and early interventions. Importantly, three major limitations prevent current research efforts on the early diagnosis of dementia from being translated into clinical practice: i) the widespread reporting of data at group– rather than individual– level, ii) the focus on single predictors of disease rather than interactive factors, and iii) the inclusion of prognostic measures such as tau- or amyloid-PET imaging that cannot be undertaken in routine clinical practice due to resource constraints. We propose to overcome these limitations by developing computational approaches that have the potential to advance precision medicine using machine learning to mine multimodal data from individual patients and determine personalised profiles of cognitive health.

In collaboration with Christopher Chen (National University of Singapore), William Jagust (University of California, Berkeley) and Pietro Lio (University of Cambridge).