Mobile Sensing of Altered EveryDay Function in Early Alzheimer's Disease (MEDEA)

This project will develop and validate a suite of app-based tests for detecting the effect of early Alzheimer's disease (AD) on everyday functions (sleep, navigation and spoken communication). This suite will transform clinical practice by i) basing diagnosis on ecologically valid, real life outcomes rather than clinic tests with limited real-world relevance and ii) delivering a low cost, scalable method for detecting early AD in the ageing population. The project will be divided into separate Work Packages (WPs). The WP1 objective will be to finalise test suite development and undertake feasibility studies in patients and controls. WP2 will focus on benchmarking of MEDEA for future clinical usage, in terms of operational (user acceptability, compliance and adherence) and diagnostic (classification accuracy, sensitivity and specificity, validation against AD biomarkers) properties. WP3 will focus on privacy, using on-device analytical approaches to maintain user privacy. Project success will additionally benefit AD research. It will deliver real life outcome measures determine the effect of interventions in clinical trials. Application of machine learning and computational statistics to the datasets produced by the test suite will aid disease modelling and uncover new phenotypic features, generating new testable hypotheses on associations between pathophysiology and real world outcomes.