BrainHealthX: Responsible AI-guided solution for early dementia prediction

Dementia is stealing the lives of >55 million people worldwide, with huge societal cost (>$1 trillion p.a.). Despite >$56 billion R&D spend over 30 years, we lack sensitive diagnostics at early stages, when interventions may work best. We co-created—with healthcare partners, clinicians and the public—a responsible, multimodal AI tool (BrainHealthx, BHx) to improve early prediction and patient stratification to optimise interventions for each patient. This RAi Enterprise Fellowship aims to scale-up BHx into a trusted, fully deployable clinical decision support system to help identify who will benefit when from which intervention, ultimately driving precision interventions and new treatments.

Activities and Achievements

Our project has the following objectives:

1. Enhance BH^xfor clinical use: validate BH^x for non-invasive, low-cost data that can be routinely collected across geographically and socio-economically diverse real-world settings.
2. Maximise BH^xutility in community settings: validate BH^x for primary care data and assess utility in community settings (GP practices, Community Diagnostic Centres: CDCs).
3. Scale-up BH^xto a clinical decision support system that adheres to medical device software standards for deployment into clinical platforms and drug discovery pipelines.

Impact

Our technology can:

a) reduce bias and misdiagnosis,

b) improve equitable access to diagnosis,

c) push dementia prediction upstream into community settings,

d) transform patient stratification for inclusion in clinical trials,

e) optimise clinical management, leading to precision interventions

Outputs and Outcomes