A research hub led by the University of Exeter is to build AI and data systems aimed at preventing and detecting dementia earlier than is currently possible.
The BRAINIAC Hub — AI and Emerging Technologies for Next Generation Brain Health Clinics — is funded by £10m from the Engineering and Physical Sciences Research Council, part of UK Research and Innovation, with another £10m from industry, charity and NHS partners.
What it is actually building
The stated goal is specific and worth reading closely: AI that combines brain imaging, genetics, blood biomarkers and digital cognitive tests into "a single, clear picture of a person's brain health".
Each of those four already exists on its own. A scan is read by a radiologist. A genetic risk score is calculated by one service. Blood biomarkers are a newer and fast-moving field. Cognitive tests are administered separately again.
What does not exist is a way of combining them. They arrive in different formats, at different times, through different departments, and no clinician has a tool that weighs all four against each other for one patient.
The scale, and the timeline
BRAINIAC brings together 11 universities and 10 NHS sites over six years. The first clinics are to start operating in 2029, initially open to people by NHS referral.
That date deserves attention, because it is three years away. This is funding for building something, not an announcement of a service. Anyone reading this hoping for a test their relative can take is reading about a plan.
Why early detection is the whole argument
Dementia is currently diagnosed late, usually once symptoms are clear enough for someone to seek help. By that point the underlying changes have been under way for years — in some cases more than a decade.
That matters because treatments which slow progression work best early, and because the things known to reduce risk — blood pressure control, hearing correction, activity, managing diabetes — are preventive measures that need time to act. A diagnosis delivered after symptoms is too late for most of them.
So a tool that flags elevated risk before symptoms is not simply an earlier version of today's diagnosis. It opens a window that currently closes before anyone knows it was there.
The harder question underneath
Early detection also creates a problem that the funding announcement does not address: what happens to a person told they are at high risk of a condition with no cure.
That is not an argument against doing it. It is an argument for the counselling, the follow-up and the clear communication of uncertainty being designed in alongside the model, rather than added once the technology works.
The hub is one of five the research council has funded in this round, each aimed at spotting disease sooner. Judged in three years, the question will not be whether the AI can combine four data sources. It will be whether ten NHS sites can act on what it finds.
What to watch for in 2029
Two details in the announcement will decide whether this reaches patients or stays in journals. The clinics open by NHS referral, which means a GP has to recognise who should be sent — and GPs are the part of the system with the least time per patient.
And the hub spans eleven universities and ten NHS sites, which is a strength for gathering varied data and a known weakness for agreeing on anything. Multi-site research projects usually founder on whose data format wins, not on whether the science works.
