AI-Mind builds pre-symptom dementia tools, targets 2035 approval

The AI-Mind consortium, coordinated at Oslo University Hospital and funded by the EU, analysed data from 1,000 European citizens over more than 50,000 hours to build two tools designed to flag dementia risk before cognitive decline arrives, according to the European Commission's Horizon magazine.

"By the time cognitive decline has progressed to dementia, opportunities to intervene may already have been lost," said Dr Ira Haraldsen, a neurologist at Oslo University Hospital and the project's coordinator. The first phase closed in early 2026, and the team expects to seek approval under the EU Medical Device Regulation by 2035.

The two tools split the work. AI-Mind Connector analyses patterns in brain electrical activity linked to elevated risk. AI-Mind Predictor takes patients already flagged with mild cognitive impairment and estimates their progression risk from cognitive function, biological markers, medical history, age, education and location.

"It takes too long for the diagnostic procedures to find the right diagnoses for people, and for them and their caretakers to get the help they need," said Vebjørn Andersson, a PhD student and dementia researcher on the team. He said the tools are meant to support clinicians rather than stand in for them: "It's not supposed to make the diagnosis and then leave it up to the healthcare professionals to treat."

Haraldsen framed the intervention window in behavioural terms. "You can do something about social isolation, you can do something about drinking and smoking habits." She also said the delivery could be lightweight: "We have a very easy concept which could be delivered on iPads and iPhones, for example."

The consortium lists Eli Lilly, Roche and Bosch among its industry partners. The Horizon piece does not publish accuracy figures for either tool or a headline budget for the project.