Significant progress has been made during the project's third reporting period. A key milestone was the successful finalisation of participant enrollment, reaching the clinical study target of over 1,000 individuals with mild cognitive impairment (MCI). This was achieved despite recruitment challenges at one site, which were effectively mitigated by reallocating recruitment efforts to other countries. The completion of the AI-Mind central database—hosted on the secure Services for Sensitive Data (TSD) platform at the University of Oslo—is well underway and near finalisation. This database integrates multimodal data from all clinical partners. Significant work has also gone into the harmonisation, standardisation, and BIDS-format storage of this data, ensuring its readiness for machine learning and clinical validation tasks.
The project has continued its close engagement with stakeholders, including clinical partners, industry representatives, and end users, through co-creation workshops and participatory design. This has guided the development of the AI-Mind Connector and Predictor tools—two AI-powered systems for early risk detection of dementia. While preliminary versions have been tested on subsets of data, the finalisation and full deployment of these tools is expected in the next period. Additionally, the project has developed a suite of educational and training materials to support end-user understanding and clinical adoption. A dedicated knowledge-sharing hub has been established on the AI-Mind website to host these resources and facilitate stakeholder engagement.
Notably, AI-Mind has carried out extensive dissemination and communication activities across multiple channels. These include participation in major international conferences, publications in peer-reviewed journals, social media outreach, and a high-visibility presence national medias. AI-Mind has also played a key role in fostering collaboration across European research project, becoming a partner in the newly funded EU-project FluiDx-AD.
Several scientific publications have been submitted during this period, including studies on functional connectivity metrics, validation of AI models, explainability in AI-supported diagnostics, and the development of ethical guidelines for trustworthy AI in clinical contexts.
Progress in innovation and exploitation has also been notable. The AI-Mind tools are being refined in accordance with the regulatory requirements for medical device software under the EU MDR. The decision-analytical model developed in WP6 for health technology assessment (HTA) has already been adopted in other EU-funded projects, demonstrating AI-Mind's broader value. Intellectual property assessments have been carried out, and discussions are ongoing with industrial partners on post-project commercialisation pathways.
In close collaboration with its industrial partners, the project continues to advance toward a groundbreaking clinical breakthrough—enabling scalable, AI-supported screening tools that can predict dementia risk in MCI patients before irreversible cognitive decline sets in. With all core components under development or in testing, AI-Mind is well-positioned for clinical implementation and real-world impact by the end of the project.
