Multimodal AI for Early Dementia Prediction | Prof. Zoe Kourtzi University of Cambridge & Prodromic

Discover how multimodal AI is transforming the early prediction of dementia and other neurodegenerative diseases in this keynote by Zoe Kourtzi, Professor of Cognitive Computational Neuroscience at the University of Cambridge and Chief Scientific Officer (CSO) at Prodromic, delivered at the Fourth Workshop on Multimodal AI (MultimodalAI'26). Early diagnosis of dementia is critical for improving patient outcomes, yet predicting which individuals will experience cognitive decline remains a major clinical challenge. This keynote presents a novel clinical AI prognostic modelling approach that combines multimodal data to generate personalised markers of cognitive decline, even before symptoms appear. The presentation demonstrates how the proposed AI approach, validated using real-world patient data from memory clinics, outperforms conventional methods such as cognitive assessments and MRI scans. Learn how multimodal AI can support earlier diagnosis, improve patient stratification for clinical pathways and trials, reduce misdiagnosis, and accelerate clinical translation and drug discovery. This keynote was recorded at the Fourth Workshop on Multimodal AI (MultimodalAI'26), organised by the UK Open Multimodal AI Network (UKOMAIN). Subscribe for more keynote talks, tutorials, and discussions on multimodal AI, machine learning, healthcare AI, and responsible AI. Follow us on LinkedIn:   / ukomain   Visit us: https://multimodalai.github.io/ #MultimodalAI #MultimodalAI26 #Dementia #ArtificialIntelligence #HealthcareAI #MachineLearning #UniversityOfCambridge #Prodromic #ClinicalAI #MedicalAI #NeurodegenerativeDisease #AIResearch #UKOMAIN