Machine learning predicts dementia 9 years in advance with 80% accuracy

A groundbreaking study by Queen Mary University of London predicts dementia up to 9 years in advance with over 80% accuracy using changes in the brain's default mode network. Published in Nature Mental Health, the research utilized fMRI scans and machine learning to identify early biomarkers for dementia. The model achieved an AUC of 0.824, outperforming traditional diagnostic methods. Social isolation was linked to DMN dysconnectivity.