A study published in Nature Aging on 1 July shows that a single, routine brain MRI can estimate how quickly a person is biologically ageing and forecast their risk of dementia, chronic illness and premature death. Duke University scientists trained an algorithm—dubbed DunedinPACNI—on 860 scans from the long-running Dunedin birth cohort to read cortical thickness, grey-matter volume and other structural features that change with age. When the model was applied to more than 50,000 additional scans, including 42,583 from the UK Biobank and 1,737 from the Alzheimer’s Disease Neuroimaging Initiative, people whose brains appeared to age faster than their peers showed poorer cognition, steeper hippocampal atrophy and a higher likelihood of progressing from mild cognitive impairment to dementia. Among initially healthy adults, those in the fastest-ageing decile were at least 18 % more likely to develop major chronic diseases over the following decade and 41 % more likely to die during the study period. The imaging ‘clock’ performed consistently across populations in North America, Europe and Latin America and showed an intraclass test–retest reliability of 0.94 in a subset of the Human Connectome Project. It also captured well-known social gradients: lower education and income were linked to faster brain-measured ageing. Because the method relies on standard hospital MRI sequences, the authors say it could be incorporated into existing imaging workflows to screen middle-aged adults long before symptoms emerge. Mahdi Moqri of Harvard Medical School, who was not involved in the work, called the tool “a direct window on structural brain ageing that blood-based markers can’t provide,” while the Duke team has released the code openly to speed independent validation.