Harvard researchers publish BrainIAC foundation model in Nature Neuroscience, predicting dementia, stroke, and cancer from 48,900 MRI scans.

Harvard-affiliated researchers have developed BrainIAC, a groundbreaking AI foundation model trained on nearly 49,000 brain MRI scans that can predict brain age, assess dementia risk, and forecast cancer survival outcomes. The self-supervised learning model represents a major advancement in medical AI, offering a generalizable tool for analyzing routine brain MRIs to identify health risks and improve early diagnosis across multiple neurological and oncological conditions.

Harvard researchers debut BrainIAC, an AI tool trained on 49,000 brain scans that predicts brain age, dementia risk, and cancer survival with superior accuracy.

A landmark Swedish study of 100,000 women, published in The Lancet, found that using AI to assist in mammogram analysis led to a 12% reduction in interval breast cancer diagnoses, marking a significant breakthrough in medical imaging.

Analysis shows AI increases demand for radiologists despite automation; Mayo Clinic employs 55% more radiologists since 2016, demonstrating Jevons Paradox in AI.