Machine learning algorithms to prevent Alzheimer's

Juan Domingo Gispert

BarcelonaBEta BRain Research Center, Spain

Dementia is caused by a variety of brain illnesses that affect memory, thinking, behavior and ability to perform everyday activities. Alzheimer’s Disease (AD) is the main cause of dementia.

Dementia is a primary global health challenge due to the ageing of the world population. 50 million people are living with it, and it is estimated that AD could reach epidemic levels by 2050, with a forecast of more than 130 million people affected.

While there is no current cure for AD, research shows its pathophysiological changes start decades before symptom onset, thus opening a window for prevention. It is estimated that 1 out of 3 cases of dementia could be prevented through lifestyle changes.

Alzheimer’s pathophysiological changes are detected via two gold-standard biomarkers, the Positron Emission Tomography (PET) and the Cerebrospinal Fluid (CSF). These are, however, very invasive and expensive, and thus not suitable for screening the general population. As a result, recruitment of cognitively unimpaired individuals for clinical trials remains a major unmet need for the realization of Alzheimer’s prevention strategies.

The aim of the project is to develop a family of machine-learning algorithms that could predict abnormality of core AD disease biomarkers from brain Magnetic Resonance Imaging (MRI). This would make it possible to significantly reduce recruitment expenses by up to 50% at no additional cost, and therefore identify the individual risk factors and provide a personalized prevention plan.