查看文章

Background

Machine learning has been successful at predicting which patients have Alzheimer’s Disease based on their MRI. We developed a 3D CNN (Convolutional Neural Net) to predict cognitive function scores from MRI images of cognitively normal older adults.

Methods

A neural net was designed to include four convolutional layers that used 3x3x3 filters and the weights were adjusted using mean squared error loss (MSEL)(n= 159). The inputs to the CNN were MPRAGE MR-T1 weighted images and the outcomes were individual cognitive score and overall subdomain scores (ie, memory-retrieval, memory-learning, language, visuospatial, and executive attention) calculated from neuropsychological tests. A cross-validated linear regression to predict cognitive function from age served as a comparative model since it is a clinically popular model.

Results

When age was a statistically significant predictor of …