Prediction of Mini-mental State Examination Scores via Machine Learning for Alzheimer’s Disease

Alzheimer’s disease (AD) is a kind of common disease among the elderly which can affect function of cognitive abilities. Machine learning (ML) has been applied to the prediction of AD recently, but due to the limitations of single modal data, its performance still can be improved. In this study, we totally downloaded 216 data of patients for three groups: AD, mild cognitive impairment (MCI) and normal control (NC). Using the joint feature set of brain cortical characteristics along with biology, risk factors, PET measures and cognitive scores, four machine learning algorithms were applied to predict the mini-mental state examination (MMSE) scores of patients at sixth month (M06) and one year later (M12). The best performance of mean absolute error (MAE), mean squared error (MSE) and root mean squared error (RMSE) reached 1.400, 2.313, 5.307, respectively. Compared with previous researches, our MRI-based joint feature set showed valuable capability in the task of MMSE scores regression.