Seyzethn Almubark
The capacity to predict the seemingly ambiguous transition from Mild Cognitive Impairment (MCI) to Alzheimer’s Disease (AD) is of critical concern in AD research. Advancement in computational systems has contributed to more robust potential to apply innovations in this sector. This study uses a Multilayer Perceptron (MLP) neural network approach to investigate and compare the utility of the Alzheimer’s Disease Assessment Scale-Cognitive subscale (ADAS-Cog), Mini-Mental State Examination (MMSE), and Functional Activities Questionnaire (FAQ) to predict a 3-year progression from MCI-to-AD. The MLP neural network is developed using the open database from the Alzheimer’s Disease Neuroimaging Initiative (ADNI). The data were based on a sample of 246 subjects with MCI whose diagnostic follow-up was available for at least the full 3-year period after the initial baseline assessment during the initial project period, i.e., ADNI-1. Classification results and analysis demonstrated that the combined features from all three neuropsychological tests outperformed a single test and the pairwise tests with an accuracy of 89.43%, a sensitivity of 89.19%, a specificity of 89.63%, and the area under the receiver operating characteristic curve (AUC) of 0.934.