On the Class-median Threshold based Optimum Boundary for Classification between Mild Cognitive Impairment and Healthy Control

Alzheimer’s disease (AD) is a progressive neurodegenerative brain disorder prevailing throughout the globe with no known medical treatment discovered so far. Researchers and medical professionals have developed different machine learning (ML) algorithms for early detection of AD based on brain MRI scans and other imaging techniques. The main motive of this proposed work focuses on developing a novel classification algorithm for detection of mild cognitive impairment (MCI) patients from the healthy control (HC) ones, based on potential biomarkers like cerebrospinal fluid (CSF), grey matter (GM), and white matter (WM) acquired from various MRI scan images of patients gathered from the ADNI dataset. In this work, a unique machine learning model is proposed and applied to the three pairs of independent biomarkers stated above at a time and finally a majority voter classification is used upon the three interim decisions for the final conclusion. A number of performance metrics, such as accuracy, sensitivity, specificity, and precision is evaluated and compared with the other state-of-the-art algorithms as well.