Investigating Alzheimers Disease Biomarkers by Applying Machine Learning Models

Abstract

ObjectiveAlzheimers Disease (AD) is a debilitating neurodegenerative disorder characterized by memory loss, cognitive decline, and the accumulation of amyloid plaques and neurofibrillary tangles. This study investigates the interplay of various biomarkers and clinical features in diagnosing AD using machine learning (ML) techniques. MethodsWe analyzed data from 191 AD patients and 59 non-AD subjects, employing classifiers including Naive Bayes (NB), Random Forest (RF), Decision Tree (DT), Support Vector Machine (SVM), and K-Nearest Neighbors (KNN). ResultsOur findings indicate that KNN, SVM, RF, and DT achieved high sensitivity (94%) and accuracy (92%), demonstrating their potential as effective diagnostic tools. Notably, significant differences in feature values between AD patients and non-AD subjects suggest that biomarker-driven approaches can enhance diagnostic precision. Key biomarkers such as neprilysin, alpha-secretase, beta-secretase, amyloid plaques and urinary formic acid emerged as critical elements. ConclusionOur results underscore the importance of selecting a targeted subset of features to streamline the diagnostic process, allowing for more efficient and cost-effective screening. While our study reveals valuable insights into AD pathology and diagnosis, future research with larger, longitudinal cohorts is essential to further elucidate these relationships and enhance our understanding of Alzheimers mechanisms, ultimately aiming for innovative therapeutic strategies.