ML-Powered Handwriting Analysis for Early Detection of Alzheimer's Disease

Alzheimer's disease (AD) is a progressive, incurable condition leading to decline of nerve cells and cognitive functions over time. Early detection is essential for improving quality of life, as treatment strategies aim to decelerate its progression. AD also impacts fine motor control, including handwriting. Utilizing machine learning (ML) with efficient data analysis methods for early detection of Alzheimer's disease (AD) through handwriting analysis holds considerable promise for clinical diagnosis, albeit a challenging undertaking. In this study, we address this complexity by employing ensemble machine learning, which amalgamates diverse ML algorithms to enhance predictive performance. Our approach involves developing an ensemble model for analysis of handwriting kinetics, utilizing the stacking technique to integrate multiple base-level classifiers. The study encompasses 174 individuals, including 89 diagnosed with Alzheimer's disease and 85 healthy participants, drawn from the DARWIN dataset (Diagnosis AlzheimeR WIth haNdwriting). To discern the most effective base classifiers, we employ both Repeated-k-fold and Monte-Carlo Cross Validation techniques. Subsequently, top k features are selected for each best-performing base classifier using analysis of variance (ANOVA) and recursive feature elimination (RFE). The final step involves consolidating predictions from the base classifiers through a stacking ensemble, resulting in an impressive performance. The ensemble model achieves 97.14% accuracy, 95% sensitivity, 100% specificity, 100% precision, 97.44% F1-score, 94.37% Matthews Correlation Coefficient (MCC), 94.21% Cohen Kappa, and 97.5% Area Under the Receiver Operating Characteristic Curve (AUC-ROC). Comparative performance analysis demonstrates that our proposed model surpasses all state-of-the-art models based on the DARWIN dataset for Alzheimer's disease prediction. These findings underscore the potential of machine learning to offer highly accurate predictions for Alzheimer's disease in an affordable and non-invasive manner, emphasizing its significant clinical utility, particularly through handwriting analysis.