AI-Driven Graphometric Assessment for Early Cognitive Decline Detection

Abstract

Alzheimer's disease is a progressive neurodegenerative disorder that gradually deteriorates cognitive function and motor skills, significantly affecting quality of life. Early identification plays a critical role in delaying disease advancement and improving patient outcomes. Graphometric analysis, specifically handwriting-based assessment, offers a non-invasive, cost-effective, and promising diagnostic path for early cognitive decline detection. Attention Deficit (AD)-related impairments in fine motor control are often reflected in handwriting patterns, making them valuable indicators for cognitive monitoring. Leveraging artificial intelligence and machine learning, a comprehensive methodology was implemented to extract, select, and evaluate critical handwriting-based features using the DARWIN dataset. To identify the most informative attributes, Recursive Feature Elimination with Cross-Validation (RFECV) and Analysis of Variance (ANOVA) were employed. These techniques allowed for the reduction of feature dimensionality and enhanced model interpretability. Robust classification models were developed using advanced validation strategies, including Repeated K-Fold and Monte Carlo Cross-Validation, ensuring reliable generalization performance. The classification framework incorporated a Voting Classifier ensemble, integrating outputs from various machine learning algorithms to maximize prediction accuracy. Notably, the Voting Classifier achieved a perfect accuracy score of 100% when trained on ANOVA-selected features, while attaining an 88.6% accuracy with RFECV-selected features. These findings underscore the potential of combining intelligent feature selection methods with ensemble learning approaches for reliable cognitive decline assessment through handwriting analysis. The integration of graphometric indicators and machine learning techniques presents a viable pathway for the development of early diagnostic tools, aiding timely intervention in Alzheimer’s progression and improving patient care strategies.