Predicting Alzheimer’s Disease: A Stroke-Based Handwriting Analysis Approach Based on Machine Learning

Abstract

Alzheimer’s disease (AD) causes most of dementia cases. Although currently there is no cure for this disease, predicting the cognitive decline of people at the first stage of the disease allows clinicians to alleviate its burden. Recently, machine learning based approaches have been developed to automatically analyze handwriting to support early diagnosis of AD. In this context, features are extracted from the coordinates of the handwriting traits, recorded using digital devices. For a given task used for data collection, typically these features take into account the whole handwriting making up the task. In this paper, we present a novel approach to predict Alzheimer’s Disease based on machine learning that analyses the single elementary traits making up handwriting, named strokes. The experimental results confirmed the effectiveness of the proposed approach.