Quantitative Tapping Analysis in People with Alzheimer’s Disease: towards a Digital Twin for Personalized Smart Health

Neurodegenerative diseases, such as Alzheimer's disease (AD), pose an increasingly significant global health challenge, characterised by the progressive deterioration of cognitive and motor functions. Traditional clinical assessments, such as the Mini-Mental State Examination (MMSE) and the Montreal Cognitive Assessment (MoCA), are often open to subjective interpretation and inter-rater variability. Objective, automated monitoring systems could help to bridge this gap. This study focuses on the Finger-Tapping Test (TT), a clinical standard for assessing motor coordination and rhythmic consistency that is traditionally performed through manual observation. An advanced framework is proposed that utilises wearable technology systems (WTS), specifically inertial measurement units (IMUs) attached to the fingers, to acquire high-resolution kinematic data. Once the features that best distinguish the Alzheimer's group from the control group have been identified, machine learning algorithms are implemented to classify healthy people from those with the disease. Initial results show that this distinction can be made with 98% accuracy. However, once the disease is confirmed, distinguishing between two severity levels (moderate versus severe) yields an accuracy of around 70%. The initial results provide a solid foundation for designing an IoT-based platform that enables real-time data collection and historical monitoring. This architecture enables the creation of a Digital Twin for each patient, equipping clinicians with a tool that can monitor disease progression, rehabilitation effectiveness, and treatment adherence.