RUSSO, MICHELA (2024) Human Movement Analysis and Artificial Intelligence for the Assessment of Parkinsonisms. [Tesi di dottorato]

Documento PDF RUSSO_MICHELA_ITEE_Thesis_37cycle_signed.pdf Visibile a [TBR] Amministratori dell'archivio Download (47MB) | Richiedi una copia

Abstract

Gait is a crucial biomarker in medicine for predicting and monitoring neurodegenerative and motor disorders, such as Parkinsonisms. These conditions are characterised by gait abnormalities, rigidity, and postural instability, often accompanied by non-motor symptoms like depression and cognitive decline. Gait analysis, combined with postural control assessment, provides quantitative insights into motor function. Detecting subtle changes, which are often difficult to identify through clinical analysis alone, is essential for effective diagnosis and management of Parkinsonian patients. Through both wearable and non-wearable technologies, in calibrated settings and home environments, even minor changes in movement, can be identified. In this context, artificial intelligence (AI) is enhancing the analysis of neurological gait data. Supervised learning improves the classification of motor patterns through labelled data, while unsupervised learning explores new features, useful for phenotyping parkinsonism variants. The main aim of this research was to explore the role of movement analysis in providing quantitative parameters to support current qualitative assessments, highlighting how movement analysis information can assist clinicians in their decision-making and in managing Parkinsonian patients. The study also examined the role of AI, specifically focusing on the question: how can AI techniques be integrated into movement analysis to enhance the accuracy and predictive power of assessments in neurodegenerative diseases? The research was conducted in close collaboration with the clinical staff of the Neurological Rehabilitation Department at San Giovanni di Dio e Ruggi d’Aragona Hospital (Salerno, Italy), ensuring that the methods and techniques used were relevant and applicable in a clinical setting.

Downloads

Actions (login required)

Modifica documento