Music-Induced Physiological Markers for Alzheimer's Disease Detection Using Machine Learning

Rodrigo Lima^1,2,3*Goncalo Barradas⁴Sergi Bermúdez i Badia^1,2,3

• ¹Faculdade de Ciências Exatas e Engenharia, Universidade da Madeira, Funchal, Portugal
• ²NOVA Laboratorio para a Ciencia da Computacao e Informatica - Polo Universidade da Madeira, Funchal, Portugal
• ³Agencia Regional para o Desenvolvimento da Investigacao Tecnologia e Inovacao, Funchal, Portugal
• ⁴University of Salford School of Health and Society, Salford, United Kingdom

ABSTRACT INTRODUCTION Alzheimer's disease (AD) involves progressive cognitive and emotional decline, and novel, non-invasive biomarker approaches are needed to support early detection, monitoring, and stage-specific interventions. This study investigates music-evoked physiological responses as potential biomarkers of AD and evaluates their translational value using machine learning. METHODS Thirty-six AD patients across different severity levels listened to emotionally evocative musical excerpts while electrodermal activity and facial electromyography (corrugator and zygomaticus muscles) were recorded. Machine learning models were trained on these signals to classify AD presence and severity and to detect residual emotion-specific physiological responses elicited by music. RESULTS Physiological reactivity to music declined with disease progression, with positive emotions eliciting more distinct responses than negative emotions. The Random Forest classifier distinguished AD from healthy controls with 70.5% accuracy, and Naïve Bayes predicted severity levels with 65.6% accuracy, showing that ML models can detect subtle, music-evoked physiological differences even in the presence of AD. DISCUSSION Music-evoked physiological signals reflect hierarchical disruption of emotion-related neural circuits in AD and have the potential to serve as complementary biomarkers of disease presence and stage. Combined with ML, these measures offer a non-invasive, ecologically valid approach to support early detection, monitoring, and the development of stage-sensitive interventions.