Harmonic Healing: Personalized AI-Composed Music Therapy for Cognitive and Emotional Regulation in Neurodegenerative Patients

Neurodegenerative disorders such as Alzheimer’s disease, Parkinson’s disease, and clinical depression present profound challenges in cognitive function, emotional regulation, and quality of life. Music therapy has long been recognized for its non-invasive therapeutic effects on memory recall, mood stabilization, and neuroplasticity. This paper presents HarmonAI, a novel framework that integrates brain signal analysis and deep generative models to compose real-time, personalized music tailored to the emotional and cognitive states of neurodegenerative patients. Utilizing pre-recorded neural datasets (EEG/fMRI), the system maps emotion-related brain activity to dynamic musical parameters via deep learning models, including convolutional and transformer-based architectures. The generated music is designed to respond adaptively to fluctuations in neural signals, creating a closed-loop therapeutic feedback system. Experimental results demonstrate promising accuracy in emotion recognition and subjective coherence of music output, with potential applications in clinical, home-based, and digital therapeutic environments. This interdisciplinary approach bridges AI, neuroscience, and music therapy to create a new frontier in affective neurotechnology.