Understanding the Effects of Naturalistic Music Listening in Brain Circuit Dynamics using Neural Manifold Learning in Patients with Alzheimer's Disease

Alzheimer's disease (AD) causes severe memory and cognitive impairment, representing the most common and costly form of dementia without a cure. While current pharmacological treatments offer some symptom relief, they often fall short of desired clinical efficacy and may pose risks like radiation exposure and adverse side effects. Thus, non-pharmacological approaches like music therapy (MT) are gaining attention for their cost-effectiveness and accessibility, particularly in treating elderly patients. While numerous studies support MT's effectiveness in reducing AD symptoms, most focus on qualitative behavioral outcomes, highlighting the need for quantitative neurological measures. Electroencephalography (EEG) emerges as a valuable tool to track neural activity before, during, and after MT sessions, enabling the extraction of neurophysiological biomarkers. This study employs EEG datasets, one assessing the effects of different music genres on AD patients and another comparing healthy individuals with AD patients. Utilizing Neural Manifold Learning (NML) and Recurrence Quantification Analysis (RQA), the study explores changes in EEG data over time. Results reveal no significant difference in recurrence rates among music genres (p = 0.4164) but significant variation among AD severity classes (p = 0.00797). These findings underscore EEG's potential for quantifying MT's neurological impact in AD treatment.