You are What (and How) you listen to.

Abstract

Depression is a rising global mental health issue, affecting individuals worldwide. This study focuses on exploring the correlation between music engagement and depression risk, with specific attention to the ”how” and ”what” of music. By analyzing individuals’ music preferences, consumption patterns, emotional responses, and lyrical content, this research aims to gain insights into their mental state and potential vulnerability to depression. The HealthyUnhealthy Music Scale (HUMS) is evaluated as a non-invasive tool for assessing depression risk based on music engagement strategies. Machine learning models, such as Support Vector Machines (SVM) and deep learning techniques, are employed to predict mental well-being using music associations. The study acknowledges the limitations of existing assessment tools, particularly the intrusive nature of some scales, and proposes the use of the HUMS scale as a promising alternative. The HUMS scale focuses on individuals’ motivations and experiences with music, considering both hedonic and eudaimonic aspects of music engagement. By assessing music listening strategies on a Likert scale, it aims to capture a holistic understanding of the role of music in individuals’ lives and its impact on mental health. Furthermore, the research delves into the influence of lyrics on music engagement and its association with depression risk. It recognizes that the emotional impact of music is not solely determined by its acoustics; lyrics play a significant role in shaping the perceived emotion of a song. By analyzing the emotional connotations and semantic themes in lyrics from individuals’ music listening history, the study aims to identify patterns associated with depressive mood. This analysis provides valuable insights into the influence of lyrical content on emotional wellbeing, enhancing our understanding of the complex relationship between music and mental health. Machine learning models are employed to analyze lyrics to determine their emotional connotations in an automated fashion and also to predict risk to depression based on consumption strategies. These models aim to identify potential markers of depressive tendencies and predict an individual’s risk of depression based on their music engagement data. The use of advanced algorithms, such as Support Vector Machines (SVM) and deep learning techniques, enables the development of predictive models that can assess depression risk through music. The findings of this study have significant implications for the field of mental health assessment and intervention. Validating the HUMS scale as a depression screening tool in the Indian context contributes to non-invasive and culturally sensitive approaches to detecting depression risk. The application of machine learning models to predict mental well-being based on music associations offers new possibilities for personalized interventions and early detection methods. Moreover, exploring the relationship between lyrics and depression risk expands our understanding of the emotional impact of music beyond its acoustic properties. This interdisciplinary research merges computer science, psychology, and musicology to explore the intricate relationship between music and mental health. By bridging the gap between technology and mental health, the study aims to drive innovation and create effective tools for assessing and addressing depression risk through music engagement. The outcomes of this research have the potential to make a positive impact on individuals’ lives by helping to alleviate the global burden of depression