The Physiology of Musical Preference

Abstract

The dominant research strategy within the field of music perception and cognition has typically involved new data collection and primary analysis techniques. As a result, numerous information-rich yet underexplored datasets exist in publicly accessible online repositories. In this paper we contribute two secondary analysis methodologies to overcome two common challenges in working with previously collected data: lack of participant stimulus ratings and lack of physiological baseline recordings. Specifically, we focus on methodologies that unlock previously unexplored musical preference questions. Preferred music plays important roles in our personal, social, and emotional well-being, and is capable of inducing emotions that result in psychophysiological responses. Therefore, we select the Study Forrest dataset “auditory perception” extension as a case study, which provides physiological and self-report demographics data for participants (N = 20) listening to clips from different musical genres. In Method 1, we quantitatively model self-report genre preferences using the MUSIC five-factor model: a tool recognized for genre-free characterization of musical preferences. In Method 2, we calculate synthetic baselines for each participant, allowing us to compare physiological responses (pulse and respiration) across individuals. With these methods, we uncover average changes in breathing rate as high as 4.8%, which correlate with musical genres in this dataset (p < .001). High-level musical characteristics from the MUSIC model (mellowness and intensity) further reveal a linear breathing rate trend among genres (p < .001). Although no causation can be inferred given the nature of the analysis, the significant results obtained demonstrate the potential for previous datasets to be more productively harnessed for novel research.