ÁñÁ«ÊÓƵ18

Music invites measurable changes in a listener’s physiology. Heart rate variability and galvanic skin response are two physiological signals where we can view these changes. Prior work has shown physiological responses to high level aspects of music like genre and structure, so we explore low level audio features found in music information retrieval (i.e. machine learning for audio). We extract audio features and physiological features from the DEAP dataset and search for correlations with exploratory data analysis techniques. Based on results, we hypothesize that vocal or vocal-like sounds and percussive or noisy sounds in music may have particularly strong effects on the body.

Brain Music Lab