Determining music features that have an effect on stress levels based on physiological signals

• Main Author: Uy, Gemilene C.
• Format: text
• Language: English
• Published: Animo Repository 2013
• Subjects: Music > Physiological effect
• Online Access: https://animorepository.dlsu.edu.ph/etd_masteral/4389
• Institution: De La Salle University

Music has been known for alleviating stress and promoting relaxation for its listeners. This research determines the correlation between music and listeners physiological responses in terms of Galvanic Skin Response (GSR), Blood Volume Pulse (BVP) and Respiration. This was done by first building a physiological and stress level model in order to determine the correlation between all the physiological signals with stress levels that were self-reported by the test subjects. After which another model, music-stress model was built, that was able to correlate music features such as harmonics, tempo and dynamics to the corresponding stress values. There were two data collection setups: The first was the controlled set-up which included two activities that induced stress: Paced Stroop Test and Mental Math Test. The second set-up was a naturalistic set-up which allowed the test subjects to do academic activities as a form of stress inducement. During the natural set-up, the test subjects were given a list of classical music to choose from and listen to. Both set-ups used test subject annotations as the stress level labels for the model. The research was able to determine a correlation with stress and the physiological signals with an accuracy of 79.82% and 73.52% for the controlled and natural set-up respectively. The research had also determined a correlation of music features with stress, determining the likelihood of a certain stress level to be elicited given a certain music.