AI Can Help Us Better Understand How Music Influences Our Emotions
Most people today have a soundtrack to their everyday lives — heavy beats for workouts and ambient sounds for work. According to a group of researchers, machine learning can be used to understand the way music influences our minds.
The scientists' new paper suggests that it might be possible to reverse-engineer the physiological effects of music.
Mapping the effects of music
In their new paper, researchers at the University of Southern California mapped out the way different factors in music, such as pitch, rhythm, and harmony, affect different types of brain activity, physiological reactions, and emotions. They say machine learning can be used to learn these reactions and predict how a piece of music will affect a person.
The research might eventually lead to the engineering of targeted musical experiences.
The study is part of the lab's broader aim to understand the way different forms of media affect people's brains and bodies.
Happy and sad songs
The researchers looked for songs on Spotify with few listens so as to avoid confounding variables, MIT Technology Review says.
Human testers were used to narrow a list of songs down to three: two that reliably induced sadness in listeners (Ólafur Arnalds’s “Fyrsta” and Michael Kamen’s “Discovery of the Camp”) and one that reliably induced happiness (Lullatone’s “Race Against the Sunset”).
One hundred participants who had not heard the songs were then asked to listen to all three.
They did this while having an fMRI scan taken or wearing a pulse, heat, and electricity sensor on their skin. The participants were also asked to rate the intensity of their emotions on a scale of 0 to 10.
Creating highly evocative music
The researchers then fed the data into several machine-learning algorithms and studied, which features most strongly predicted specific responses. They learned, for example, that the brightness of a song (measured via the level of its medium and high frequencies) were strong indicators for affecting a listener's heart rate and brain activity.
While the research is in its very early stages, the scientists believe it could eventually be used to design music specifically catered to an individual. It can better allow film directors to create a movie soundtrack that will have the effect they desire on an audience. Also, it can be used to help patients with mental health challenges and stimulate specific parts of their brains.