Why Do We Like the Music We Like?

Consider this question: Why do we like the music we like? Were we born with musical preferences, or did we develop them over time through life experiences? This post, Part 2 of a 3-Part series, explores a theory that emphasizes the latter possibility. Known as predictive coding, this model argues that musical preferences are largely learned. Next month, we will look at a contrasting notion that stresses innate musical tendencies. Both approaches lean into the concept of embodiment, the overall theme of this series.

How does the brain turn the basic elements of music into sentiments (emotions and feelings)? How do rhythm and melody – timing and frequency patterns of sound waves – become experiences of joy, sadness, excitement, or comfort? According to predictive coding, the musical qualities we learn to value inform a certain brain system which musical elements to reward.

Our reactions to music link our brain with our body

Source: Image by Elizabeth Lemos / Pixabay

This system, located in the limbic region of the brain, is called the reward prediction error system (RPE), or simply the reward system.¹ The RPE responds to predictions about the musical features we anticipate for two key reasons. First, the auditory cortex and the reward system are highly connected to one another.² Second, a key role of the human brain is to forecast what will happen next; its ability to predict the next musical note is an expression of this function.

As we listen to music, the auditory regions of the brain make predictions about upcoming sounds and share those predictions with the reward system. The RPE becomes active when the actual music differs from what was anticipated. This mismatch generates emotions and feelings.

Robert Zatorre, Ph.D., a professor at McGill University, is one of the world’s leading music neuroscientists and a major proponent of the predictive coding model. He notes that the brain’s auditory system analyzes musical patterns and makes predictions about them. The reward system then evaluates the results of those predictions, creating positive or negative emotional responses depending on whether expectations are met, not met, or exceeded.³

Zatorre also notes that every musical style follows its own set of rules, or syntax, which the auditory system can begin to understand even in infancy. The brain’s ability to make predictions also appears very early in life.⁴ Our general understanding of the music we hear is then complemented by personal taste, which develops through our experiences over time. As a result, the brain’s analyses and predictions about music are grounded in learning—through both broad knowledge and individual life experience.⁵

How do these subconscious predictions become the sentiments about music that enter our awareness? The auditory and reward systems are closely connected.⁶ After forming predictions, the auditory system signals to the RPE which musical features to reward. When rewarding features are recognized, the RPE stimulates the release of a suite of neurotransmitters, especially dopamine, the chemical most closely associated with pleasure.

We cannot consciously experience the release of dopamine itself. Instead, dopamine triggers physical reactions throughout the body, such as speeding up the pulse and causing the skin hairs to stand on end. Using its map of the body,⁷ the brain then becomes aware of these physical changes—the flutter of the heart, goose bumps on the skin—by truly feeling them. This is what is meant by the embodiment of our response to music. In this case, the response is pleasure, which encourages us to seek out and listen to more of the music we enjoy.⁸

The predictive coding model generates two major critiques. First, it is a “top-down” model wherein musical preferences arise mainly from activity on the brain’s surface (cortex), the outer layer associated with higher-order thinking.⁹ According to this theory, musical preferences are learned rather than innate, shaped more by culture and experience than by biology.¹⁰ If this is true, then people could potentially be taught not only to understand, but also to enjoy, any musical system.

Second, predictive coding argues that emotional responses to music come from mismatches between expectation and reality; the brain predicts one thing, but the music delivers another. This explanation makes sense the first time we hear a song, or maybe even the tenth time. But what about the hundredth time? By then, the music is completely familiar, and there are no surprises left. Yet we may still love the song just as much as ever. If there are no more prediction errors, how does the emotional response remain so powerful?

Neuroscience Essential Reads

Social intelligence Arises Between Minds

How Social Context Influences Brain Disorders

Part 3 of this series, where we’ll explore how biology shapes musical preferences and responses, will appear next month.