The Pleasure Arrives Early

Most people who listen to music seriously know the experience. A particular moment in a particular piece, and a wave of shivers runs up the back of the neck or down the arms. It arrives at the same place every time, which is the interesting part: something in the music is reliably producing something in the body.

The first question is where. In 2001 Anne Blood and Robert Zatorre, in Montreal, had volunteers bring music that gave them chills and scanned them with PET while they listened. Blood flow rose in regions associated with reward and emotion, the circuitry that responds to food and to drugs. That was a striking result for something with no obvious biological use, but blood flow is an indirect measure. It said these regions were busy, not what they were doing.

The second question is what chemical. In 2011 Valorie Salimpoor and colleagues in the same lab used a PET tracer, raclopride, that competes with dopamine for the same receptors, so that less binding means more dopamine released. They screened volunteers for people who reliably got chills from music, had them bring their own, and compared it with music that left them cold. Dopamine was released in the striatum during the music they loved and not during the neutral music.

Then they did something cleverer. PET is slow, so they scanned the same people with functional MRI, which can follow activity second by second, and lined it up with the moments of chills, marked by changes in skin conductance, heart rate and breathing. Two regions came apart. The caudate was more involved in the seconds before the peak. The nucleus accumbens was more involved during the peak itself.

That is the finding worth keeping. The pleasure has a time structure, and part of it comes before the moment it seems to be about. While the build is still building, while the singer is drawing breath for the high note, the reward system is already working on a prediction.

This fits what music does structurally. In 1991 John Sloboda asked listeners to pin down the exact moments in pieces that had given them physical reactions. Shivers clustered at new or unprepared harmonies and at sudden changes of texture and loudness. Those are moments where an expectation has been set up and is then answered or overturned. A chill is not a response to a pleasant sound. It is a response to a prediction, and the waiting is part of the bargain.

It also explains why nobody's list of chills is the same as anybody else's. Expectation depends on what you have heard before, so what is a long-awaited arrival to one listener is a cliché to a second and noise to a third. That is why the Montreal studies let participants choose their own music. There is no universal chill.

Scans show correlation, so the next step was a drug. In 2019 Laura Ferreri and colleagues gave twenty-seven people, in three separate sessions, levodopa, which boosts dopamine signalling, risperidone, which blocks it, and a placebo. Under levodopa, listeners rated music as more pleasurable and were willing to pay more for it. Under risperidone, both fell. That is about as close to cause as this kind of research gets.

And the system can fail selectively. In 2014 Ernest Mas-Herrero and colleagues described people with what they called specific musical anhedonia. They perceive music perfectly well and respond normally to winning money, but music gives them no pleasure at all, and their heart rate and skin conductance do not respond to it either. The route from music into the reward system is a route, and it can be missing on its own.

What is music for? That is still argued, and none of this settles it. What it settles is the mechanism. Music has found its way onto machinery built to make an animal want food and go on looking for it, and it works that machinery largely by making you wait.