The Neuroscience of Music

There was a nice article in The Times on Sunday about the research of Daniel Levitin, a neuroscientist at McGill (and former record producer) who studies the neural substrate of music:

Observing 13 subjects who listened to classical music while in an M.R.I. machine, the scientists found a cascade of brain-chemical activity. First the music triggered the forebrain, as it analyzed the structure and meaning of the tune. Then the nucleus accumbus and ventral tegmental area activated to release dopamine, a chemical that triggers the brain's sense of reward.

The cerebellum, an area normally associated with physical movement, reacted too, responding to what Dr. Levitin suspected was the brain's predictions of where the song was going to go. As the brain internalizes the tempo, rhythm and emotional peaks of a song, the cerebellum begins reacting every time the song produces tension (that is, subtle deviations from its normal melody or tempo).

"When we saw all this activity going on precisely in sync, in this order, we knew we had the smoking gun," he said. "We've always known that music is good for improving your mood. But this showed precisely how it happens."

In my forthcoming book, Proust Was A Neuroscientist, I discuss The Rite of Spring in the context of this research. How does a symphony cause a riot? Why was Stravinsky's music so magisterially upsetting? The answer, I believe, is that The Rite was so dangerous because it was so new. Nobody had ever written a symphony like this before.

This sort of brazen newness doesn't feel good. The brain is designed to learn by association: if this, then that. Music works by subtly toying with our expected associations, enticing us to make predictions about what note will come next, and then confronting us with our prediction errors. In fact, the brain stem contains a network of neurons that only responds to surprising sounds. When the musical pattern we know is violated, these cells begin the neural process that ends with the release of dopamine. (As Levitin notes, dopamine is also the chemical source of our most intense emotions, which helps to explain the strange emotional power of music, especially when it confronts us with newness and dissonance.) By tempting us with fragile patterns, music taps into our most basic brain circuitry and activates our most primal emotions.

Stravinsky simply took this idea to its logical extreme. His musical violations were more violent than anything that had come before. (Although , to be fair, Wagner and Schoenberg had also caused violent riots.) When he sat down to write The Rite, Stravinsky wanted to dismantled the edifice of classical music. It was time to listen to something completely new.

Of course, The Rite is now a revered work of modernism. While it still sounds aggressive, that aggression is no longer so shocking. Forty years after The Rite caused a riot, Walt Disney included it in Fanatasia, a cartoon for children. How does this happen? How does a violent work of art become suitable for toddlers? The answer returns us to the brain, and its ability to learn new patterns. After a few listens, the painful Augurs chords that begins Stravinsky's symphony ceases to be so painful. Our dopamine neurons have learned how to listen.

More like this

Over time, I have been involved with several different music genres, both as performer and listener. I would be interested in the implications this research has for the appeal of musical genres that have looser theory rules (and, presumably, more "surprise" in terms of development), such as hardcore, or, perhaps more highbrow, the outer reaches of jazz.

I can appreciate that perhaps when Coltrane started to get more and more "out" there, his work was received with similar surprise (and, in part, distaste) as Stravinsky's.

The Rite caused a riot? Don't forget, it is also a ballet. I thought that the "riot" was equally caused by the scantily clad dancers doing a "radical departure from classical ballet... arms and legs were sharply bent...dancers danced more from their pelvis than their feet" (Wikipedia).

One should be careful of mythological musical stories; Stravinsky was a notorious self promoter, and the riot story seems a little too good to be true, no? Music so radical that the French could not stand it! We are still talking about the Rite as a radical work nearly 100 years after the premier thanks in part to the riot story.

Awesome post. Does this also apply to other transgressive musical performers? (Punk rock, early rap, etc.)

Think about the connection between certain genres of music and the movements associated with them. I'd be curious to know if these movements correspond at all to the patterns seen in the cerebellum. For instance, when a metalhead is headbanging to Opeth or In Flames, does it generate fundamentally different patterns of physical movement in the cerebellum than, say, someone crunking out to Missy Elliot or Ludacris? Both movements are similar, an up and down motion at each of the downbeats, it's just that the other patterns that go between those are different. For metalheads, there is only that up-and-down, whereas for a crunk dancer, the up-and-down motion is intercalated with many other movements. So do you think that the patterns would be different?

By Toaster Sunshine (not verified) on 02 Jan 2007 #permalink

noone's got a point. Staged riots weren't that unusual at the time eg Gorlitz, Paderevsky's manager, was said to have given 50 tickets to students on condition that they stampede "as though overcome with a mad desire to get a nearer view of P. performing his magic." [reference in Schonberg's "the Great Pianists"]

What is unusual is how the riot story has stuck to the Rite apparently to the exclusion of other contemporary works. Material for a dissertation there, maybe.

In the more placid 1960's and 70's I was regularly involved in first performances and they were routinely booed. The booing was half-hearted though, as if the audience just felt it was expected of them. Never had anything approaching a riot.

Thanks for all your comments. I agree that The Rite's riot has become a cliche of modernism. Although the riot was absolutely a real, historical event, it has certainly displaced other important modernist events. (Diaghilev, the head of the Ballet Russes, was a great salesman, and knew how to generate PR.) It's also important to note that Nijinsky's choreography was extremely controversial and provocative, and was no doubt part of the reason that the crowd was so agitated. (In the performance, Nijinsky basically inverted classical ballet technique.) That said, I don't think it's a coincidence that the riot started right after the famous Augurs chord began. In 1913, that chord was an unimaginably difficult sound.

Of course, Stravinsky didn't incite the first musical riot. At the time of the riot, newspapers noted that the riot was only the worst since Wagner. And just a few years before, Schoenberg had incensed a few crowds with his atonal experiments.

So new music - especially when it defies our expectations - has a long history of inciting boos, if not riots. (My favorite modern example would be Dylan going electric and getting called a Judas. But you can interpet all sorts of avante-garde events through the prism of the brain. That's what artists do: they stretch our neurons, and force us to confront new sorts of sensations.)

As for the question of whether or not more "open-ended" music is more difficult, I would hypothesize that it is. The mind abhores ambiguity, and always wants its sensory patterns to have neat resolutions. While all music requires some sort of narrative tension - we wait for the tonic chord to arrive, and end the dissonant suspense - I think that the more ambiguous the patterns of music become, the harder it is to learn to listen to it. This is why Wagner was so difficult: he was the Coltrane of his time, and loved musical ambiguity.

The problem with "expectation" theories of music is that there is a very weak correlation between our perception of musicality and whether or not we can predict which note is going to be played next. A new tune can be enjoyable the first time you hear it, or the tenth time, and after maybe the hundredth time in a month you might get a bit sick of it.

We can compare this to humour, where a joke is funny the first time, and not at all the second or third time. Jokes very obviously depend on some difference between what the audience expects and how the joke actually turns out.

Too much of the same music can become boring, but there is a similar "boredom effect" for food and sex. This suggests that the effect of expectations on perception of musicality is very much a secondary effect, and the primary determinant of musicality is independent of whether or not we have experienced that item of music before (just like with food or sex).

If our inability to predict the next note played was critical to our enjoyment of music, then presumably someone playing a tune would experience no enjoyment whatsoever from the sound of their own playing (since if they don't know what is going to happen next then how can they be playing it?).

An alternative account of expectation is that music is pleasurable, and like all sources of pleasure we learn to anticipate it. In the case of music this implies the construction of an internal theory of music in our brains. Such a theory will always be less than perfect, so when we hear good new music, it will be better than our theory predicts, and thus better than expected. And just as with any other kind of pleasure, the best pleasures are the surprising ones. But the surprise itself is not the major cause of the pleasure. A new recipe won't be pleasurable to eat just because it is surprising. It still has to taste nice.

If the violation of expectations is so crucial to enjoyment of music, howcum people still cherish Elvis & Beatles albums?

By Pierce R. Butler (not verified) on 28 Apr 2007 #permalink

As a classical music lover for over 50 years, I think that the predictability, rather than the unpredictability of the music is what makes it so enjoyable. I do not think one can enjoy and release much dopamine the first time one listens to Beethoven's quartets (any of them). It is the repeated playing of a new song by X on the radio that pushes it up the ladder of popularity and joy. Stravinsky's Rite of Spring is the exeption to the rule (unpredictability) that proves the rule that music's joy is its predictability.

Where the brain and music are concerned, let's not forget that memory, and thus, other brain regions (hippocampus) also play an important part in the joy of music. Many of us asociate specific past events in our lives with a specific piece of music. Occasionally, the first time we have heard a piece of music is the memory that connect us to other "historical" events that occurred concomitantly.

I also wonder when neuroscience will be able to explain the "immortality" of classical music compared to the mortality of other genres. Elvis and the Beatles will fade away with the generation that grew up with them. The music of Mozart, Mendelssohn, Schubert, Beethoven, Dvorak, Stravinsky and many more outsurvived them and continues to induce the secretion of untold amounts of dopamine.

By S. Rivlin (not verified) on 29 Apr 2007 #permalink

S. Rivlin: I offered the examples of the King & those Liverpool lads, rather than the classical composers you cite, specifically because those latter will mostly be heard through fresh interpretations in new performances, thus allowing for the possibility of novelty which is proposed here as an attraction of music.

The recordings which are so popular now, otoh, haven't changed by an eighth-note in 40 years.

By Pierce R. Butler (not verified) on 01 May 2007 #permalink