The Neuroscience of Improvisation


by Carl Sherman

June 13, 2011

The discussion began wordlessly.

Jazz guitarist Pat Metheny took the stage and, accompanied by bassist Larry Grenadier, delivered a five-minute cascade of notes and chords, slowing, accelerating, changing subtly and abruptly, fresh and unpredictable but somehow connected, snatches of a single melody emerging from and hiding beneath an avalanche of music.

“What’s happening in the brain when Pat does what he just did?” asked moderator John Schaefer, host of a New York radio show about music. He had a panel of four scientists to shed light on what he called “the neuroscience of improvisation.”

The symposium, “Music and the Spark of Spontaneity,” part of the four-day World Science Festival in New York City, ranged widely over the subject: the origins of music itself, its relationship to language, how musical proficiency is learned—but always returned, like Metheny’s extended riff, to its central theme.

Why improvisation? It may represent “the essence of creativity,” one panel member suggested: a sustained flight of the mysterious process that generates art, literature—and great ideas in science. “Jazz is a good model… it’s amenable to experimentation,” panelist Charles Limb said in a telephone interview after the symposium.

Limb, a Dana Foundation grantee who teaches at both Johns Hopkins Medical School and the Peabody Conservatory of Music, has used fMRI to look into the brain of the improviser.

“Music affects the brain globally,” he said at the symposium. “We were trying to figure out what part of the brain is involved specifically when you improvise.” In his study, he had several highly proficient jazz musicians play a keyboard adapted for use in the scanner, and compared brain activity while they improvised and played a memorized passage.

One distinctive feature of the improvising brain was a pattern of dissociation in the prefrontal cortex: increased activity in the medial part of this brain region, and reduced activity in lateral areas.

In a paper reporting these findings, Limb and co-author Allen Braun noted that the medial prefrontal cortex has been associated with “internally motivated behavior” and “autobiographical narrative,” while lateral areas are thought to be involved in monitoring, judging, and correcting behavior. Their deactivation, they speculated, might suggest suspension of conscious control, enabling freer, more spontaneous thoughts and actions.

“You have to take such interpretations with a large caveat,” Limb said. “I get uncomfortable when people try to make this seem clear and tidy; it’s not.”

But the interpretation seems plausible, Limb granted, in light of musicians’ descriptions of the experience of improvising. “For an experienced jazz guy, there are a whole set of musical options—you could go to F#, you could switch to Dorian mode.” Metheny said during the event. “But I’m not thinking about any of them… I remain somewhat detached. It’s as if I’m listening.”

 Gary Marcus, a professor of psychology at New York University, explained that from a cognitive scientist’s point of view, “the brain is capable of two kinds of fundamental processes, controlled and conscious, and automatic and unconscious. Conscious processes are slow, one thing at a time. You can’t play music that way. It takes years of practice to make what might originally be conscious become unconscious.

“You have to acquire incredible skill in music to be both fast and accurate—to play off each other as Pat and Larry do. But once it’s there, it’s effortless, and it doesn’t have to be conscious.”

 Cognitive neuroscientist Jamshed Bharucha, of Tufts University, spoke similarly about creative spontaneity. “You can’t improvise if you haven’t been engaged in learning, practice, guidance, discipline—a lot of hard work. There’s a myth about creativity as magic. As a scientist, I don’t believe in magic, but in years and years of mastering what’s already there, so you can go beyond it.”

 For Metheny, it was “a 10-year process of becoming familiar with the language to get to the point that you’re not messing around with the nuts and bolts—you’re free to just talk.”

The musician also described feeling out of touch, in a sense, while improvising. “I try to be considerate of the audience—to get to the gig early, rehearse, be ready to do what we do. But from the first note, I play for myself…if I started worrying about what critics say, or a friend says, or what the audience is thinking, I’d be paralyzed.”

This, too, appears to correspond to a brain-activity pattern, said neuroscientist-musicologist Aaron Berkowitz of Harvard University. In his fMRI study, trained musicians and non-musicians were scanned while they played simple improvisations. Contrary to what researchers expected, non-musicians produced as many variations as musicians, and brain regions associated with music were activated equally in both groups.

“The only difference we saw was that for some kinds of improvisation, musicians were turning off an area of the brain, and non-musicians were not: the right temporoparietal junction, which is…typically deactivated in situations of goal-directed behavior” to inhibit distraction by irrelevant stimuli that might impair performance.

“This is exactly what Pat is expressing,” Berkowitz said. “I could imagine that when he came in and started to play, he shut everything down and was one with the instrument.”

The discussion broadened to music’s relationship to language. “Both are infinite combinatorial systems,” said Marcus. “A small number of elements, phonemes, and musical notes that can recombine in a lot of different ways.”

Which came first? It’s a contentious debate, participants agreed, and Berkowitz cited theories “that suggest a common predecessor, in animal calls.” One species of monkey has three alarm calls that refer to danger from different directions, he said. “Sounds are more to express emotion; words often refer to things that aren’t as emotionally changed.” And some forms of speech “fall between music and language, like poetry and a preacher’s sermon.”

Charles Limb has been investigating just such a hybrid: rap. “It’s what kids are doing spontaneously when growing up... and improvisation is a strong theme. It incorporates language and rhythmic music very equally.”

Limb has been scanning the brains of rappers the same way he looked at jazz musicians: comparing fMRIs when they recited memorized passages to when they “freestyled,” or improvised in rhyme. Although the study is still in progress, preliminary data suggest “major changes in brain activity when you go from memorized rap to freestyle.”

Can studies of improvisation unlock more general secrets of creativity? Limb hopes to do similar investigations of artists as they draw or paint.

 The moderator ended with an inevitable question about art and science: “It is worth the effort to measure and quantify something abstract and artistic… to demystify what we enjoy the mystery of?”

Limb saw nothing “threatening or reductionist” in the work of neuroscientists. “Humans are hardwired to seek art, and there are very few things that engage the brain on the level that music does. To understand the neural basis of creativity teaches us something fundamental about who we are, why we’re here.”

Improvisation “shows us what the mind can do,” Marcus added. “The ability of human beings to improvise tells us a lot about the ultimate scope of our capabilities.”