To Tug Hearts, Music First Must Tickle the Neurons

 

“The song has that triplet going on underneath that pushes it along, and at a certain point I wanted it to stop because the story suddenly turns very serious,” Mr. Simon said in an interview.

“The stopping of sounds and rhythms,” he added, “it’s really important, because, you know, how can I miss you unless you’re gone? If you just keep the thing going like a loop, eventually it loses its power.”

An insight like this may seem purely subjective, far removed from anything a scientist could measure. But now some scientists are aiming to do just that, trying to understand and quantify what makes music expressive — what specific aspects make one version of, say, a Beethoven sonata convey more emotion than another.

The results are contributing to a greater understanding of how the brain works and of the importance of music in human development, communication and cognition, and even as a potential therapeutic tool.

Research is showing, for example, that our brains understand music not only as emotional diversion, but also as a form of motion and activity. The same areas of the brain that activate when we swing a golf club or sign our name also engage when we hear expressive moments in music. Brain regions associated with empathy are activated, too, even for listeners who are not musicians.

And what really communicates emotion may not be melody or rhythm, but moments when musicians make subtle changes to the those musical patterns.

Daniel J. Levitin, director of the laboratory for music perception, cognition and expertise at McGill University in Montreal, began puzzling over musical expression in 2002, after hearing a live performance of one of his favorite pieces, Mozart’s Piano Concerto No. 27.

“It just left me flat,” Dr. Levitin, who wrote the best seller “This Is Your Brain on Music” (Dutton, 2006), recalled in a video describing the project. “I thought, well, how can that be? It’s got this beautiful set of notes. The composer wrote this beautiful piece. What is the pianist doing to mess this up?”

Before entering academia, Dr. Levitin worked in the recording industry, producing, engineering or consulting for Steely Dan, Blue ?yster Cult, the Grateful Dead, Santana, Eric Clapton and Stevie Wonder. He has played tenor saxophone with Mel Tormé and Sting, and guitar with David Byrne. (He also performs around campus with a group called Diminished Faculties.)

After the Mozart mishap, Dr. Levitin and a graduate student, Anjali Bhatara, decided to try teasing apart some elements of musical expression in a rigorous scientific way.

He likened it to tasting two different pots de crème: “One has allspice and ginger and the other has vanilla. You know they taste different but you can’t isolate the ingredient.”

To decipher the contribution of different musical flavorings, they had Thomas Plaunt, chairman of McGill’s piano department, perform snatches of several Chopin nocturnes on a Disklavier, a piano with sensors under each key recording how long he held each note and how hard he struck each key (a measure of how loud each note sounded). The note-by-note data was useful because musicians rarely perform exactly the way the music is written on the page — rather, they add interpretation and personality to a piece by lingering on some notes and quickly releasing others, playing some louder, others softer.

The pianist’s recording became a blueprint, what researchers considered to be the 100 percent musical rendition. Then they started tinkering. A computer calculated the average loudness and length of each note Professor Plaunt played. The researchers created a version using those average values so that the music sounded homogeneous and evenly paced, with every eighth note held for an identical amount of time, each quarter note precisely double the length of an eighth note.

They created other versions too: a 50 percent version, with note lengths and volume halfway between the mechanical average and the original, and versions at 25 percent, 75 percent, and even 125 percent and 150 percent, in which the pianist’s loud notes were even louder, his longest-held notes even longer.

Study subjects listened to them in random order, rating how emotional each sounded. Musicians and nonmusicians alike found the original pianist’s performance most emotional and the averaged version least emotional.

But it was not just changes in volume and timing that moved them. Versions with even more variation than the original, at 125 percent and 150 percent, did not strike listeners as more emotional.

“I think it means that the pianist is very experienced in using these expressive cues,” said Dr. Bhatara, now a postdoctoral researcher at the Université Paris Descartes. “He’s using them at kind of an optimal level.”

And random versions with volume and note-length changes arbitrarily sprinkled throughout made almost no impression.

All of this makes perfect sense to Paul Simon.