What makes a hit record? Dr. Gregory Berns thinks that the teenage brain may hold an answer to that question. This summer, Berns, the distinguished professor of neuroeconomics and director of the Center for Neuropolicy at Emory, released results from a study he began six years ago. As a part of the research, he had teens listen to a variety of songs culled from MySpace to determine from brain imaging what they liked. Turns out, the research proved to be something of a hit record predictor. “Apologize,” by One Republic, became a huge hit after Berns had completed his research. Here, he talks about music, the mind and “American Idol.”
Q: What made you decide to do the research?
A: The original goal was to understand teenaged decision-making, specifically how peer pressure affects that. We were interested to look at teenagers because conventional wisdom would say that teenagers are more susceptible to peer pressure than adults. Music was a pretty good thing to use for the study because everyone likes music and has strong preferences about it. The idea was to have the kids go in the MRI scanner and play clips of music through headphones and see how that affected their brain activity and their behavior.
Q: You found the songs on MySpace, specifically a song by the group One Republic. That's not an obscure group. So how did you ensure that the kids weren't familiar with the songs you selected?
A: The reason we went to MySpace is because we couldn't use songs that they had already heard on the radio, and we didn't want them to immediately recognize them. So we looked at six different genres to sample all tastes of music — alternative or indie, country, hip-hop/rap, jazz/blues, metal and rock. We picked 20 songs in those categories and checked to make sure the groups weren't signed to a label and, in some cases, older artists who were not likely to be known to teenagers. When we started the study, One Republic was not signed. When we played the songs for the kids in the scanner, they rated the songs on two things. One, "Do you like the song?" And two, "How familiar is it to you?"
Q: What did their brains do to indicate a hit?
A: There are two ways to measure whether someone likes something. The conventional way is to ask someone whether they like it. And we did that, asking them on a scale of one to five how much they liked it. Then we had the brain data, where we were specifically looking in parts of the brain that are known to respond to rewarding things, like food or money, called the reward centers. Most of the time, what a person says they like and what their brain says they like are in agreement, but not always.
Q: Did the other winning songs have some sort of a standard formula respective to their genre? You could say that a lot of hits these days have the same beat track or otherwise sound a lot alike.
A: Listening to the radio, you would certainly have that impression. We had exactly that same question so we did a lot of analysis of the songs themselves. But the short answer is no. There are a couple of companies out there that claim they can take a demo from an artist and run it through their algorithms and say it's like such-and-such a hit. We tried many of those algorithms to see if there was anything in common between the songs that sold a lot or elicited the most brain activity, and we couldn't find any pattern.
Q: When did you discover you were actually onto something?
A: When I heard "Apologize" by One Republic on "American Idol." I was watching with my kids, half listening, and Kris Allen sung that song, and I recognized it and said, "Man, we used that in our study years ago." I don't listen to Top 40 radio so I wasn't aware that the song had become a huge hit. So I went on the Internet and found that One Republic got signed after we were done with collecting our data. And I realized we had this very valuable data set of brain response of a bunch of teenagers to a group of songs before anything really happened to those bands and the artists. So that immediately raised the question, "Could we have predicted that song becoming a hit?" Which caused us to go back and look at the brain data in a different light, then go and pull data on the songs themselves and how many of the songs were sold. We found that there was a very modest link between the brain data and the sales.
Q: So what's your larger goal here?
A: What we're really interested in is figuring out how many of your likes and dislikes are really your own. Society influences what you like and think. So the big picture we're going after is what is it that makes ideas popular?
The Sunday conversation is edited for length and clarity.
About the Author
