Scientists Now Trying to Predict Radio Hits


In 2006, Emory University scientist Gregory Berns conducted a study on peer pressure by getting a group of teenagers to listen to brief clips of then-unknown songs and rate their likability while undergoing brain scans. Years later, it turned out that "Teenage ratings of likeability had little predictive value ... but their brains were effectively tuned to commercial success ... the researchers usually could tell from the strength of brain activity alone which songs would later sell at least 20,000 copies." Berns: "The punch line is that brain responses correlated with units sold ... It is far from being a hit predictor, but it was statistically significant." This is of course an early stage of the research, but eventually we may be able to immediately tell, off scanning teenagers' brains alone, which singles are destined for greatness and which ones will be duly discarded. Finally, science is useful! [WSJ]