“That the neural activity of 33 people can predict if millions of others listened to new songs is quite amazing. Nothing close to this accuracy has ever been shown before.”

Have you ever read Isaac Asimov’s short story “Franchise“? The short sci-fi story tells the story that in the future, the United States has converted to an “electronic democracy” where the computer Multivac selects a single person to answer a number of questions. Multivac will then use the answers and other data to determine what the results of an election would be, avoiding the need for an actual election to be held. The story centers around the very average and reluctant Norman Muller of Bloomington, Indiana, who was chosen as the “Voter of the Year” in the 2008 U.S. presidential election and ends with the ironic statement that the US public has “exercised once again their free, untrammeled franchise (voting)”.

Sounds far fetched? Think again.

Researchers in the US have used a comprehensive machine learning technique applied to brain responses and were able to predict hit songs with 97% accuracy. Not only that, but they can do it with as few as testing the song on 33 people.

“By applying machine learning to neurophysiologic data, we could almost perfectly identify hit songs,” said Paul Zak, a professor at Claremont Graduate University and senior author of the study published in Frontiers in Artificial Intelligence. “That the neural activity of 33 people can predict if millions of others listened to new songs is quite amazing. Nothing close to this accuracy has ever been shown before.”

This ability to determine things before they hit markets of scale is being called “Neuroforcasting”.

In the experiment, they equipped participants with off-the-shelf sensors who listened to a set of 24 songs and were asked about their preferences and some demographic data. Researchers used different statistical approaches to assess the predictive accuracy of neurophysiological variables. Linear statistical model identified hit songs at a success rate of 69%. When they applied machine learning to the data they collected, the rate of correctly identified hit songs jumped to 97%. They also applied machine learning to the neural responses to the first minute of the songs. In this case, hits were correctly identified with a success rate of 82%.

This ability to predict hits is called ‘neuroforecasting’ and could even be used instead of recommender systems like you see in Netflix and Pandora.

“If in the future wearable neuroscience technologies, like the ones we used for this study, become commonplace, the right entertainment could be sent to audiences based on their neurophysiology. Instead of being offered hundreds of choices, they might be given just two or three, making it easier and faster for them to choose music that they will enjoy,”