Music reconstructed from brain signals: researchers have reconstructed a Pink Floyd song

In recent years, we have seen significant advances in technology that can decipher brain signals and interpret neural activity. However, a new study conducted by American scientists takes this technology to the next level. Researchers were able to reconstruct a “recognizable version” of a Pink Floyd song based on pulses of activity passing through a specific part of the temporal lobe of the brain in volunteers as they listened to the hit Another Brick in the Wall Part 1.

For this study, the team of scientists recruited 29 people who had already been implanted with brain electrodes to treat epilepsy. All participants had 2,668 electrodes checked for neural patterns while they listened to the Pink Floyd song. All of this data was then analyzed using machine learning and a regression-based decoding model. This allowed the researchers to flip the system and identify “Another Brick in the Wall” by the way the brain reacted to it. The recovered track is a bit muddy and distorted, but it’s not hard to recognize what the song is.

This research opens up new possibilities in deciphering brain patterns and developing brain-machine interfaces. Restoring music perception could be useful for people with brain damage or those who have lost the gift of speech. They will be able to think about the words they want to say, as well as the pitch, tone and lyrical flow of those words.

“For example, the results of music perception may contribute to the development of a general auditory decoder that incorporates prosodic elements of speech based on a relatively small number of well-placed electrodes,” the researchers wrote in their published paper.

This discovery has great potential for the development of new technologies and methods in brain research. It also emphasizes the importance of understanding the neural basis of music perception and its future applications.

Notify of

Inline Feedbacks
View all comments
Would love your thoughts, please comment.x