Communications of the ACM
New Algorithm Allows for Potential 'Brain-Reading'
A new machine learning algorithm developed at the D'Or Institute for Research and Education in Brazil can identify pieces of music based on scans of the listener's brain.
Credit: Thomas Terceira
Researchers at the D'Or Institute for Research and Education in Brazil have developed a machine learning algorithm capable of identifying pieces of music from functional magnetic resonance imaging (fMRI) scans of the listener's brain.
The team first mapped brain responses triggered by listening to the music, and then used the collected information to identify novel musical pieces based on fMRI imaging data alone.
FMRI visualizes cerebral blood flow and neuronal activation, because when an area of the brain is in use, blood flow to that region increases.
The implication of the research is that by interpreting the right mapping of musical features to the brain, scientists can predict and decode any unique musical piece.
The researchers say the model was based on analyzing six participants who listened to 40 distinct pieces of music. Through this method, the algorithm encoded the listeners' fMRI responses for individual pieces of music, evaluating specific features such as tonality, dynamics, rhythm, and timbre.
From Digital Journal
View Full Article
Abstracts Copyright © 2018 Information Inc., Bethesda, Maryland, USA
No entries found