Decoding Music from Brain Activity: Exploring the Neural Correlates of Music Perception
Matteo Ferrante*, Matteo Ciferri*, Nicola Toschi
fMRI experiment
5 participants listen to 540 songs while brain activity is recorded with fMRI
CLAP
model
An encoding model of brain activity was built to predict audio responsive regions from audio features extracted with CLAP model. The responsive regions were further be used as inputs for decoding models to decode music from brain activity.
Encoding Pipeline
Music Brain
Decoding model
Brain activity is decoded with a retrieval system that outputs musical genre and a candidate song
Decoding Pipeline
Abstract
This study investigates the relationship between music and brain activity patterns, aiming to bridge the gap between music perception and its neural representation. We leverage the GTZen music fMRI dataset, encompassing 5 subjects who listened to 540 tracks (15s each) from 10 genres while undergoing 3T fMRI scans (TR = 1.5s). Despite the limitations of fMRI's temporal resolution, music elicits robust brain responses.
To explore this concept, we constructed a decoding pipeline capable of retrieving musical information from brain activity. Preprocessing was conducted using fMRIprep. Subsequently, an encoding model was built using CLAP to transform audio into feature representations. This was followed by a Ridge regression to predict brain activity (averaged across 15s listening blocks). This model served to identify brain regions responsive to music.
Voxel values from these responsive regions were then used in three experiments:
These findings demonstrate the feasibility of decoding musical information from brain activity patterns. While further research is needed to refine the decoding process, this approach holds promise for advancing our understanding of music perception and its potential applications in music therapy and other domains.
Jazz Stimulus
Decoded
Metal Stimulus
Decoded
Disco Stimulus
Decoded
Pop Stimulus
Decoded