Brain2Music: Reconstructing Music from Human Brain Activity
This addresses the problem of decoding complex auditory experiences from brain data for neuroscience and AI applications, though it is incremental as it builds on existing music generation models.
The paper tackles reconstructing music from human brain activity captured via fMRI, using music retrieval or the MusicLM model conditioned on fMRI embeddings, resulting in generated music that resembles the original stimuli in semantic properties like genre and mood.
The process of reconstructing experiences from human brain activity offers a unique lens into how the brain interprets and represents the world. In this paper, we introduce a method for reconstructing music from brain activity, captured using functional magnetic resonance imaging (fMRI). Our approach uses either music retrieval or the MusicLM music generation model conditioned on embeddings derived from fMRI data. The generated music resembles the musical stimuli that human subjects experienced, with respect to semantic properties like genre, instrumentation, and mood. We investigate the relationship between different components of MusicLM and brain activity through a voxel-wise encoding modeling analysis. Furthermore, we discuss which brain regions represent information derived from purely textual descriptions of music stimuli. We provide supplementary material including examples of the reconstructed music at https://google-research.github.io/seanet/brain2music