Text-to-music generation models capture musical semantic representations in the human brain
Timo I. Denk, Yu Takagi, Takuya Matsuyama, Andrea Agostinelli, Tomoya Nakai, Christian Frank, Shinji Nishimoto · 2025
Read Text-to-music generation models capture musical semantic representations in the human brain by Timo I. Denk, Yu Takagi, Takuya Matsuyama, Andrea Agostinelli, Tomoya Nakai, Christian Frank, Shinji Nishimoto (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/pmc-12770365.txt.
Description
Timo Denk and co-authors show that text-to-music models capture how the human brain represents music in a 2025 study. Aligning brain scans with model features is a growing way to probe how minds encode the world.
Volunteers heard music inside a scanner while the team recorded brain activity. Music was then rebuilt two ways: by retrieval, and by feeding brain data into the MusicLM generator. The rebuilt tracks matched the originals in genre, instruments, and mood. Encoding models found that music-derived and text-derived meaning overlapped heavily around the auditory cortex. Features from the music model predicted primary auditory activity better than general models. Music meaning, it seems, has a shared address in brain and machine.
Catalog record
| CREATOR | Timo I. Denk, Yu Takagi, Takuya Matsuyama, Andrea Agostinelli, Tomoya Nakai, Christian Frank, Shinji Nishimoto |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 71 KB |
| VERSION | 1.0 |
| LICENSE | CC BY 4.0 |
| RIGHTS | CC BY |
| SOURCE | Europe PMC |