Fine-tuning speech models on fMRI data shifts semantic information to late layers and aligns them with semantic brain regions, while early layers preserve acoustic processing.
These analyses investigate the hierarchy of in- formation processing in brain-tuned vs
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
citation-role summary
background 1
citation-polarity summary
fields
cs.CL 1years
2025 1verdicts
CONDITIONAL 1roles
background 1polarities
background 1representative citing papers
citing papers explorer
-
Brain-tuned Speech Models Better Reflect Speech Processing Stages in the Brain
Fine-tuning speech models on fMRI data shifts semantic information to late layers and aligns them with semantic brain regions, while early layers preserve acoustic processing.