From Black Box to Biomarker: Sparse Autoencoders for Interpreting Speech Models of Parkinson's Disease
Fuente:
arXiv
Enregistré dans:
| Auteurs principaux: | , , , , |
|---|---|
| Format: | Preprint |
| Publié: |
2025
|
| Sujets: | |
| Accès en ligne: | |
| Tags: |
Ajouter un tag
Pas de tags, Soyez le premier à ajouter un tag!
|
| _version_ | 1866916858142130176 |
|---|---|
| author | Plantinga, Peter Chen, Jen-Kai Sattari, Roozbeh Ravanelli, Mirco Klein, Denise |
| author_facet | Plantinga, Peter Chen, Jen-Kai Sattari, Roozbeh Ravanelli, Mirco Klein, Denise |
| contents | Speech holds promise as a cost-effective and non-invasive biomarker for neurological conditions such as Parkinson's disease (PD). While deep learning systems trained on raw audio can find subtle signals not available from hand-crafted features, their black-box nature hinders clinical adoption. To address this, we apply sparse autoencoders (SAEs) to uncover interpretable internal representations from a speech-based PD detection system. We introduce a novel mask-based activation for adapting SAEs to small biomedical datasets, creating sparse disentangled dictionary representations. These dictionary entries are found to have strong associations with characteristic articulatory deficits in PD speech, such as reduced spectral flux and increased spectral flatness in the low-energy regions highlighted by the model attention. We further show that the spectral flux is related to volumetric measurements of the putamen from MRI scans, demonstrating the potential of SAEs to reveal clinically relevant biomarkers for disease monitoring and diagnosis. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2507_16836 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | From Black Box to Biomarker: Sparse Autoencoders for Interpreting Speech Models of Parkinson's Disease Plantinga, Peter Chen, Jen-Kai Sattari, Roozbeh Ravanelli, Mirco Klein, Denise Audio and Speech Processing Machine Learning Speech holds promise as a cost-effective and non-invasive biomarker for neurological conditions such as Parkinson's disease (PD). While deep learning systems trained on raw audio can find subtle signals not available from hand-crafted features, their black-box nature hinders clinical adoption. To address this, we apply sparse autoencoders (SAEs) to uncover interpretable internal representations from a speech-based PD detection system. We introduce a novel mask-based activation for adapting SAEs to small biomedical datasets, creating sparse disentangled dictionary representations. These dictionary entries are found to have strong associations with characteristic articulatory deficits in PD speech, such as reduced spectral flux and increased spectral flatness in the low-energy regions highlighted by the model attention. We further show that the spectral flux is related to volumetric measurements of the putamen from MRI scans, demonstrating the potential of SAEs to reveal clinically relevant biomarkers for disease monitoring and diagnosis. |
| title | From Black Box to Biomarker: Sparse Autoencoders for Interpreting Speech Models of Parkinson's Disease |
| topic | Audio and Speech Processing Machine Learning |
| url | https://arxiv.org/abs/2507.16836 |