Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Adapting Self-Supervised Speech Representations for Cross-lingual Dysarthria Detection in Parkinson's Disease

Abner Hernandez, Eunjung Yeo, Kwanghee Choi, Chin-Jou Li, Zhengjun Yue, Rohan Kumar Das, Jan Rusz, Mathew Magimai Doss, Juan Rafael Orozco-Arroyave, Tomás Arias-Vergara, Andreas Maier, Elmar Nöth, David R. Mortensen, David Harwath, Paula Andrea Perez-Toro

Abstract

The limited availability of dysarthric speech data makes cross-lingual detection an important but challenging problem. A key difficulty is that speech representations often encode language-dependent structure that can confound dysarthria detection. We propose a representation-level language shift (LS) that aligns source-language self-supervised speech representations with the target-language distribution using centroid-based vector adaptation estimated from healthy-control speech. We evaluate the approach on oral DDK recordings from Parkinson's disease speech datasets in Czech, German, and Spanish under both cross-lingual and multilingual settings. LS substantially improves sensitivity and F1 in cross-lingual settings, while yielding smaller but consistent gains in multilingual settings. Representation analysis further shows that LS reduces language identity in the embedding space, supporting the interpretation that LS removes language-dependent structure.

Adapting Self-Supervised Speech Representations for Cross-lingual Dysarthria Detection in Parkinson's Disease

Abstract

The limited availability of dysarthric speech data makes cross-lingual detection an important but challenging problem. A key difficulty is that speech representations often encode language-dependent structure that can confound dysarthria detection. We propose a representation-level language shift (LS) that aligns source-language self-supervised speech representations with the target-language distribution using centroid-based vector adaptation estimated from healthy-control speech. We evaluate the approach on oral DDK recordings from Parkinson's disease speech datasets in Czech, German, and Spanish under both cross-lingual and multilingual settings. LS substantially improves sensitivity and F1 in cross-lingual settings, while yielding smaller but consistent gains in multilingual settings. Representation analysis further shows that LS reduces language identity in the embedding space, supporting the interpretation that LS removes language-dependent structure.
Paper Structure (11 sections, 2 equations, 1 figure, 3 tables)

This paper contains 11 sections, 2 equations, 1 figure, 3 tables.

Table of Contents

  1. Datasets
  2. Method
  3. Self-supervised Representations
  4. Centroid-Based Language Shift
  5. Experiments
  6. Experimental settings
  7. Cross-Lingual Transfer with Language Shift
  8. Multilingual Transfer with Language Shift
  9. Discussion
  10. Conclusion
  11. Generative AI Use Disclosure

Figures (1)

  • Figure 2: UMAP visualization of HuBERT PD speaker representations with and without shifting. Circles denote raw source representations and triangles show representations shifted to the target-language (LS$\rightarrow$TGT).