Evaluating Parkinson's Disease Detection in Anonymized Speech: A Performance and Acoustic Analysis

TL;DR

This study evaluates how different speech anonymization techniques impact the ability to detect Parkinson's disease, balancing privacy preservation with diagnostic accuracy using acoustic and performance analyses.

Contribution

It compares two anonymization methods, revealing that kNN-VC maintains crucial PD-related acoustic features better than STT-TTS, enabling effective privacy-preserving PD detection.

Findings

kNN-VC preserves macro-prosodic features with minimal detection score reduction

STT-TTS offers better privacy but significantly impairs PD detection

Acoustic analysis identifies weaknesses in kNN-VC for future anonymizer design

Abstract

Automatic detection of Parkinson's disease (PD) from speech is a promising non-invasive diagnostic tool, but it raises significant privacy concerns. Speaker anonymization mitigates these risks, but it may suppress the pathological information necessary for PD detection. We assess the trade-off between privacy and PD detection for two anonymizers (STT-TTS and kNN-VC) using two Spanish datasets. STT-TTS provides better privacy but severely degrades PD detection by eradicating prosodic information. kNN-VC preserves macro-prosodic features such as duration and F0 contours, achieving F1 scores only 3-7\% lower than original baselines, demonstrating that privacy-preserving PD detection is viable when using appropriate anonymization. Finally, an acoustic distortion analysis characterizes specific weaknesses in kNN-VC, offering insights for designing anonymizers that better preserve PD information.