Evaluating the Usefulness of Non-Diagnostic Speech Data for Developing Parkinson's Disease Classifiers

2025-05-24

AI Summary

This study investigates the feasibility of using non-diagnostic, natural conversational speech—specifically turn-taking (TT) interactions—for automated Parkinson’s disease (PD) detection, challenging the prevailing paradigm reliant on diagnostic speech tasks (e.g., sustained vowels or read speech). Method: We employ acoustic feature extraction, binary classification modeling, and cross-dataset transfer evaluation. To mitigate biases, we apply resampling and data augmentation to balance gender and diagnostic status distributions. Contribution/Results: Our systematic evaluation demonstrates that models trained solely on TT data achieve PD classification performance comparable to those trained on the canonical diagnostic dataset PC-GITA. Moreover, TT-trained models exhibit superior generalization to PC-GITA—revealing an asymmetry in cross-dataset generalizability. Variance analysis identifies inter-subject variability—not dataset or model architecture—as the dominant source of cross-validation instability. These findings provide methodological grounding and empirical evidence for unobtrusive, everyday-speech-based PD screening.

Technology Category

• Natural Language Processing: Conversational AI/Dialog Systems
• Data Mining & Knowledge Management: Conversational Systems for Recommendation & Retrieval
• Machine Learning: Multimodal Learning

Application Category

• Web Mining and Content Analysis: Large pretrained models with web data
• Economics, Online Markets and Human Computation: Data quality aspects of human-annotated datasets
• User Modeling, Personalization and Recommendation: User modeling and simulation for interactive and conversational systems

Abstract

Speech-based Parkinson's disease (PD) detection has gained attention for its automated, cost-effective, and non-intrusive nature. As research studies usually rely on data from diagnostic-oriented speech tasks, this work explores the feasibility of diagnosing PD on the basis of speech data not originally intended for diagnostic purposes, using the Turn-Taking (TT) dataset. Our findings indicate that TT can be as useful as diagnostic-oriented PD datasets like PC-GITA. We also investigate which specific dataset characteristics impact PD classification performance. The results show that concatenating audio recordings and balancing participants' gender and status distributions can be beneficial. Cross-dataset evaluation reveals that models trained on PC-GITA generalize poorly to TT, whereas models trained on TT perform better on PC-GITA. Furthermore, we provide insights into the high variability across folds, which is mainly due to large differences in individual speaker performance.

Problem

• Exploring non-diagnostic speech data for Parkinson's disease detection
• Comparing Turn-Taking dataset with diagnostic datasets like PC-GITA
• Investigating dataset characteristics affecting PD classification performance

Innovation

• Using non-diagnostic speech data for PD detection
• Concatenating audio recordings improves classification
• Cross-dataset evaluation reveals generalization differences