Segment-Level Mandarin Chinese Speech-Based Cognitive Impairment Detection via an Autoencoder with Contrastive Learning

2026-06-18

AI Summary

This study addresses the challenges of scarce labeled data and poor cross-dataset generalization in detecting cognitive impairment from Chinese speech. To this end, the authors propose a segment-level representation learning framework that integrates autoencoding and contrastive learning. The approach segments speech into clips, converts them into spectrograms, and leverages both offline and online data augmentation to jointly optimize reconstruction and discriminative objectives, thereby learning robust and discriminative speech representations. Experiments across four independent Chinese datasets demonstrate that the proposed method significantly enhances model generalization under limited annotation and cross-domain settings, achieving consistently superior performance in both binary classification and the more clinically challenging three-class classification tasks.

Abstract

\noindent\textbf{Background and Objective:} Speech has emerged as a low-cost and non-invasive digital biomarker with considerable potential for cognitive impairment detection. However, limited labeled data and cross-dataset variability remain major challenges for robust speech-based screening systems. \par\noindent\textbf{Methods:} We developed a segment-level representation learning framework for speech-based cognitive impairment detection. Speech recordings were divided into short segments and converted into spectrogram representations. To improve robustness under limited-data conditions, offline and online augmentation strategies were combined with autoencoder-based representation learning and contrastive objectives to enhance discriminative latent representations. \par\noindent\textbf{Results:} Experiments conducted on four independent Mandarin Chinese speech datasets demonstrated stable and competitive performance in both binary and three-class classification tasks, with particularly notable improvements in the clinically challenging three-class setting. Ablation studies further supported the effectiveness of the proposed framework. \par\noindent\textbf{Conclusions:} The findings suggest that segment-level speech representation learning may provide a scalable and practical approach for cognitive impairment screening in resource-constrained clinical settings.

Problem

• cognitive impairment detection
• speech-based biomarker
• limited labeled data
• cross-dataset variability
• Mandarin Chinese speech

Innovation

• segment-level representation
• autoencoder
• contrastive learning
• speech-based biomarker
• data augmentation