Enhancing and Exploring Mild Cognitive Impairment Detection with W2V-BERT-2.0

TL;DR

This paper introduces W2V-BERT-2.0 for multilingual MCI detection, addressing transcription limitations by directly using speech features, and proposes visualization and inference methods that improve classification accuracy.

Contribution

The study presents a novel speech feature extraction approach with W2V-BERT-2.0 and introduces visualization and inference techniques tailored for MCI detection, enhancing baseline performance.

Findings

W2V-BERT-2.0 achieves competitive results in MCI detection.

The proposed inference logic significantly improves classification accuracy.

Analysis reveals speaker bias and data split sensitivity issues.

Abstract

This study explores a multi-lingual audio self-supervised learning model for detecting mild cognitive impairment (MCI) using the TAUKADIAL cross-lingual dataset. While speech transcription-based detection with BERT models is effective, limitations exist due to a lack of transcriptions and temporal information. To address these issues, the study utilizes features directly from speech utterances with W2V-BERT-2.0. We propose a visualization method to detect essential layers of the model for MCI classification and design a specific inference logic considering the characteristics of MCI. The experiment shows competitive results, and the proposed inference logic significantly contributes to the improvements from the baseline. We also conduct detailed analysis which reveals the challenges related to speaker bias in the features and the sensitivity of MCI classification accuracy to the data split, providing valuable insights for future research.