Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Bias and Fairness in Self-Supervised Acoustic Representations for Cognitive Impairment Detection

Kashaf Gulzar, Korbinian Riedhammer, Elmar Nöth, Andreas K. Maier, Paula Andrea Pérez-Toro

TL;DR

A systematic bias analysis of acoustic-based CI and depression classification using the DementiaBank Pitt Corpus highlights the need for fairness-aware model evaluation and subgroup-specific analysis in clinical speech applications, particularly in light of demographic and clinical heterogeneity in real-world applications.

Abstract

Speech-based detection of cognitive impairment (CI) offers a promising non-invasive approach for early diagnosis, yet performance disparities across demographic and clinical subgroups remain underexplored, raising concerns around fairness and generalizability. This study presents a systematic bias analysis of acoustic-based CI and depression classification using the DementiaBank Pitt Corpus. We compare traditional acoustic features (MFCCs, eGeMAPS) with contextualized speech embeddings from Wav2Vec 2.0 (W2V2), and evaluate classification performance across gender, age, and depression-status subgroups. For CI detection, higher-layer W2V2 embeddings outperform baseline features (UAR up to 80.6\%), but exhibit performance disparities; specifically, females and younger participants demonstrate lower discriminative power (\(AUC\): 0.769 and 0.746, respectively) and substantial specificity disparities (\(Δ_{spec}\) up to 18\% and 15\%, respectively), leading to a higher risk of misclassifications than their counterparts. These disparities reflect representational biases, defined as systematic differences in model performance across demographic or clinical subgroups. Depression detection within CI subjects yields lower overall performance, with mild improvements from low and mid-level W2V2 layers. Cross-task generalization between CI and depression classification is limited, indicating that each task depends on distinct representations. These findings emphasize the need for fairness-aware model evaluation and subgroup-specific analysis in clinical speech applications, particularly in light of demographic and clinical heterogeneity in real-world applications.

Bias and Fairness in Self-Supervised Acoustic Representations for Cognitive Impairment Detection

TL;DR

A systematic bias analysis of acoustic-based CI and depression classification using the DementiaBank Pitt Corpus highlights the need for fairness-aware model evaluation and subgroup-specific analysis in clinical speech applications, particularly in light of demographic and clinical heterogeneity in real-world applications.

Abstract

Speech-based detection of cognitive impairment (CI) offers a promising non-invasive approach for early diagnosis, yet performance disparities across demographic and clinical subgroups remain underexplored, raising concerns around fairness and generalizability. This study presents a systematic bias analysis of acoustic-based CI and depression classification using the DementiaBank Pitt Corpus. We compare traditional acoustic features (MFCCs, eGeMAPS) with contextualized speech embeddings from Wav2Vec 2.0 (W2V2), and evaluate classification performance across gender, age, and depression-status subgroups. For CI detection, higher-layer W2V2 embeddings outperform baseline features (UAR up to 80.6\%), but exhibit performance disparities; specifically, females and younger participants demonstrate lower discriminative power (: 0.769 and 0.746, respectively) and substantial specificity disparities ( up to 18\% and 15\%, respectively), leading to a higher risk of misclassifications than their counterparts. These disparities reflect representational biases, defined as systematic differences in model performance across demographic or clinical subgroups. Depression detection within CI subjects yields lower overall performance, with mild improvements from low and mid-level W2V2 layers. Cross-task generalization between CI and depression classification is limited, indicating that each task depends on distinct representations. These findings emphasize the need for fairness-aware model evaluation and subgroup-specific analysis in clinical speech applications, particularly in light of demographic and clinical heterogeneity in real-world applications.
Paper Structure (18 sections, 6 equations, 4 figures, 6 tables)

This paper contains 18 sections, 6 equations, 4 figures, 6 tables.

Table of Contents

  1. Introduction
  2. Related Work
  3. Contribution of this Study
  4. Data
  5. Methods
  6. Acoustic representations
  7. Mel Frequency Cepstral Coefficients
  8. Extended Geneva Minimalistic Acoustic Parameter Set
  9. Wav2Vec 2.0
  10. Optimization and Classification
  11. Fairness Analysis
  12. Experiments and results
  13. Feature Performance
  14. CI vs. NCI
  15. D-CI vs. ND-CI
  16. ...and 3 more sections

Figures (4)

  • Figure 1: Subject distribution based on CI and gender.
  • Figure 2: Workflow diagram illustrating the key steps in the data analysis.
  • Figure 3: Comparison of classifier performance across W2V2 embedding layers in CI vs. NCI and D-CI vs. ND-CI classifications for imbalanced and CI-balanced datasets. Best results are highlighted for each classifier.
  • Figure 4: SVM decision score distributions across demographic and clinical subgroups for Wav2Vec 2.0 hidden layer 9 using CI-gender balanced dataset. Histograms and Kernel Density Estimate (KDE) curves illustrate the separation between NCI and CI classes. The dashed line at 0 represents the SVM decision threshold. AUC values are provided for each subgroup to indicate classification performance.