With the emergence of AI techniques for depression diagnosis, the conflict between high demand and limited supply for depression screening has been significantly alleviated. Among various modal data, audio-based depression diagnosis has received increasing attention from both academia and industry since audio is the most common carrier of emotion transmission. Unfortunately, audio data also contains User-sensitive Identity Information (ID), which is extremely vulnerable and may be maliciously used during the smart diagnosis process. Among previous methods, the clarification between depression features and sensitive features has always serve as a barrier. It is also critical to the problem for introducing a safe encryption methodology that only encrypts the sensitive features and a powerful classifier that can correctly diagnose the depression. To track these challenges, by leveraging adversarial loss-based Subspace Decomposition, we propose a first practical framework \name presented for Trustable Audio Affective Computing, to perform automated depression detection through audio within a trustable environment. The key enablers of TAAC are Differentiating Features Subspace Decompositor (DFSD), Flexible Noise Encryptor (FNE) and Staged Training Paradigm, used for decomposition, ID encryption and performance enhancement, respectively. Extensive experiments with existing encryption methods demonstrate our framework's preeminent performance in depression detection, ID reservation and audio reconstruction. Meanwhile, the experiments across various setting demonstrates our model's stability under different encryption strengths. Thus proving our framework's excellence in Confidentiality, Accuracy, Traceability, and Adjustability.
Figures & tables
Fig. 1: The kernel insight of our implementation
Method
PRE
REC
F1
ACC
Unsupervised Autoencoder [ 29 ]
85%
80%
83%
-
MFCC-based RNN [ 27 ]
75%(70%)
95%(26%)
84%(38%)
74%
Feature Fusion based on spectral analysis [ 22 ]
83%
100%
91%
85%
Attentive LSTM Network [ 46 ]
-
-
94.01%
90.2%
Graph Neural Network [ 31 ]
92.36±2.53%
92.18±1.55%
92.23±2.01%
92.21±1.86%
MFCC & CNN generated spectrogram features [ 8 ]
86.4%(93.4%)
93.6%(85.9%)
89.85%(89.5%)
90.26%
TABLE I: A brief summary of representative methods.
Fig. 2: General design of our proposed model.
Our Arthitecture - TAAC
ACC
REC
PRE
F1
Unencrypted
84.75%
0.88
0.86
0.87
Encrypted (Strength 5)
80.17%
0.88
0.85
0.86
Encrypted (Strength 10)
83.62%
0.92
0.84
0.88
Encrypted (Strength 25)
78.28%
0.91
0.8
0.85
Other Methods
ACC
REC
PRE
F1
Chaos Maps-Based Encryption
64.00%
0.72
0.83
0.77
TABLE II: Evaluation results on depression detection task.
Fig. 3: Confusion matrix on depression detection with encStrength 25.
Fig. 4: Confusion matrix on depression detection with encStrength 10.
Data - Method
MSE
MAE
PSNR
Encrypted ( TAAC )
7.56
2.75
-30.02dB
Reconstructed ( TAAC )
2.7e-4
1.0e-2
14.46dB
Encrypted (Chaos Maps-Based)
1.4e+7
2.7e+3
-1.91dB
Reconstructed (Chaos Maps-Based)
82.90
0.55
50.3dB
Encrypted (Homomorphic)
0.91
0.95
-20.82dB
Reconstructed (Homomorphic)
0.00
0.00
98.77dB
TABLE III: Statistical differences between the encrpyted audio, the reconstructed audio and the original audio.
Our Arthitecture - TAAC
ACC
FAR
FRR
EER
Encrypted (Strength 5)
75.40%
0.48
0.01
0.26
Encrypted (Strength 10)
58.47%
0.83
0.01
0.3
Encrypted (Strength 25)
50.53%
0.97
0.02
0.49
Other Methods
ACC
FAR
FRR
EER
Unencrypted
100.00%
0
0
0
Chaos Maps-Based Encryption
51.06%
0.75
0.23
0.35
TABLE IV: Evaluation results on PII extraction task.
Fig. 5: Radar chart comparison of the three encryption methods.
Speech-based mental health screening offers scalable depression detection, yet clinical deployment faces a significant barrier: users' privacy concerns about demographic information exposure. Current techniques struggle to resolve this conflict. Adversarial training often fails against unseen threats, whereas Differential Privacy tends to compromise diagnostic performance by injecting noise across all features. This paper presents InfoShield, which minimizes mutual information between speech representations and sensitive attributes while preserving depression classification accuracy. We identify that standard MINE estimators struggle with sequential speech due to temporal-static misalignment, and introduce TimeAwareMINE with cross-modal attention to align acoustic frames with attribute embeddings. Experiments on the Androids Corpus show InfoShield reduces gender inference from 92.6% to 55.5% and age inference from 55.7% to 30.3% with limited utility loss (6% F1 reduction), achieving F1=0.784 compared to prior SOTA's 0.723.
Xueyang Wu, Siyuan Liu, Kezhuo Yang +1
Shenzhen NeurStar Inc., China · University of York, United Kingdom · Shanghai Jiao Tong University, China
Automatic depression detection using speech signals with acoustic and textual modalities is a promising approach for early diagnosis. Depression-related patterns exhibit sparsity in speech: diagnostically relevant features occur in specific segments rather than being uniformly distributed. However, most existing methods treat all frames equally, assuming depression-related information is uniformly distributed and thus overlooking this sparsity. To address this issue, we proposes a depression detection network based on Adaptive Cross-Modal Gating (ACMG) that adaptively reassigns frame-level weights across both modalities, enabling selective attention to depression-related segments. Experimental results show that the depression detection system with ACMG outperforms baselines without it. Visualization analyses further confirm that ACMG automatically attends to clinically meaningful patterns, including low-energy acoustic segments and textual segments containing negative sentiments.
Hangbin Yu, Yudong Yang, Rongfeng Su +2
Shenzhen Institutes of Advanced Technology, Chinese Academy of Sciences, China · University of Chinese Academy of Sciences · Key Laboratory of Biomedical Imaging Science and System, Chinese Academy of Sciences, China
Speech recordings used for dementia detection inherently expose speaker identity, raising critical privacy concerns. Existing methods typically address only singular threats and fail to resolve the privacy--utility trade-off. We propose a multi-level framework designed to neutralize two distinct eavesdropping vectors. At the signal level, a Cumulative Signal Attack (CSA) concentrates perturbations in keyword-aligned regions to maximize transcription error (Word Error Rate WER = 1.00) while preserving vital prosodic biomarkers. At the feature level, a Gradient Reversal Layer (GRL) with Mutual Information (MI)-guided noise injection suppresses speaker-discriminative dimensions while retaining dementia-relevant diagnostic structure. Evaluated on the DementiaBank Pitt Corpus, our framework achieves near-chance speaker identification (Equal Error Rate EER = 0.59, F1 = 0.003) while maintaining strong dementia classification performance (F1 = 0.78, AUC = 0.86).
Henriette Flore Kenne, Raphael Anaadumba, Mohammad Arif Ul Alam
Richard A Miner School of Computer and Information Sciences, University of Massachusetts Lowell, Lowell, USA