JMIR Publications

ABSTRACT

Background:

Suicidal ideation (SI) among university students is a growing public health concern. Self-report screening can be limited by concealment and delayed disclosure. We evaluated a leakage-resistant, proof-of-concept pipeline to detect SI from standardized smartphone-recorded speech.

Objective:

evaluate a leakage-resistant, proof-of-concept pipeline to detect SI from standardized smartphone-recorded speech.

Methods:

Questionnaire data and speech recordings were collected via a WeChat mini-program. After screening and clinical confirmation, 96 participants (48 SI; 48 controls) were included. Each participant read 16 standardized sentences. Acoustic features were extracted using openSMILE (v3.0.2), yielding a 570-dimensional feature vector per utterance. To prevent leakage from multiple recordings per speaker, we used participant-level 5-fold cross-validation, assigning all recordings from each participant to a single fold. Seven machine-learning algorithms were evaluated using AUC, accuracy, and F1-score.

Results:

Acoustic-based models discriminated SI from control participants. Random Forest achieved AUC = 0.813 (accuracy = 0.748), and Naïve Bayes achieved AUC = 0.806 (accuracy = 0.757). Feature families related to pitch, MFCCs, and harmonicity contributed to model performance.

Conclusions:

Standardized read speech captured via smartphones shows promise for SI discrimination under a leakage-aware evaluation design. External validation and testing with more naturalistic speech are warranted. Clinical Trial: ChiCTR2500106625