Related Experiment Video
Updated: May 23, 2025
07:12
Individualized rTMS Treatment for Depression using an fMRI-Based Targeting Method
Published on: August 2, 2021
3.4K
Privacy-Conscious Internet Behavior for Depression Detection With Cross-Scale Adaptive Transformer
IEEE Journal of Biomedical and Health Informatics
|May 21, 2025
Summary
This study introduces a privacy-preserving AI model to detect depression in college students using online behavior patterns. The novel approach offers a non-intrusive screening method, enhancing mental health support accessibility.
Area of Science:
- Digital Mental Health
- Computational Psychiatry
- Machine Learning in Healthcare
Background:
- Depression is a major cause of suicide among college students, necessitating scalable screening.
- Online behavior shows potential for depression detection, but privacy concerns are a barrier.
Purpose of the Study:
- To develop a privacy-conscious model for detecting depression using irregular time series data from online behavior.
- To address privacy limitations while leveraging internet usage patterns for mental health screening.
Main Methods:
- Proposed a privacy-conscious cross-scale adaptive transformer for irregular time series data.
- Incorporated adaptive sampling for temporal resolution unification and cross-scale attention for pattern recognition.
- Excluded sensitive personal information, focusing on application categories and usage patterns.
Main Results:
- The proposed adaptive transformer model outperformed classic irregular time series models.
- Demonstrated effective depression-related behavioral pattern capture using weakly private online data.
Conclusions:
- The developed method offers a promising, non-intrusive approach for depression detection in college students.
- Privacy-conscious analysis of online activity patterns can enhance mental health screening scalability and accessibility.
