Related Experiment Video
Updated: Aug 4, 2025
07:46
Gait Analysis of Age-dependent Motor Impairments in Mice with Neurodegeneration
Published on: June 18, 2018
12.0K
Using gait videos to automatically assess anxiety
Yeye Wen1,2, Baobin Li3, Xiaoqian Liu2
1School of Electronic, Electrical and Communication Engineering, University of Chinese Academy of Sciences, Beijing, China.
Frontiers in Public Health
|April 3, 2023
Summary
Anxiety can be reliably assessed using 2D gait analysis. This study developed a machine learning model from gait videos, demonstrating its effectiveness for objective anxiety detection.
Area of Science:
- Psychiatry
- Computer Science
- Biomedical Engineering
Background:
- Global rise in anxiety disorders necessitates objective assessment methods.
- Current anxiety identification techniques lack maturity and validated reliability.
- Need for dependable, automated tools for anxiety detection.
Purpose of the Study:
- To propose a reliable and valid automatic anxiety assessment model.
- To explore the efficacy of using 2D gait analysis for anxiety detection.
- To establish a foundation for non-invasive anxiety monitoring.
Main Methods:
- Collected 2D gait videos and GAD-7 scores from 150 participants.
- Extracted static, dynamic, time-domain, and frequency-domain gait features.
- Developed and evaluated machine learning models for anxiety prediction.
Main Results:
- Wavelet decomposition layers significantly impacted frequency-domain modeling.
- Dynamic time-frequency gait features were more influential than static features.
- The model achieved a high correlation (0.725) with GAD-7 scores, showing strong reliability and validity, particularly in women.
Conclusions:
- Anxiety assessment via 2D gait video modeling is reliable and effective.
- The developed model offers a basis for real-time, non-invasive anxiety assessment.
- Gait analysis presents a promising avenue for objective mental health monitoring.
