Assessment method of depressive disorder level based on graph attention network

¹ Department of Automation, Faculty of Information Technology, Beijing University of Technology, Beijing 100124
² Beijing International Collaboration Base on Brain Informatics and Wisdom Services, Beijing 100124, China
³ Engineering Research Center of Intelligent Perception and Autonomous Control, Ministry of Education, Beijing 100124
⁴ Engineering Research Center of Digital Community, Ministry of Education, Beijing 100124

^* Corresponding author: limi@bjut.edu.cn

Abstract

This paper presents an approach to predict the depression self-rating scale of Patient Health Questions-9 (PHQ-9) values from pupil-diameter data based on the graph attention network (GAT). The pupil diameter signal was derived from the eye information collected synchronously while the subjects were viewing the virtual reality emotional scene, and then the scores of PHQ-9 depression self-rating scale were collected for depression level. The chebyshev distance based GAT (Chebyshev-GAT) was constructed by extracting pupil-diameter change rate, emotional bandwidth, information entropy and energy, and their statistical distribution. The results show that, the error (MAE and SMRE)of the prediction results using Chebyshev-GAT is smaller then the traditional regression prediction model.