军事医学2025，Vol.49Issue(12)：890-897,8.DOI:10.7644/j.issn.1674-9960.2025.12.002
图卷积神经网络在电磁辐射职业人员认知功能障碍评估中的应用研究
Application of graph convolutional networks for assessing cognitive dysfunction in occupational personnel exposed to electromagnetic radiation
孙月 1董霁 1师博洋 1田含科 1赵黎 1王惠 1王浩宇 1彭瑞云1
作者信息
- 1. 军事科学院军事医学研究院,北京 100850
- 折叠
摘要
Abstract
Objective To construct a cognitive dysfunction evaluation model for occupational personnel exposed to electromagnetic radiation based on graph convolutional networks(GCN),providing new methods for early warning.Methods One hundred electromagnetic radiation related occupational personnel were recruited as study subjects.Four types of data were collected:general demographic characteristics,cognitive-related neuropsychological indicators,Montreal cognitive assessment(MoCA)data and serum biochemical indicators.Based on these data,three models were constructed using different types of GCNs(basic-GCN,RA-GCN,and RF-GCN)and their performance was evaluated.Feature ablation experiments were conducted to analyze the impact of demo graphic characteristics and serum biochemical indicators on model performance.Results The RF-GCN model demonstrated higher assessment performance than the other models with an accuracy of 0.710±0.089.The feature ablation experiments showed that removing any demo graphic characteristics or serum biochemical indicators led to a decline in performance.Conclusion The RF-GCN model constructed in this study provides a novel method for cognitive dysfunction assessment in occupational personnel exposed to electromagnetic radiation.关键词
认知障碍/深度学习/图卷积神经网络/电磁辐射/职业人员Key words
cognitive dysfunction/deep learning/graph convolutional network/electromagnetic radiation/occupational personnel分类
医药卫生引用本文复制引用
孙月,董霁,师博洋,田含科,赵黎,王惠,王浩宇,彭瑞云..图卷积神经网络在电磁辐射职业人员认知功能障碍评估中的应用研究[J].军事医学,2025,49(12):890-897,8.
