Scholarlink

<p>Previously reported association between internet gaming disorder (IGD) and social anxiety (SA) was based on subjective questionnaires, while objective assessment of socioemotional stress using electroencephalography (EEG) was hindered by the absence of a reproducible realistic environment. Harnessing virtual reality (VR) technology, we investigated the characteristics of social scenario-triggered EEG signals in high-risk IGD individuals to develop a machine learning (ML) model for risk prediction. Thirty college students considered at high risk of IGD (HIGD) and thirty low-risk counterparts (LIGD) first completed three emotion-related questionnaires. EEG data in response to VR-based socioemotional stress were gathered using a VR headset with eight electrodes. The frequency components of time-domain signals and intra-brain synchronization of EEG information were analyzed with non-parametric tests. MATLAB’s fitcauto and fscmrmr were used on 564 EEG features per participant to classify the two groups, with k-nearest neighbors (kNN) identified as the optimal algorithm. The HIGD group exhibited more severe depression (<i>p</i> = 0.001) and SA (<i>p</i> = 0.02) than the LIGD group. Compared with the LIGD group, the HIGD group demonstrated lower neural activities at F3, Fz, F4, C3, C4, P3, and P4. Besides, HIGD individuals showed less intense delta/theta-band and frontal gamma-band synchronizations than in the LIGD group. Furthermore, EEG-based kNN identified 20 key EEG features for detecting HIGD with an accuracy of 97% and an area under the curve of 0.97. The findings highlighted the feasibility of assessing EEG responses to socioemotional stress in individuals with HIGD and SA, and applying an EEG-based kNN for early identification.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Application of EEG-Based Machine Learning in Time–Frequency and Brain Connectivity Analysis Among Individuals at High Risk of Internet Gaming Disorder with Social Anxiety

  • 期刊论文

  • 发表日期：

    2025年10月7日

查看全文
  • Pin-Yang Yeh,
  • Chun-Ling Lin,
  • Cheuk-Kwan Sun,
  • Shih-Yi Tung

摘要

Previously reported association between internet gaming disorder (IGD) and social anxiety (SA) was based on subjective questionnaires, while objective assessment of socioemotional stress using electroencephalography (EEG) was hindered by the absence of a reproducible realistic environment. Harnessing virtual reality (VR) technology, we investigated the characteristics of social scenario-triggered EEG signals in high-risk IGD individuals to develop a machine learning (ML) model for risk prediction. Thirty college students considered at high risk of IGD (HIGD) and thirty low-risk counterparts (LIGD) first completed three emotion-related questionnaires. EEG data in response to VR-based socioemotional stress were gathered using a VR headset with eight electrodes. The frequency components of time-domain signals and intra-brain synchronization of EEG information were analyzed with non-parametric tests. MATLAB’s fitcauto and fscmrmr were used on 564 EEG features per participant to classify the two groups, with k-nearest neighbors (kNN) identified as the optimal algorithm. The HIGD group exhibited more severe depression (p = 0.001) and SA (p = 0.02) than the LIGD group. Compared with the LIGD group, the HIGD group demonstrated lower neural activities at F3, Fz, F4, C3, C4, P3, and P4. Besides, HIGD individuals showed less intense delta/theta-band and frontal gamma-band synchronizations than in the LIGD group. Furthermore, EEG-based kNN identified 20 key EEG features for detecting HIGD with an accuracy of 97% and an area under the curve of 0.97. The findings highlighted the feasibility of assessing EEG responses to socioemotional stress in individuals with HIGD and SA, and applying an EEG-based kNN for early identification.