Scholarlink

<p>Alzheimer's disease (AD) is a neurodegenerative disorder that poses a serious global threat to human health. Accurate detection of AD is critical for improving patient outcomes, yet current detection methods still exhibit significant limitations in accuracy, necessitating further improvement. In this paper, we propose an AD detection method based on multi-frequency electroencephalography (EEG) and a multi-functional connectivity graph convolutional network (MFE-FCGCN). The method conducts multi-frequency analysis of power spectral density (PSD) features across five EEG frequency bands (Delta, Theta, Alpha, Beta, Gamma) and constructs two functional connectivity networks based on mutual information and Pearson correlation coefficients. This network comprehensively captures abnormalities in brain network structures induced by AD, across different frequency bands and connectivity modes. By leveraging a multi-dimensional feature extraction and fusion strategy, the model effectively identifies EEG pattern changes associated with AD, enhancing detection accuracy. Experimental results demonstrate that this method achieves a classification accuracy of 95.09% and an AUC of 98.36% on the ds00450 dataset from the OpenNeuro database, significantly outperforming traditional approaches and validating its superior performance in AD detection. Code is available at <a href="https://github.com/brief-city/MFE-FCGCN-model">https://github.com/brief-city/MFE-FCGCN-model</a>.</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

Multi-frequency EEG and multi-functional connectivity graph convolutional network based detection method of patients with Alzheimer’s disease

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2025年6月30日

查看全文
  • Yujian Liu,
  • Libing An,
  • Haiqiang Yang,
  • Shuzhi Sam Ge

摘要

Alzheimer's disease (AD) is a neurodegenerative disorder that poses a serious global threat to human health. Accurate detection of AD is critical for improving patient outcomes, yet current detection methods still exhibit significant limitations in accuracy, necessitating further improvement. In this paper, we propose an AD detection method based on multi-frequency electroencephalography (EEG) and a multi-functional connectivity graph convolutional network (MFE-FCGCN). The method conducts multi-frequency analysis of power spectral density (PSD) features across five EEG frequency bands (Delta, Theta, Alpha, Beta, Gamma) and constructs two functional connectivity networks based on mutual information and Pearson correlation coefficients. This network comprehensively captures abnormalities in brain network structures induced by AD, across different frequency bands and connectivity modes. By leveraging a multi-dimensional feature extraction and fusion strategy, the model effectively identifies EEG pattern changes associated with AD, enhancing detection accuracy. Experimental results demonstrate that this method achieves a classification accuracy of 95.09% and an AUC of 98.36% on the ds00450 dataset from the OpenNeuro database, significantly outperforming traditional approaches and validating its superior performance in AD detection. Code is available at https://github.com/brief-city/MFE-FCGCN-model.