Scholarlink

<p>Early and accurate diagnosis of Parkinson’s disease (PD) is essential for timely intervention and improved patient outcomes. Traditional diagnostic methods, such as clinical assessments and neuroimaging often fail to detect PD in its early stages leading to delayed treatment. Electroencephalography (EEG)-based detection has emerged as a promising alternative, but existing methods suffer from poor artifact handling, suboptimal feature extraction, and limited classification accuracy. To address these limitations, this study proposes an EEG-Based Parkinson’s disease Diagnosis via Verifiable Convolutional Neural Networks and Self-Competition Particle Swarm Optimization (EEG-PDD-VCNN-SCPSOA)<b>.</b> EEG signals recorded at Hospital University Kebangsaan Malaysia were pre-processed using Dual Central Difference Kalman Filtering to reduce eye-blinking artifacts. Robust feature extraction was achieved using the Concise Empirical Wavelet Transform, integrating linear predictive coefficients and correlation coefficients. The extracted features were classified using a VerifiableConvolutional Neural Network (VCNN) optimized with Self-Competition Particle Swarm Optimization (SCPSOA) to enhance diagnostic accuracy. The proposed EEG-PDD-VCNN-SCPSOA model was evaluated using accuracy, precision, F1-score, recall, and Matthews correlation coefficient. Experimental results demonstrate that the proposed approach achieves 18.97%, 24.57%, and 32.68% higher accuracy and 19.84%, 24.93%, and 31.62% higher precision compared to three existing methods: DPD-EEG-FAWT (decision support system for PD diagnosis using FAWT and entropy features), EML-PDD-EEG (entropy-based machine learning model for fast diagnosis), and PDD-EEG-TML (monitoring PD and EEG-based affective analysis). These findings highlight the effectiveness of EEG-PDD-VCNN-SCPSOA in enhancing automated PD diagnosis.</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

EEG-based Parkinson’s disease diagnosis via verifiable convolutional neural networks and self-competition particle swarm optimization

  • 期刊论文

  • 发表日期：

    2025年8月16日

查看全文
  • Chetan Balaji,
  • D. S. Suresh

摘要

Early and accurate diagnosis of Parkinson’s disease (PD) is essential for timely intervention and improved patient outcomes. Traditional diagnostic methods, such as clinical assessments and neuroimaging often fail to detect PD in its early stages leading to delayed treatment. Electroencephalography (EEG)-based detection has emerged as a promising alternative, but existing methods suffer from poor artifact handling, suboptimal feature extraction, and limited classification accuracy. To address these limitations, this study proposes an EEG-Based Parkinson’s disease Diagnosis via Verifiable Convolutional Neural Networks and Self-Competition Particle Swarm Optimization (EEG-PDD-VCNN-SCPSOA). EEG signals recorded at Hospital University Kebangsaan Malaysia were pre-processed using Dual Central Difference Kalman Filtering to reduce eye-blinking artifacts. Robust feature extraction was achieved using the Concise Empirical Wavelet Transform, integrating linear predictive coefficients and correlation coefficients. The extracted features were classified using a VerifiableConvolutional Neural Network (VCNN) optimized with Self-Competition Particle Swarm Optimization (SCPSOA) to enhance diagnostic accuracy. The proposed EEG-PDD-VCNN-SCPSOA model was evaluated using accuracy, precision, F1-score, recall, and Matthews correlation coefficient. Experimental results demonstrate that the proposed approach achieves 18.97%, 24.57%, and 32.68% higher accuracy and 19.84%, 24.93%, and 31.62% higher precision compared to three existing methods: DPD-EEG-FAWT (decision support system for PD diagnosis using FAWT and entropy features), EML-PDD-EEG (entropy-based machine learning model for fast diagnosis), and PDD-EEG-TML (monitoring PD and EEG-based affective analysis). These findings highlight the effectiveness of EEG-PDD-VCNN-SCPSOA in enhancing automated PD diagnosis.