Scholarlink

<p>Brain disorder issues are quite common in Parkinson’s disease (PD) patients following Deep Brain Stimulation (DBS). Therefore, identifying a predictor to classify the patient’s brain dysfunction status is crucial. The primary goal of this research is to create a morlet wavelet-based Electroencephalogram (EEG) biomarker that accurately distinguishes four types of cognition by extracting essential features and establishing a link between Quantitative EEG (QEEG) in all categories. This research focuses on the feature extraction of EEG data using the morlet wavelet approach to identify the whole spectrum of Cognition. The dataset has been received from the North Shore Institute of Health Science in Chicago. The XGBoost outperformed other algorithms such as Decision tree (DT) and Random Forest (RF) classifiers, achieving 100% training accuracy and 95.43% testing accuracy (sensitivity = 0.94, specificity = 0.98), as well as an AUC score of 1.0. The suggested model outperforms all other techniques evaluated in the literature in terms of accuracy, sensitivity, and specificity. The study also contains a comparative statistical analysis of quantitative EEG features, which uses Welch’s ANOVA and the Post-hoc Games-Howell test to determine a link between EEG measures. First, regions of interest (ROIs) are chosen based on feature significance scores, which are considered as spatially informative regions for cognition in PD. This study reveals that the frontopolar, frontal, and parietal regions show substantial differences in delta, theta, alpha&#xa0;and beta power between all groups. Specifically, delta and theta power increase gradually from HC to PDD groups, whereas alpha and beta power drop in the same manner.</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

Advanced cognitive classification in Parkinson’s disease through morlet wavelet analysis of EEG-derived key brain regions

  • 期刊论文

  • 发表日期：

    2026年1月3日

查看全文
  • Aanchal Sharma,
  • Anu Gupta,
  • Sukesha Sharma

摘要

Brain disorder issues are quite common in Parkinson’s disease (PD) patients following Deep Brain Stimulation (DBS). Therefore, identifying a predictor to classify the patient’s brain dysfunction status is crucial. The primary goal of this research is to create a morlet wavelet-based Electroencephalogram (EEG) biomarker that accurately distinguishes four types of cognition by extracting essential features and establishing a link between Quantitative EEG (QEEG) in all categories. This research focuses on the feature extraction of EEG data using the morlet wavelet approach to identify the whole spectrum of Cognition. The dataset has been received from the North Shore Institute of Health Science in Chicago. The XGBoost outperformed other algorithms such as Decision tree (DT) and Random Forest (RF) classifiers, achieving 100% training accuracy and 95.43% testing accuracy (sensitivity = 0.94, specificity = 0.98), as well as an AUC score of 1.0. The suggested model outperforms all other techniques evaluated in the literature in terms of accuracy, sensitivity, and specificity. The study also contains a comparative statistical analysis of quantitative EEG features, which uses Welch’s ANOVA and the Post-hoc Games-Howell test to determine a link between EEG measures. First, regions of interest (ROIs) are chosen based on feature significance scores, which are considered as spatially informative regions for cognition in PD. This study reveals that the frontopolar, frontal, and parietal regions show substantial differences in delta, theta, alpha and beta power between all groups. Specifically, delta and theta power increase gradually from HC to PDD groups, whereas alpha and beta power drop in the same manner.