Scholarlink

<p>Parkinson’s disease (PD) is a brain condition causing movement, mental health, sleep, and pain issues. It progresses over time and has no cure, so early detection is crucial. Hence, the proposed framework implements a 2D cross-chaotic Lemur optimization with long short-term memory (2DC2L2OSTM) approach to detect neurological movement disorders based on electroencephalogram (EEG) and clinical data, which are not focused on prevailing approaches. Initially, the preprocessed EEG signals are processed for power spectral density estimation, and then the Bezier curve fuzzy logic system (Bc-FLS) is applied to remove artifacts. Next, the analysis of variance (ANOVA) is utilized to analyze temporal patterns through motor symptom detection and heat map generation. At the same time, the fusion ranking Spearman correlation (Fr-SC) is applied for correlation calculation. In the meantime, the unit-Rayleigh principal component analysis (U-PCA) is utilized to extract the features of the detected subbands (alpha, beta, gamma, and delta) and calculate the correlation. Meanwhile, the features are extracted as of the preprocessed clinical data, and then, based on the extracted features, the 2DC2L2OSTM is utilized to classify neurological movement disorders. Thus, the experiment results illustrate that the proposed framework attained 98.97% accuracy and 13450&#xa0;ms of training time, which indicates that the proposed framework efficiently identifies neurological movement disorders compared to the other prevailing approaches.</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

Multimodal data analysis-based neurological movement disorders prediction using U-PCA and 2D Cross-Chaotic Lemur optimization with Long Short-Term Memory

  • 期刊论文

  • 发表日期：

    2025年9月30日

查看全文
  • T. R. Lekhaa,
  • Manoj Kumar,
  • Hazem Gouda,
  • M. Bala Anand

摘要

Parkinson’s disease (PD) is a brain condition causing movement, mental health, sleep, and pain issues. It progresses over time and has no cure, so early detection is crucial. Hence, the proposed framework implements a 2D cross-chaotic Lemur optimization with long short-term memory (2DC2L2OSTM) approach to detect neurological movement disorders based on electroencephalogram (EEG) and clinical data, which are not focused on prevailing approaches. Initially, the preprocessed EEG signals are processed for power spectral density estimation, and then the Bezier curve fuzzy logic system (Bc-FLS) is applied to remove artifacts. Next, the analysis of variance (ANOVA) is utilized to analyze temporal patterns through motor symptom detection and heat map generation. At the same time, the fusion ranking Spearman correlation (Fr-SC) is applied for correlation calculation. In the meantime, the unit-Rayleigh principal component analysis (U-PCA) is utilized to extract the features of the detected subbands (alpha, beta, gamma, and delta) and calculate the correlation. Meanwhile, the features are extracted as of the preprocessed clinical data, and then, based on the extracted features, the 2DC2L2OSTM is utilized to classify neurological movement disorders. Thus, the experiment results illustrate that the proposed framework attained 98.97% accuracy and 13450 ms of training time, which indicates that the proposed framework efficiently identifies neurological movement disorders compared to the other prevailing approaches.