Scholarlink

<p>Ability assessment in schools is generally based on psychometric measures of general intelligence. Viewed from the subjective point&#xa0;of view, psychometric measures are not devoid of disadvantages and problems. Based on empirical wavelet transform (EWT), the current study proposes a new scheme to analyze and classify the traits of EEG signals from high-ability, average-ability, low-ability, and ADHD students. Compared to conventional strategies, EWT offers more adaptiveness in deconstructing EEG signals, hence making it possible to derive nonlinear traits more representative of the dynamic processes of the brain. Nonlinear traits are then retrieved from the rebuilt rhythms using EWT. Unlike traditional methods such as FFT and DWT, which rely on fixed basis functions, EWT dynamically decomposes EEG signals to capture nonlinear traits that represent the complex and varied processes of the brain. This adaptability not only enhances the resolution of the analysis but also facilitates a more effective extraction of oscillatory modes relevant to cognitive states. After selecting the measured traits by the&#xa0;Least Absolute Shrinkage and Selection Operator (LASSO), they are fed to SVM and KNN classifiers. These experiments showed that the recommended scheme resulted in a good accuracy of 78.91% attained by the recommended method utilizing the LOSOCV strategy in classifying EEG of students with high ability, average ability, low ability, and ADHD. Moreover, the recommended scheme was externally validated with an unseen database for EEG and was able to obtain a striking accuracy of 94.87% to detect ADHD subtypes and healthy controls. While the outcomes of this investigation have been promising, considerable work is still needed to further develop and expand it.</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 Classification of Students with High Ability, Average Ability, Low Ability, and ADHD Through Empirical Wavelet Transform

  • 期刊论文

  • 发表日期：

    2025年8月15日

查看全文
  • Zhihong Zheng,
  • Lin Weng

摘要

Ability assessment in schools is generally based on psychometric measures of general intelligence. Viewed from the subjective point of view, psychometric measures are not devoid of disadvantages and problems. Based on empirical wavelet transform (EWT), the current study proposes a new scheme to analyze and classify the traits of EEG signals from high-ability, average-ability, low-ability, and ADHD students. Compared to conventional strategies, EWT offers more adaptiveness in deconstructing EEG signals, hence making it possible to derive nonlinear traits more representative of the dynamic processes of the brain. Nonlinear traits are then retrieved from the rebuilt rhythms using EWT. Unlike traditional methods such as FFT and DWT, which rely on fixed basis functions, EWT dynamically decomposes EEG signals to capture nonlinear traits that represent the complex and varied processes of the brain. This adaptability not only enhances the resolution of the analysis but also facilitates a more effective extraction of oscillatory modes relevant to cognitive states. After selecting the measured traits by the Least Absolute Shrinkage and Selection Operator (LASSO), they are fed to SVM and KNN classifiers. These experiments showed that the recommended scheme resulted in a good accuracy of 78.91% attained by the recommended method utilizing the LOSOCV strategy in classifying EEG of students with high ability, average ability, low ability, and ADHD. Moreover, the recommended scheme was externally validated with an unseen database for EEG and was able to obtain a striking accuracy of 94.87% to detect ADHD subtypes and healthy controls. While the outcomes of this investigation have been promising, considerable work is still needed to further develop and expand it.