Scholarlink

<p>Attention-Deficit Hyperactivity Disorder (ADHD) is a neurodevelopmental condition that manifests in early childhood and often persists into adulthood. One approach to diagnosing ADHD is through quantitative analysis of the Electroencephalogram (EEG) signal, also known as Quantitative Electroencephalography (QEEG). This paper proposes a novel diagnostic framework that quantifies cognitive ability through spectral, power and connectivity QEEG pooled across temporal, spectral and spatial dimensions of an EEG signal evoked by visual stimuli; the generalizability of this approach is proved using a subject-wise data splitting strategy. The spectral features represent the frequency characteristics of the EEG signal, while power features represent the amplitude characteristics. Connectivity refers to the inter-hemispheric brain region connectivity which is measured using averaged correlation and coherence statistics computed between all possible channel pairs in the left and right hemispheres. Ensemble-based feature selection is then employed for selection of optimal features from the concatenated feature list. Two ADHD datasets have been used for the evaluation. The first dataset is smaller and balanced, and involves 121 participants (ADHD:61, Healthy:60), while the second dataset is larger and imbalanced, and involves 144 participants (ADHD:23,773, Healthy:10,129). The proposed method achieves the highest accuracies of 98.57 ± 0.88% and 98.68 ± 0.52% for the two datasets, respectively, using the kNN classifier with tenfold cross-validation. The proposed method also achieves high accuracies of 98.18 ± 0.98% and 97.69 ± 0.52% for the leave-one-subject-out (LOSO) validation on the two datasets, establishing the generalizability of the method across subjects.&#xa0;The results establish the diagnostic efficacy of the proposed method and its potential for influencing future research.</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

Generalizable temporal-spectral-spatial quantitative electroencephalogram based diagnosis of attention-deficit hyperactivity disorder in children

  • 期刊论文

  • 发表日期：

    2025年9月9日

查看全文
  • Ruchi Holker,
  • Seba Susan

摘要

Attention-Deficit Hyperactivity Disorder (ADHD) is a neurodevelopmental condition that manifests in early childhood and often persists into adulthood. One approach to diagnosing ADHD is through quantitative analysis of the Electroencephalogram (EEG) signal, also known as Quantitative Electroencephalography (QEEG). This paper proposes a novel diagnostic framework that quantifies cognitive ability through spectral, power and connectivity QEEG pooled across temporal, spectral and spatial dimensions of an EEG signal evoked by visual stimuli; the generalizability of this approach is proved using a subject-wise data splitting strategy. The spectral features represent the frequency characteristics of the EEG signal, while power features represent the amplitude characteristics. Connectivity refers to the inter-hemispheric brain region connectivity which is measured using averaged correlation and coherence statistics computed between all possible channel pairs in the left and right hemispheres. Ensemble-based feature selection is then employed for selection of optimal features from the concatenated feature list. Two ADHD datasets have been used for the evaluation. The first dataset is smaller and balanced, and involves 121 participants (ADHD:61, Healthy:60), while the second dataset is larger and imbalanced, and involves 144 participants (ADHD:23,773, Healthy:10,129). The proposed method achieves the highest accuracies of 98.57 ± 0.88% and 98.68 ± 0.52% for the two datasets, respectively, using the kNN classifier with tenfold cross-validation. The proposed method also achieves high accuracies of 98.18 ± 0.98% and 97.69 ± 0.52% for the leave-one-subject-out (LOSO) validation on the two datasets, establishing the generalizability of the method across subjects. The results establish the diagnostic efficacy of the proposed method and its potential for influencing future research.