Novel entropy difference-based EEG channel selection technique for automated detection of ADHD

TL;DR

This work addresses automated ADHD detection from EEG by introducing Entropy Difference (EnD) for objective channel selection, quantified as $EnD(C) = |H_{ADHD}(C) - H_{HC}(C)|$, to identify the most informative channels. It evaluates three feature pipelines (DWT, EMD, SLBP) on EnD-selected channels, combined with SVM, k-NN, and Ensemble classifiers, and demonstrates that EnD-based ranking consistently outperforms entropy-based ranking. The key finding is that using the top three EnD-ranked channels with Chi-square feature selection yields very high accuracy, with SLBP features achieving up to 99.29% on 10-fold cross-validation using k-NN, indicating strong discriminative power and low computational burden. This approach offers a practical, scalable EEG-based ADHD screening pathway and has potential for extension to other brain disorders and integration with explainable AI.

Abstract

Attention deficit hyperactivity disorder (ADHD) is one of the common neurodevelopmental disorders in children. This paper presents an automated approach for ADHD detection using the proposed entropy difference (EnD)- based encephalogram (EEG) channel selection approach. In the proposed approach, we selected the most significant EEG channels for the accurate identification of ADHD using an EnD-based channel selection approach. Secondly, a set of features is extracted from the selected channels and fed to a classifier. To verify the effectiveness of the channels selected, we explored three sets of features and classifiers. More specifically, we explored discrete wavelet transform (DWT), empirical mode decomposition (EMD) and symmetrically-weighted local binary pattern (SLBP)-based features. To perform automated classification, we have used k-nearest neighbor (k-NN), Ensemble classifier, and support vectors machine (SVM) classifiers. Our proposed approach yielded the highest accuracy of 99.29% using the public database. In addition, the proposed EnD-based channel selection has consistently provided better classification accuracies than the entropy-based channel selection approach. Also, the developed method

Figures (7)

• Figure 1: Block diagram of the proposed methodology for the automated detection of ADHD detection from EEG signals.
• Figure 2: Sample plot of multi-channel EEG segment.
• Figure 3: Comparison of ENS (left), k-NN (middle), and SVM (right) classifier accuracy obtained from EMD-based features for EN- and EnD-based channel selection method (before feature selection).
• Figure 4: Comparison of ENS (left), k-NN (middle), and SVM (right) classifier accuracy obtained from DWT-based features for En- and EnD-based channel selection method (before feature selection).
• Figure 5: Comparison of ENS (left), k-NN (middle), and SVM (right) classifier accuracy obtained from SLBP-based features for En- and EnD-based channel selection method (before feature selection).