Classification of ADHD/Normal Participants Using Frequency Features of ERP’s Independent Components

This study investigates the Event Related Potentials (ERP) obtained from Independent Components of EEG (ERPIC) while participants performed a sustained attention task. EEG signals were recorded from 50 adult participants including ADHD and normal subjects while performing Continuous Performance Test (CPT). Signals were recorded from 21 Ag/AgCl electrodes according to the international 10-20 standard. Independent Component Analysis (ICA) was used as the processing method. For ERP extraction, average of each group of signals which were time-locked to the onset of stimuli was calculated. Several frequency features were extracted from different ERPICs. High accuracy (92%) was achieved in classification of clinical and non-clinical participants using combination of two features in a K-Nearest Neighbors (KNN) classifier. Nine pairs of features resulted in such accuracy, while most of the best features are related to the power in γ band which is consistent with the previous studies. Regarding the ERP groups, most of the best features are related to wrong answeredtargets and to time block ERPICs. The results revealed a promising relation between clinical situation of the participants and some parameters of brain independent components which can be used for further evaluations of the sustained attention level.

کلیدواژه ها:

Attention Deficit / Hyperactivity Disorder (ADHD) ، K-Nearest Neighbors (KNN) Classifier ، Continuous PerformanceTest (CPT) ، Event Related Potentials (ERP) ، Feature Extraction ، Independent Component Analysis (ICA) ، Sustained Attention