Detecting different tasks using EEG-source-temporal features

Abstract

This study proposes a new type of features extracted from Electroencephalography (EEG) signals to distinguish between different tasks. EEG signals are collected from six children aged between two to six years old during opened and closed eyes tasks. For each time-sample, Time Difference of Arrival (TDOA) is applied to EEG time series to compute the source-temporal- features that are assigned to x, y and z coordinates. The features are classified using neural network. The results show an accuracy of around 100% for eyes open task and around (83%-95%) for eyes closed tasks for the same subject. This study highlights the use of new types of features (source-temporal features), to characterize the brain functional behavior.