EEG Based Network Connectivity Classification in 7 and 9 Years- Old Children

Abstract

© 2018 IEEE. Investigating the brain neural pathways requires extensive knowledge of childrens' cognitive development. Significant variations in the cognitive process of a child, across ages, were assessed through the success in recognizing various stimuli. Longitudinal EEG data were gathered from 45 healthy children at the ages of seven and nine years. During the EEG data acquisition, children were asked to respond to the Flanker stimuli for investigating the development of the response conflict process. In each age group, the coherence and imaginary component of coherency were used to assess the network connectivity of each child. The congruent and incongruent stimuli were tried within delta, theta, alpha and beta bands. Following that, efficacies of various classification algorithms were tested in discriminating the coherency data of the two age groups. It was observed that brain connectivity was more helpful in distinguishing between two age groups using the incongruent Flanker stimuli. For the incongruent condition, the imaginary part of the coherency provides better features for classification. Using the features derived from the theta, alpha and beta bands, a classification accuracy of more than 94.31% could be achieved using the naïve Bayes classifier.