Functional Connectivity Network based on Graph Analysis of Scalp EEG for Epileptic Classification

the proposed study presents a novel fully automated data-driven approach for differentiating epileptic subjects from normal controls using graph-based functional connectivity networks calculated using scalp EEG. A set of fourteen density-related, graph distance-based and spectral topological features extracted from the network graph is employed for the classification process. The proposed algorithm demonstrated an accuracy of 87.5% with a sensitivity of 75% and specificity of 100% when tested on 8 subjects. The study showed that graph-based functional connectivity networks in epileptic subjects were significantly different from those of controls (p<0.05). The study has the potential for aiding neurologists in decision making for diagnostic purposes solely based on scalp EEG.