Analysing epileptic EEGs with a visibility graph algorithm

This paper analyzes the human epileptic lectroencephalogram (EEG) based on a visibility graph algorithm. A single-channel EEG is mapped into a visibility graph (VG). Then its mean degree and degree distribution on the VG are extracted. It is shown that the mean degree on a VG from an epileptic subject is larger than that on a healthy subject based on the VG. The number of nodes having five degree on a VG from a healthy subject is significantly different from the number of nodes having the same degree on the VG from an epileptic subject. The mean degree and the number of nodes with five and eight degrees are used to discriminate the healthy EEGs, seizure EEGs and inter-ictal EEGs. Experimental results demonstrate that the visibility graph algorithm has a high classification accuracy to identify these three types of EEGs.