Application of extreme learning machine to epileptic seizure detection based on lagged Poincare plots

Epilepsy is a serious brain disorder affecting nearly 1 % of the world's population. Detecting the epileptic seizures in EEGs is not only the first step for the diagnosis, but also a significant evidence for the treatment follow-up in epilepsy patients. In recent years, automatic seizure detection using epileptic EEGs has been developed with the significance of relieving the heavy workload of traditional visual inspection for diagnosing epilepsy. The appropriate feature extraction method and efficient classifier are recognized to be crucial in the successful realization. This paper first designs a novel lagged-PoincaA center dot e-based feature extraction method on the basis of a class of lagged PoincaA center dot e plots, as well as the scatter-degree and distribution-uniformity of them which are explored to characterize the lag-T PoincaA center dot e plots from the quantitative point of view. Then we propose an automatic seizure detection method LPBF-ELM which integrates the lagged-PoincaA center dot e-based feature LPBF and extreme learning machine (ELM). Experimental results on Bonn database demonstrate that the proposed method LPBF-ELM does a good job in epileptic seizure detection while preserving the efficiency and simplicity.