Classiﬁcation of Seizure and Non seizure EEG Signals Using Empirical Mode Decomposition
Classiﬁcation of Seizure and Non seizure EEG Signals Using Empirical Mode Decomposition.Classiﬁcation of Seizure and Non seizure EEG Signals Using Empirical Mode Decomposition. An EEG data set, which is publicly available online in  is used in this work. The data set consists of ﬁve subsets each containing 100 single-channel EEG signals, each one having 23.6 s duration. These signals have been selected from continuous multi channel EEG recording after visual inspection for artifacts.
The subsets Z and O have been recorded extra cranial, whereas the subsets N, F, and S have been recorded intra cranial. The subsets Z and O have been acquired from surface EEG recordings office healthy volunteers with eyes open and closed, respectively. The signals in two subsets have been measured in seizure-free intervals from ﬁve patients in the epileptic zone (subset F) and from the hippo cam-pal formation of the opposite hemisphere of the brain.The subset contains reactivity,selected from all recording sites exhibiting activity.
These are partial seizures, probably complex partial, which may then have developed to secondary generalized tonic conic seizures. The sampling frequency of EEG signals in the data set is 173.61 Hz. The signals have been randomized with regard to the recording location. The montage is a common average reference. Typical EEG signals (one from each subset) are shown in Fig. 1. In this work, the subsets Z, O, N, and F are combined to form non seizure (NS) class and subset S forms the seizure (S) class.
The empirical mode decomposition method is an adaptive, and data-dependent method. The EMD method does not require any condition about the stationary and linearity of the signal. The aim of the EMD method is to decompose the nonlinear and non stationary signal x(t) into a sum of intrinsic mode functions (IMFs). Each IMF satisﬁes two basic conditions: 1) the number of extrema and the number of zero crossings must be the same or differ at most by one.