Estimating complexity of spike-wave discharges with largest Lyapunov exponent in computational models and experimental data

Abstract

Here we consider the possibility to characterize the signal complexity of electroencephalo-grams using calculation of largest Lyapunov exponent explicitly from time series. This would help in detection of seizures, understanding and modeling epileptic activity. Baseline activity and spike-wave discharges (SWDs) were considered as regimes. Three channels relevant for absence epilepsy were studied: the parietal cortex, the ventroposterial medial nucleus of thalamus, and the reticular thalamic nucleus. Experimental data and two types of models were investigated. The result show that SWDs often treated as more or less regular oscillations are characterized by large positive Lyapunov expo-nent, not very different from the value obtained for baseline activity. The mesoscale network model of epilepsy is mostly able to reproduce this phenomenon, including absolute values. The more simple neuron mass model exhibits Lyapunov exponent during SWDs twice smaller than in baseline.

Keywords

• absence epilepsy
• lyapunov exponent
• time series analysis
• eeg
• spike-wave discharges
• genetic absence epilepsy model
• computational models