EEG Signal Analysis Based on Time-variant Coupled Network Lattice Model

Brain electrical activity is widely accepted as the typical non-stationary signal. In addition, there exist more evidences that both EEG and ERP signals are chaotic signal produced by the nonlinear dynamics system. To investigate the time-varying nonlinear dynamics of the ERP under specified cognitive tasks, a novel model based on the time-variant coupled map lattice system is proposed for this purpose. The time-variant largest Lyapunov exponent (LLE) is also defined as the quantitative parameters to reveal the global characters of system and discover new information. Several simulations and real ERP signals under different fixed location cue were examined in terms of LLE spectra for studying the signal´s dynamic complexity. The experimental results show that the brain chaos changes with time under different attention tasks of the information processing. The influence of the LLE with the scopes size of cues mainly occurs in P2 period. No simple linearly relationship between the scopes of cues and the influence of complexity is found.