Analysis of the chaotic characteristics of sleep EEG patterns from dominant Lyapunov exponents

Many biological oscillating systems, including the brain, exhibit the property of low dimensional chaos and EEG is now considered to be the outcome of a chaotic process. A defining feature of chaos is its Lyapunov exponents. The authors have evaluated the running dominant Lyapunov exponent of sleep EEG records using Wolf´s method. The results show that during sleep there are various transitions and the degree of chaoticity is dependent on the stage of sleep. The running dominant Lyapunov exponent curve yields an objective sleep pattern that is comparable to visual sleep scoring pattern or hypnogram. The different transitional states of sleep can be discerned from the values of the dominant Lyapunov exponents