The Effect of Long-Term Apnea on the Autonomic Nervous System

The complexity of physiological system can be well represented by the complexity of time series. In this paper, complexity analysis was performed on the inter-beat-interval(IBI) of apnea syndrome and normal sleep population. Hurst exponent, sample entropy and fractal dimension were used as three indicators of complexity measurement. The results show that optimal combination of statistical and nonlinear features with the Support Vector Machine (SVM) can achieve the apnea recognition accuracy of up to 75.51%. In the complexity measurement, t-test shows that the average of Hurst exponent, sample entropy or fractal dimension of the normal sleep group was significantly larger than that of the apnea group. The results show that long-term apnea syndrome will cause certain damage to the autonomic nerve system.