心拍変動時系列の極値の個数を用いた睡眠時呼吸周波数の簡易高速連続推定手法

抄録

It is reported that frequency component of approximately 0.25Hz of heart rate time series (RSA) is corresponding to the respiratory frequency. In this paper, we proposed that continuous estimation method of respiratory fequency during sleep using the number of extreme points of heart rate time series in real time. Equation for calculation of the method is very simple and the method can continuously calculate frequency by window width of about 18 beats. To evaluate accuracy of proposal method, RSA frequency was calculated using proposal method from the heart rate time series during supine rest. Result, minimum error rate was observed when RSA had time lag for about 11s and error rate was about 13.8%. Result of estimating RSA frequency time series during sleep, it varied regularly during non-REM and varied irregularly during REM. This result is similar as report of previous study about respiratory variability during sleep. Therefore, it is considered that proposal method possible to apply respiratory monitoring system during sleep.