Estimation of Sleep Transition using Quantitative Evaluation of Lorenz Plot

Abstract

In the present study, it is intended to helpfully estimate the depth of sleep during all night with RR intervals (RBIs) in order to monitor a daily life sleep in real-time. Analysis of electrocardiogram's RBIs was used a Lorenz plot (LP). To evaluate quantitatively the changes in these distributions, RRIs are projected a LP on a y = x axis, y = -x axis, and considered as an ellipse which was composed of a long axis σ(x), a short axis σ(-x). We calculated the shifting of the mean (center C) and standard deviation (area S) for each sleep levels. Its analysis periods for 60sec were between n-60th and nth and shifted for every 1 second of data. As a result, center C showed the process toward light sleep level, area S showed the transition phase toward deep sleep level. Sum differences of Center C and area S were set each the thresholds and estimated transition phases of sleep level using the algorithm. A mean of concordance rate was 60.9% between estimated values and measured values which were actual transition of sleep levels. Therefore, the transitional sleep level can be evaluated based on heart rate variability by using LP.