AbstractHeart rate variability (HRV) refers to the physiological phenomenon of variation in heartbeat duration, which can be characterized using various metrics. Considering a complex systems approach, in this study we used network modeling to quantitatively evaluate the relationship between different HRV metrics during sleep. Polysomnography recordings were performed on 24 healthy participants and their cardiac activity was sampled from the N2, N3, and REM sleep stages. Fifty-eight HRV metrics were calculated, and the relationship between each was assessed using mutual information (MI). One network was created for each sleep stage; HRV metrics constituted its nodes, and MI values were used to establish its edges. Repeated measures ANOVA was applied to each metric to assess variation between sleep stages. It was found that all three networks had characteristics of complex networks. Several communities of shared similar metrics were found across the three sleep stages. Of these, one community had the same metrics in stages N2 and N3, but in REM sleep was divided into three communities. REM sleep exhibited significant differences compared to the other sleep stages in several metrics. These preliminary findings allow us to suggest the application of this method in other HRV research contexts, which will determine its scope and limitations.
Read full abstract