Abstract:Heart rate is one of the most easily accessed human physiological data. In recent years, the analysis of sleep function based on heart rate variability has become a new popular feature of wearable devices used for daily health management. Consequently, it is needed to explore various types of short-term characteristic parameters which can be applied to the heartbeat interval time series within the standard sleep staging time window (about 30 s). Utilizing the recently reported limited penetrable horizontal visibility graph (LPHVG) algorithm, together with a weighted limited penetrable horizontal visibility graph (WLPHVG) algorithm proposed in this paper, the short-term heartbeat interval time series in different sleep stages are mapped to networks respectively. Then, 6 characteristic parameters, including the average clustering coefficient C, the characteristic path length L, the clustering coefficient entropy E-c, the distance distribution entropy E-d, the weighted clustering coefficient entropy E-Cw and the weight distribution entropy E-w are calculated and analyzed. The results show that the values of these characteristic parameters are significantly different in the states of wakefulness, light sleep, deep sleep and rapid eye movement, especially in the case of the limited penetrable distance L-p = 1, indicating the effectiveness of LPHVG and WLPHVG algorithm in sleep staging based on short-term heartbeat interval time series. In addition, a preliminary comparison between proposed algorithm and the basic visibility graph (VG) algorithm shows that in this case, the LPHVG and WLPHVG algorithm are superior to the basic VG algorithm both in performance and in calculation speed. Meanwhile, based on the LPHVG and WLPHVG algorithm, the values of network parameters (the clustering coefficient entropy Ec and the weighted clustering coefficient entropy E-Cw) are calculated from heartbeat interval time series of healthy young and elder subjects in different sleep stages, to further study the aging effect on and sleep regulation over cardiac dynamics. It is found that despite an overall level difference between the values of E-c and E-Cw in young and elder groups, the stratification patterns across different sleep stages almost do not break down with advanced age, suggesting that the effect of sleep regulation on cardiac dynamics is significantly stronger than the effect of healthy aging. In addition, compared with the clustering coefficient entropy E-c based on LPHVG algorithm, the weighted clustering coefficient entropy E-Cw based on WLPHVG algorithm shows higher sensitivity to discriminating subtle differences in cardiac dynamics among different sleep states. Overall, it is shown that with the simple mapping criteria and low computational complexity, the proposed method could be used as a new auxiliary tool for sleep studies based on heart rate variability, and the corresponding network parameters could be used in wearable device as new auxiliary parameters for sleep staging.

Time Labeled Visibility Graph for Privacy-Preserved Physiological Time Series Classification

High-Performance Personalized Heartbeat Classification Model for Long-Term ECG Signal

Leveraging Visibility Graphs for Enhanced Arrhythmia Classification with Graph Convolutional Networks

CLPVG: Circular Limited Penetrable Visibility Graph As a New Network Model for Time Series

Assessment of time irreversibility in a time series using visibility graphs

Amplitude-Independent Machine Learning for PPG through Visibility Graphs and Transfer Learning

Complex Network Recognition of Electrocardiograph Signals in Health and Myocardial Infarction Patients Based on Multiplex Visibility Graph

Gated temporal convolutional neural network and expert features for diagnosing and explaining physiological time series: A case study on heart rates

Machine Learning on Visibility Graph Features Discriminates the Cognitive Event-Related Potentials of Patients with Early Alzheimer's Disease from Healthy Aging

Time-aware Context-Gated Graph Attention Network for Clinical Risk Prediction

Graph learning with label attention and hyperbolic embedding for temporal event prediction in healthcare

Adaptive Visibility Graph Neural Network and its Application in Modulation Classification

Graph Neural Network Based Multi-Label Hierarchical Classification for Disease Predictions in General Practice.

Research of Short-Term Heart Rate Variability During Sleep Based on Limited Penetrable Horizontal Visibility Graph

GRU-TV: Time- and velocity-aware GRU for patient representation on multivariate clinical time-series data

STGNet: Spatio-temporal graph neural networks considering inherent properties of physiological signals for camera-based remote photoplethysmography

Research on Healthy Anomaly Detection Model Based on Deep Learning from Multiple Time-Series Physiological Signals

Spatio-Temporal 3-Dimentional Variations Modeling with Self-Attention for Multi-Label ECG Classification

Conv-RGNN: An efficient Convolutional Residual Graph Neural Network for ECG classification

Visualization deep learning model for automatic arrhythmias classification

Concept-based AI interpretability in physiological time-series data: Example of abnormality detection in electroencephalography