A novel geometric heart rate variability mapping method for enhanced bradycardia prediction in preterm infants

Abstract

Background

Geometric representation of heart rate variability (HRV) evaluates the autonomic nervous system (ANS) and the cardiovascular system. However, geometric features have drawbacks, especially for short-length signals and events. This study aimed to evaluate the potential of existing geometric features in infant bradycardia analysis. We also proposed a new mapping method for HRV analysis.

Methods

Data were analyzed for 10 preterm infants with a post-conceptional age of 293/7 to 342/7 weeks. The subjects' electrocardiogram (ECG) was analyzed before and during bradycardia. Standard Poincare Plot Parameters, Triangle Phase Space Mapping (TPSM), Parabolic Phase Space Mapping (PPSM), Triple Geometric features, and the proposed mapping method were compared.

Results

The proposed method achieved the best results, with sensitivity, specificity, and accuracy of 94.37±4.22 %, 95.12±4.06 %, and 95.28±3.84 %, respectively. These results show the robustness of θ to predict bradycardia. The worse result was related to the A parameter of PPSM, with sensitivity, specificity, and accuracy of 64.34±8.12, 71.28±7.66, and 73.23±7.49, respectively.

Conclusion

This study demonstrates the potential of geometric features in infant bradycardia analysis. The proposed mapping method shows robustness in predicting bradycardia with high sensitivity, specificity, and accuracy. The findings of this study have implications for the use of HRV analysis in evaluating the autonomic nervous and cardiovascular systems in preterm infants.