Artificial intelligence for heart rate variability analyzing with arrhythmias

Introduction. Existing standards of Heart Rate Variability (HRV) technology limit its use to sinus rhythm. A small number of extrasystoles is allowed, if the device used has special procedures for the detection and replacement of ectopic complexes. However, it is important to expand the indicated limits of the applicability of the HRV technology. This specially regards the cases when the HRV technology looks promising in the diagnostics, as, for example, in atrial fibrillation and atrial flutter. Materials and Methods. All ECG measurements were performed on XAI-MEDICA® equipment and software. Processing of the obtained RR Series was carried out using the software Kubios® HRV Standard. All recommended HRV characteristics for Time-Domain, Frequency-Domain and Nonlinear were calculated. The purpose of the work. The article presents an artificial intelligence (AI) procedure for detecting episodes of arrhythmias and reconstruction of core patient’s rhythm, and demonstrates the efficacy of its use for the HRV analysis in patients with varying degrees of arrhythmias. The results of the study. It was shown efficiency of developed artificial intelligence procedure for HRV analyzing of patients with different level of arrhythmias. These were demonstrated for Time-Domain, Frequency-Domain and Nonlinear methods. The direct inclusion into review of Arrhythmia Episodes and the use of the initial RR Series leads to a significant distortion of the results of the HRV analysis for the whole set of methods and for all considered options for arrhythmia. Conclusion. High efficacy of operation of the procedure AI core rhythm extraction from initial RR Series for patients with arrhythmia was reported in all cases.