Evaluation of Time- and Frequency-Domain Features of ECG Signals

Proceedings of ‏The 3rd International Conference on Applied Research in Engineering, Science and Technology

Year: 2020

DOI:

[Fulltext PDF]

Evaluation of Time- and Frequency-Domain Features of ECG Signals

Evgeniya Gospodinova, Galya Georgieva-Tsaneva, Mitko Gospodinov, Diana Dimitrova

 

ABSTRACT: 

Heart rate variability (HRV) is one of the most studied, informative, non-invasive, easily applicable and promising methods for analyzing and evaluating the state of the autonomic nervous system. This method is based on the study of the RR intervals obtained from digital ECG signals. The purpose of the article is to show the results of the study of two groups of subjects: 21 controls and 21 patients with arrhythmia by applying Time-Domain and Frequency-Domain methods. The Time-Domain method includes statistical and geometrical measurements. Statistical measurements use two types of variables to evaluate the data examined. The first type of variables is related to the duration of normal RR intervals (SDNN, SDANN and SDNN index) and  and the second type of variables is related to the difference in the duration of the adjacent RR intervals (NN50, pNN50, RMSSD). Geometric measurements (HRVTi, TINN) allow a graphical representation of the distribution of RR intervals. These measurements are less affected by the quality of the recorded data and can be considered as an alternative to the statistical parameters. The frequency domain analysis shows the periodic oscillations of the RR series in the context of different frequencies and was performed using the following two methods: Fast Fourier Transform (FFT) and the autoregressive (AR) method. The results obtained show that there are a significant statistical difference between the study groups and HRV is decreased in the patients with arrhythmia. Time-Domain and Frequency-Domain methods are standardized, with the limits of norm-pathology being known, making they the preferred.

Keywords: Heart Rate Variability, Time-Domain analysis, Frequency-Domain analysis, ECG signal.