Electrocardiogram (ECG) analysis of features such as the mean P-wave duration requires ECG preprocessing in the form of creating a single-beat template representative of the average heartbeat. ECG templates enable analysis of global ECG characteristics independent of inter-beat variability.
This work introduces two different template generation methods. The first one depends only on RR-intervals for extracting the single heartbeats (ECG snippets), therefore called RR-based approach. The second utilizes the information of P-wave onset and T-wave offset in the ECG snippet extraction and is therefore called PT-based approach. The quality of the templates generated with the two methods was visually assessed.
In cases of high heart rate, the RR-based approach yielded low-quality templates that comprised multiple heartbeats in contrast to single-beat templates in the PT-based approach. Moreover, the templates created with the PT-based approach had reduced isoelectric line length in the beginning and end of the signal, condensing the information of the template. Additionally, baseline wander artifacts were reduced in the PT-based approach compared to the RR-based approach. In conclusion, this work offers robust PT-based single-beat ECG template generation available as open source code.