Abstract
Introduction: An analysis of Heart Rate Variability (HRV) is widely used in clinical and research. To properly calculate HRV parameters, the RR intervals series must be properly preprocessed. There are already automated tools for correcting these artifacts; however, they are not fully transparent and fully customizable. To address these limitations, we introduce CoRRection, a semi-automatic tool for RR interval correction, which integrates both automatic and manual approaches providing greater control over which intervals are corrected.
Material and Methods: The application offers detection and correction methods. Additionally, it allows to manually remove an artifact before applying a correction method. It also allows to clean artifacts in shorter segments which enables applying different detection methods in one example, e.g., gathered during intense exercise. The application provides a test to assess signal stationarity (Augmented Dickey-Fuller test). After signal processing the report is created, containing the number of probes deleted and modified.
Results: The proposed CoRRection application allows both technical and non-technical users to preprocess RR signals with a better control over the process by enabling the use of semi-automatic approach. A few examples of studies which required Correction’s unique approach of semi-automatic correction were presented. The need for different methods and not only considering different examinations, but also different segments within one recording was emphasized.
Conclusions: We have presented a new application designed to preprocess RR intervals signal with a few popular detection and correction methods. This approach enables a good compromise between a precise manual approach and a less time-consuming automatic approach. The semi-automatic method of artifact correction empowers users to explore multiple identification and correction methods (and to choose the one best fitted to the data or measurement conditions), and offers a better understanding of the examination preprocessing tools. This may also result in a better trust in the automatic approach among the users.