Quantification of the QT variability related to HRV: robustness study facing automatic delineation and noise on the ECG

In this work we evaluated the robustness of an automatic system to study the relation between HRV and QTV. RR and QT series were measured by an automatic delineator and a low order linear model was used to explore their short term interactions and to quantify the QTV fraction related to HRV simulated series and artificial ECG signals were used to quantify the method performance, considering noise contamination of SNR from 30 to 10 dB. The errors found in the estimation of the QTV fraction related to HRV point out a non relevant performance decrease resulting from automatic delineation. The joint performance of delineation plus variability analysis is satisfactory for records presenting SNRs over 20 dB.