Dirk Muessig

Fast communication: A simple method to quantify the morphological similarity between signals

We propose a simple index, termed adaptive signed correlation index (ASCI), to quantify the morphological similarity between signals. The ASCI between two signals is calculated by trichotomizing each signal based on predefined three signal subspaces, then calculating the signed correlation of the trichotomized vectors. Examples are shown to compare ASCI with conventional correlation coefficients with respect to the effects of signal perturbation and additive noise. The ASCI provides a robust and efficient measure of morphological similarity and has particular applications in embedded systems involving biological signal analysis.

Biventricular capture verification by means of morphological analysis of intracardiac electrogram

AIMS The aim of this study is to develop a new feature for automatic biventricular capture verification in cardiac resynchronization therapy (CRT) devices, by means of morphological analysis of the intracardiac electrogram (IEGM). METHODS AND RESULTS The algorithm performs capture classification based on a novel adaptive signed correlation index (ASCI), which measures morphological similarity between the post-pace IEGM and a template waveform representing captured paces. To evaluate the performance of the algorithm, CRT pacemakers were implanted in six dogs. During a mean follow-up of 23 days, 175 biventricular threshold tests were conducted with various configurations of pace/sense polarities. Biventricular IEGMs were recorded and downloaded for offline analysis. Template signals for each pace/sense configuration in each chamber were created for individual dogs during the first follow-up. Each pace was annotated for capture or non-capture by visual examination of the IEGM. A total of 9991 capture paces and 4474 non-capture paces were included for morphological analysis. The calculated ASCI values were well separated for capture and non-capture paces irrespective of right/left pacing chambers, pace/sense configurations, pacing amplitude, individual dogs, and temporal proximity of the capture templates. Overall, the classification accuracy of the algorithm remained ≥99% for any ASCI cut-off value choosing between 0.18 and 0.52. CONCLUSION This study demonstrated the feasibility to perform automatic biventricular capture verification based on morphological analysis of the IEGM.

Risk assessment of R-on-T event based on modeled QT-RR relationship.

Background: R‐on‐T event is a well‐known trigger of ventricular tachycardia (VT) and ventricular fibrillation (VF). We propose a method to estimate the risk of R‐on‐T event from the inter‐beat (RR) intervals based on modeled QT‐RR relationship.