Abstract
BACKGROUND: Phase variance analysis, which quantifies the spatial heterogeneity of local activation, has been proposed as a method to characterize complex conduction patterns in atrial fibrillation (AF). However, the influence of recording duration on its reliability remains unclear. This study aimed to determine how recording duration affects the accuracy and reproducibility of identifying heterogeneous conduction regions by phase variance analysis. METHODS: Intracardiac electrogram recordings (30 s) were obtained from 16 left atrial sites in 5 patients with persistent AF. A reference phase variance map (REF) was generated by averaging instantaneous phase variance maps over the full 30-s dataset to represent the frequency of local conduction heterogeneity. Thirteen recording durations (0.01-25 s) were defined as experimental conditions. For each duration, 100 randomly selected segments were extracted, and estimated phase variance maps (ESTs) were computed using the same temporal averaging method. Accuracy was evaluated by calculating the structural similarity index measure (SSIM) between ESTs and REF, while reproducibility was assessed using pixel-wise standard deviations across subsamplings. RESULTS: Longer recording durations produced higher SSIM values and lower standard deviations. With durations ≥ 10 s, SSIM consistently exceeded 0.95, including outliers, and standard deviations fell below 0.015, indicating accurate and stable estimation of complex conduction patterns. CONCLUSIONS: Adequate recording duration enables accurate and reproducible estimation of heterogeneous conduction regions using phase variance analysis, even in persistent AF with complex conduction patterns. This method may support optimal ablation target selection and improve treatment outcomes.