Dynamical cross-correlations between RR and QT intervals in long-term electrocardiogram recordings

Abstract

Understanding the dynamic relationship between RR and QT intervals is crucial for interpreting electrocardiograms (ECGs) and managing cardiac conditions. Here we focus specifically on the short-range, beat-to-beat coupling between RR and QT intervals, which we investigate in long-term ECG recordings from 202 healthy subjects using time-lagged cross-correlation (TLCC) analysis. This approach allows us to quantify how QT intervals relate to both preceding and succeeding RR intervals across a range of short time lags. Data were preprocessed with the smoothness priors method to remove long-term trends while preserving relevant short-term dynamics. Understanding the dynamic relationship between RR and QT intervals is crucial for interpreting electrocardiograms (ECGs) and managing cardiac conditions. We investigated cross-correlation between RR and QT intervals in long-term ECG recordings from 202 healthy subjects using time-lagged cross-correlation analysis to explore how QT intervals correlate with both preceding and succeeding RR intervals across various time lags. Data was preprocessed with the smoothness priors method to remove long-term trends while preserving relevant short-term dynamics. Our results reveal that, as expected, the highest RR-QT cross-correlations occur when the intervals overlap the most. Notably, we also observed a strong cross-correlation between the present QT interval and the preceding RR interval, with the overall correlations skewed toward negative lags. This suggests that past RR intervals exert a stronger influence on current QT intervals. These findings align with previous research on RR-QT transfer entropy. We found no significant differences in RR-QT cross-correlations between sexes. However, aging was associated with significantly reduced cross-correlations across all lags. This effect is consistent with existing literature on age-related cardiac changes and provides further insight into the impact of aging on RR-QT coupling. Additionally, we observed a slight positive correlation between heart rate and cross-correlation at all lags, with cross-correlation decreasing as the lag increased. These findings enhance our understanding of the dynamic RR-QT relationship by characterizing its short-range structure in a model-free way. The empirical patterns offer a reference for normal RR-QT dynamics in healthy individuals and may support future development or validation of dynamical QT/RR models.