Nonlinear predictive interpolation. A new method for the correction of ectopic beats for heart rate variability analysis.

Heart rate variability (HRV) analysis is a technique that uses the beat-to-beat variations in RR intervals as a measure of the level of activity of the autonomic nervous system. However, the presence of ectopic beats can alter measures of HRV by introducing mathematical artifact and thus, prevent accurate determinations of HRV. Simple exclusion of those portions of data that contain ectopy from analysis inappropriate because (1) it can lead to a substantial reduction in the amount of data available for analysis and (2) if the presence (and frequency) of ectopic beats is correlated with specific alterations in autonomic tone, then the utilization of only ectopy-free data for analysis will lead to bias in the measures. For this reason, methods for the correction of ectopic beats have been devised and applied in the determination of HRV. The authors therefore propose a new method for the correction of ectopic beats: nonlinear predictive interpolation. Using the fact that beat-to-beat changes in heart rate occur in a deterministic fashion as the only assumption, the authors apply the methods of chaos theory in order to locate ectopy-free portions of the RR interval sequence that describe trajectories in phase space that are locally similar to that of the ectopy-containing segments. The authors then determine which of these trajectories most closely approximates that of a particular ectopy-containing segment, and use it to determine replacement RR intervals for the ectopic beats.