TIME-DOMAIN SYSTEM FOR ASSESSING VARIATION IN HEART-RATE

The paper presents an alternative to Fourier techniques for assessment of heart rate variation. The system operates on R-R interval data (intervals between successive R-waves of an ECG). Artefacts are first detected and corrected by recombination of deviant intervals. Next, the highest frequency variation is recognised by identifying changes in the sign of the slope of successive R-R intervals as 'peaks' and 'troughs'. Lower frequency variation is determined from the variation in higher frequency peaks. The period of variation is computed as the sum of the intervals between two peaks, and extent of variation is represented by the peak-to-trough difference. Variation in a given frequency range is then quantified as the median extent of all variation within that range. Assessment of variation at the respiratory frequency by the peak/trough method is highly correlated with results derived from spectral techniques, while the two methods show consistently weaker correlations in the assessment of lower-frequency variation.