Twenty-four hour variation in heart rate variability indices derived from fractional differintegration.

Assuming that RR time-series behave as a fractionally differintegrated Gaussian process, García-González et al. (2003) recently proposed new indices for quantifying variability and structure in RR data. One of these was the 'fractional noise quantifier' (fnQ), measuring the departure of an RR time-series from a monofractal structure (i.e. a measure of its multifractality). Sixty-nine participants (aged = 34·5 ± 12·4 years, body mass index (BMI) = 23·9 ± 2·9 kg m(-2) , maximal oxygen uptake rate (V˙O2peak ) = 42·4 ± 10·9 ml min(-1)  kg(-1) , 39 males) provided continuous beat-to-beat ECG recordings for a 24-h period. Fractional differintegration was used to quantify fnQ, and heart rate variability was calculated in the time domain. All variables were evaluated during consecutive 1-h periods and also during four 6-h blocks corresponding to morning, afternoon, evening and night periods. Apart from RR, circadian trends in all variables were independent of gender (P = 0·11-0·59). Apart from fnQ, all variables exhibited circadian variation (0·0005<P<0·012). Although fnQ was statistically uniform during the 24-h period, it showed a trend towards elevated values during evening and night. The main finding of this study was that fnQ was elevated by around 10% during the evening and night, although this was not statistically significant. This suggests that the structure of RR time-series in healthy individuals is most strongly 'multifractal' during evening and night periods. fnQ appears to be a plausible surrogate measure of multifractality in RR time-series.