• DocumentCode
    3748992
  • Title

    Fractal pattern of heart rate variability revealing unknown very low frequency properties

  • Author

    Dorota Kokosi?ska;Jan Jakub Giera?towski;Jan Jacek ?ebrowski

  • Author_Institution
    Faculty of Physics, Warsaw University of Technology, Poland
  • fYear
    2015
  • Firstpage
    597
  • Lastpage
    600
  • Abstract
    The subject of our research was an analysis of heart rate variability based on non-linear method Multiscale Multifractal Analysis (MMA). The analysis of HRV night-time recordings involved 5 groups of patients (285 subjects): 38 healthy patients, 103 with aortic valve stenosis, 36 with hypertrophic cardiomyopathy, 32 with atrial fibrillation, 59 coronary disease and 17 with congestive heart failure. The end result of MMA is the Hurst surface h(q,s) measuring a persistence. This is the 3D plot of the local Hurst exponent h describing the scaling of the variance of the signal as a function of q, which is a magnitude of the fluctuations and the parameter s - a measure of the time scale (convertible to frequency). In our research, we assessed the shape and form of the Hurst surface and based on the differences of these features we constructed 6 criteria. These criteria intended as a diagnostic tool for screening examination, allow to classify patients as healthy when all the criteria were fulfilled or ill when at least one criterion was negative. We also prepared an additional criterion, distinguishing group of patients with atrial fibrillation and detecting heart rate variability pattern for this group. In general for all of groups (285 patients), we obtained 76% of correct results (i.e. the accuracy). The percent of correct results for coronary disease: 70%, for hypertrophic cardiomyopathy: 61% patients, for atrial fibrillation: 86%, for aortic valve stenosis: 79% and 80% - for congestive heart failure patients. These results allow us to draw a conclusion that Multiscale Multifractal Analysis can be used as an effective screening examination method (general 6 criteria) as well as there are clear heart rate variability multifractal pattern, which we can detect using our criteria. This multifractal pattern can reflect a variety of physio-pathological processes, which at this stage of our research is not able to specify, so need further research.
  • Keywords
    "Fractals","Valves","Heart rate variability","Time measurement","Physiology"
  • Publisher
    ieee
  • Conference_Titel
    Computing in Cardiology Conference (CinC), 2015
  • ISSN
    2325-8861
  • Print_ISBN
    978-1-5090-0685-4
  • Electronic_ISBN
    2325-887X
  • Type

    conf

  • DOI
    10.1109/CIC.2015.7410981
  • Filename
    7410981