Author_Institution :
Fac. of Phys., Warsaw Univ. of Technol., Warsaw, Poland
Abstract :
Purpose: The subject of our research was the analysis of human heart rate, using the recently published method MMA. The main goal was an attempt to obtain a correct diagnosis, based mainly on the results of MMA, which we used as a screening examination method.Materials and methods: We analyzed 38 heart rate variability nighttime recordings of healthy patients and 236 recordings of ill patients in four groups: 103 patients with aortic valve stenosis, 36 patients with hypertrophic cardiomyopathy, 15 patients with atrial fibrillation and 82 patients with cardiac arrest. We applied MMA - method developed at our lab, describing the scaling properties of fluctuations as a function of the multifractal parameter q and the scale s. The end result of the MMA is the Hurst surface h(q,s), where h is the local Hurst exponent, q is value offluctuation and s is series length. We prepared 6 criteria quantifying mainly the local shape of the surface. The criteria were intended as a screening examination method and allow us to classify patients as healthy, when all of the criteria are fulfilled or ill, when at least one criterion was negative. Results: In order to check reliability of applied method and defined criteria, we calculated measures of diagnostic test as sensitivity, specificity, positive predictive value and negative predictive value were respectively as follows: for patients with aortic valve stenosis: 81%, 74%, 89%, 58%, hypertrophic cardiomyopathy: 47%, 74%, 63%, 60%, atrialfibrillation: 100%, 74%, 60%, 100% and for patients with cardiac arrest: 69%, 74%, 85%, 53%. Conclusion: These results show that analysis of human heart rate based on MMA is promising. However, we believe that this method still requires improvement and a lot of tests in order to obtain higher values of measures of diagnosis accuracy.
Keywords :
bioelectric potentials; cardiology; diseases; fluctuations; patient diagnosis; Hurst exponent; Hurst surface; aortic valve stenosis; atrial fibrillation; cardiac arrest; diagnosis accuracy; diagnostic test; fluctuation scaling property; heart rate variability; human heart rate analysis; hypertrophic cardiomyopathy; multifractal parameter; multiscale multifractal analysis; screening examination method; Accuracy; Fractals; Heart rate variability; Physiology; Time series analysis; Valves;