International Journal of Computer Applications (0975 – 8887) Volume 14– No.5, January 2011 40 The Effect of Aging on Nonlinearity and Stochastic Nature of Heart Rate Variability Signal Computed using Delay Vector Variance Method Srinivas Kuntamalla Department of Physics National Institute of Technology Warangal, INDIA - 506004 L. Ram Gopal Reddy Department of Physics National Institute of Technology Warangal, INDIA - 506004 ABSTRACT Heart rate variability analysis is fast gaining acceptance as a potential non-invasive means of autonomic nervous system assessment in research as well as clinical domains. In this study, a new nonlinear analysis method is used to detect age related changes in the degree of nonlinearity and stochastic nature of heart rate variability signals. The data obtained from an online and widely used public database (i.e., MIT/BIH physionet database), of young and elderly subjects is used in this study. The method used is the delay vector variance (DVV) method, which is a unified method for detecting the presence of determinism and nonlinearity in a time series and is based upon the examination of local predictability of a signal. From the results it is clear that there is no significant change in the minimum target variance values for young and elderly subjects and also the values are very small, which indicates that there is a strong deterministic component over the stochastic one in both the groups. There is a significant decrease in the degree of nonlinearity from younger to elder subjects (p- value, 0.0002). This indicates that there is no change in the stochastic or deterministic nature of the signals but there is a considerable change in the degree of nonlinearity with aging. General Terms Nonlinear analysis, Heart rate variability. Keywords Nonlinearity, stochastic nature, heart rate variability, delay vector variance 1. INTRODUCTION Heart rate variability (HRV) analysis is gaining acceptance as a potential non-invasive means of autonomic nervous system assessment in research as well as clinical domains. Usually HRV data is obtained from R peaks of electrocardiogram (ECG). The heart beat instants are taken at these points and consequently the beat to beat intervals are determined as the time interval from one R peak to the next one. Therefore, these intervals are called R-R intervals and are plotted against their beat number, which is called a tachogram (Figure 1). In 1996, the Taskforce of the ESC/NASPE published standards in HRV analysis proposing several time and frequency parameters based on short-term (5- min) and long-term (24-h) HRV data [1]. Although HRV has been the subject of many clinical studies investigating a wide spectrum of cardiological and non-cardiological diseases and clinical conditions, a general consensus of the practical use of HRV in medicine has been reached only in two clinical scenarios: depressed HRV can be used as a predictor of risk after acute myocardial infarction and as an early warning sign of diabetic neuropathy. The HRV can be analyzed using several methods which are broadly classified as time domain and frequency domain methods. Time domain measures are simple statistical operations on R-R intervals, such as standard deviation of normal R-R intervals (SDNN), root mean square of successive R-R interval differences (RMSSD) and the percentage change of normal R-R intervals that differ by > 50 ms (PNN50) etc. Frequency domain analysis includes FFT or AR based power spectral density measures which provide information on how variance distributes as a function of frequency. Three main spectral components are distinguished in a spectrum calculated from short-term recordings in both absolute and normalized units: very low frequency (VLF) (≤0.04 Hz), Low frequency (LF) (0.04 - 0.15 Hz), and high frequency (HF) (0.15 – 0.4 Hz) components. Heart rate dynamics are nonlinear in nature and it is proved that nonlinear analysis of it provides more appropriate information for the physiological interpretation of heart rate variability [2]. However, variety of contradictory reports in this domain indicates that there is a need for a more rigorous investigation of methods. The nonlinear analysis of HRV is a valuable tool in both clinical practice and physiological research reflecting the ability of the cardiovascular system [3]. Poincare plot, Approximate Entropy, Sample Entropy, Detrended Fluctuation Analysis, Correlation Dimension and Sequential trend analysis are some of the nonlinear analysis methods of HRV. In recent years, as the presence of nonlinearity and determinism in a biomedical signal is used as an index for risk stratification in many diseases [4], detecting the nature of physiological time series received large attention. Physiological time series are generated by complex systems for which it is not possible to solve or even set up the equations governing the dynamics, and generally assumed that such time series significantly display nonlinearity. There are two types of methods available for testing the nonlinearity in a time series [5] (i) fitting a linear or nonlinear model to the time series and their accuracies are evaluated (ii) comparing nonlinearity measures computed for the signal with those computed for linearised versions of the signal (surrogates) The delay vector variance (DVV) is a unified method for detecting the presence of determinism and nonlinearity in a time series and is based upon the examination of local predictability of a signal. Further, it spans the complete range of local linear