AGE DEPENDENT CHANGES IN NON-LINEAR DYNAMICS OF THE HEART RATE VARIABILITY SIGNAL

E.L. Fallen, N.S. Salem, M. Kamath
Division of Cardiology and Department of Electrical and Computer Engineering,
McMaster University, Hamilton, Ontario, Canada

Abstract: It is known that aging affects both time domain and frequency domain components of heart rate variability.  To determine if these age related effects are due to loss of complexity, we analyzed several components of non-linear dynamics in 93 healthy human subjects across a wide age range from 5 years to 78 years. There was a continuous decay in the power law scaling (b) from -1.162 in the pediatric group (age 5-12 yrs) to -1.95 in the elderly group (>60 yrs); p<0.001. Approximate entropy decreased from 1.456 to 1.272 (p<0.001) and both the short term and long term detrended fluctuation exponent a₁ and a₂  (DFA) increased from 0.774 to 1.138 and from 0.667 to 0.86 respectively (p<0.001). These data indicate that nonlinear dynamic indices, especially power law scaling and approximate entropy,  decay with age while the DFA appears to decouple with age.  Aging progressively diminishes the complexity of integrative neurocardiac control.

INTRODUCTION

The frequency composition of the heart rate variability (HRV) signal has been used to measure autonomic efferent modulation of sinus node activity (1). It  is known that aging influences the different spectral components of the signal  but as a linear construct  the power spectrum of heart rate variability may not adequately reflect changes in non-linear dynamics. To determine whether complexity as measured by non-linear dynamical indices decays with age we developed several algorithms to characterize non-linear dynamics in the  HRV signal in a group of healthy subjects across a wide age range.

METHODS

Continuous HRV signals were computed from ECG recordings obtained under both controlled laboratory conditions and ambulatory Holter records from  93 healthy subjects (41 males and 52 females) ranging in age from 5 to 78 years. The minimal acceptable length of data for computation was 1000 R-R intervals. Data was analyzed  by four techniques:  Power law scaling;  Approximate entropy;  Short and long term detrended fluctuation analysis (DFA).

RESULTS

There was a continuous and systematic decay in power law scaling (b) from -1.162±0.388 in the pediatric group (age 5-12 yrs) to -1.95±0.6 in the elderly group (age >60 yrs); p<0.001. Approximate entropy decreased with age from 1.456±0.093 for the Pediatric group to 1.272±0.135 for the elderly; p<0.001. The short term DFA scaling exponent increased with age (0.774±0.204 to 1.138±0.289; p<0.001 and the long term DFA increased from 0.667±0.082 to 0.86±0.172 (p<0.001).

DISCUSSION

Decreased complexity of signals from physiologic processes represent the decoupling of integrative physiologic components (2). This loss of complexity may result in the isolation of individual biophysical components necessary for both short and long term regulation and adaptation of cardiovascular function. This maladaptation could signal susceptibility to disease. In the light of our findings it appears that the power law scaling is the best predictor of the effect of aging on cardiac autonomic regulation. Maladaptation is reflected in the trends for approximate entropy which also clearly declined with age. Our findings are consistenet with others in showing similar changes in the DFA with age (3,4). This loss of fractal organisation may reflect degradation of integrated physiologic systems with aging and less crossover between the components in the younger subjects. The latter suggests a balance between many different physiologic inputs operating over different time scales. Surrogate data analysis demonstrated that the hypothesis that the HRV signal is generated by a linear stochastic process is not always rejected.

REFERENCES

 [1] M.V. Kamath, E.L. Fallen. Power spectral analysis of heart rate variability: a noninvasive signature of cardiac autonomic function. Crit Rev Biomed Eng 21:245, 1993

[2]  A.L. Lipsitz, A.L. Goldberger. Loss of complexity and aging. JAMA 267:1806, 1992.

[3] C.K. Peng, S. Havlin, et al. Quantification of scaling exponents and crossover phenomena in nonstationary heart beat time series. Chaos 5:82, 1995.

[4] N. Iyengar, C. Peng et al. Age related alterations in the fractal scaling of cardiac interbeat interval dynamics. Am J Physiol 271:R1078, 1996.