IEEE TRANSACTIONS ON BIOMEDICAL ENGINEERING, VOL. 50, NO. 7, JULY 2003 797 Very Low-Frequency Heart Rate Variability Wave Amplitude and Sympathetic Stimulation—Characterization and Modeling Ron Joseph Leor-Librach*, Member, IEEE, Sarah Eliash, Elieser Kaplinsky, and Ben-Zion Bobrovsky Abstract—The purpose of this study was to assess the relation- ship between very low-frequency heart rate variability (LFHR) wave amplitude and the degree of sympathetic stimulation. We de- veloped a computerized system for the controlled increase of heart rate (HR) by isoproterenol (ISP), with which we obtained a series of stabilized HR levels in conscious freely moving rats. We found that LFHR amplitude rises gradually as a function of the average HR for each level until it reaches a point where additional increases in average HR are associated with gradual decrease in LFHR ampli- tude. We successfully built and fitted a model of LFHR amplitude to the experimental results. The fact that our model fits the exper- imental data well may suggest a possible relationship between our LFHR amplitude findings and the basic physiologic properties of the HR-ISP system inherent in our model. Index Terms—Heart rate control, heart rate variability, isopro- terenol, physiological modeling. I. INTRODUCTION S PONTANEOUS heart rate variability (HRV) has been studied extensively in recent years in diverse clinical con- ditions [1]. HRV is a recognized measure of autonomic control and is used as a clinical tool for the diagnosis of autonomic neuropathy and for risk stratification after acute myocardial infarction [1]–[3]. Frequency analysis usually shows 3 spectral peaks. A very low-frequency heart rate and blood pressure variability peak at 0.003–0.04 Hz, which is related to vascular resistance [4], baroreflexes [5], [6] and chemoreceptors [7], [8]. A low-frequency heart rate variability peak at 0.04–0.15 Hz is related to baroreflexes and represents sympathetic and parasympathetic activity. The high-frequency peak at 0.15–0.4 Hz, is associated with respiration and represents parasympa- thetic activity [1]. Orthostatic tilt test [9]–[11] and mild effort [12] increase LFHR amplitude while maximal effort [12], and pronounced sympathetic stimulation [13], decrease LFHR amplitude. In one study the infusion of isoproterenol (ISP) caused an increase in LFHR amplitude that was ascribed to the nervous system response to vasodilation [14], while in another it caused a decrease in LFHR amplitude [9]. Manuscript received May 17, 2001; revised December 27, 2002. Asterisk in- dicates corresponding author. *R. J. Leor-Librach is with the Heart Institute, Laniado Hospital, Sanz Med- ical Center, Netanya, P.O. Box 744, Netanya 42107, Israel, and also with the Heart Institute Sheba Medical Center, Tel Hashomer, Ramat Gan 52621, Israel (e-mail: rellb@post.tau.ac.il). S. Eliash and E. Kaplinsky are with the Department of Physiology, Sackler School of Medicine, Tel Aviv University, Ramat Aviv, Tel Aviv 69978, Israel. B.-Z. Bobrovsky is with the Department of Electrical Engineering-Systems, Tel Aviv University, Ramat Aviv, Tel Aviv 69978, Israel. Digital Object Identifier 10.1109/TBME.2003.813547 Beta-blockade during orthostatic tilt test decreases LFHR amplitude [11]. These apparent contradictions suggest a possible dependency of LFHR amplitude on the degree of sympathetic activity. We developed a computerized system for the controlled in- crease of heart rate (HR), with which we can gradually obtain several levels of stabilized HR [15], [16]. Since the mean HR level in our experimental system was directly influenced by the rate of ISP infusion, it was used as an equivalent of the level of sympathetic activity. The purpose of this study was to use our computerized system for the precise characterization of the rela- tionship between LFHR amplitude and the level of sympathetic stimulation. In this manner we tried to resolve the inconsisten- cies between the results from previous studies that addressed this issue and gain more insight into the system. II. METHODS A. Experimental Protocol It is hard to predict the steady-state HR while using a con- stant flow intravenous infusion of medications. Furthermore, the pharmacodynamic response varies with time because of the emergence of tolerance, changes in metabolism and adverse re- actions. We therefore developed a computerized system for HR increase by closed loop control ISP infusion [15], [16]. This system enabled us to achieve well-separated and distinct levels of stabilized HR. All tests were performed on freely moving conscious rats, operated on under anesthesia at least 24 h prior to the test. Rats were anesthetized by intraperitoneal injection of Avertin, 200 mg/kg-(Tribromoethanol dissolved in tertiary amyl alcohol) [17]. The catheters were implanted according to the method of Chieuh and Kopin [18]. Blood pressure (BP) was measured from the catheter in the caudal artery, drug infusion was via a catheter in the jugular vein. ECG was measured from electrodes implanted subcutaneously. The C programming language was used for all computer sam- pling and control routines. A first routine continuously sampled the ECG and pressure signals through an A/D converter, at a frequency of 1.5 kHz. This routine calculated and stored the in- stantaneous HR and systolic and diastolic BP peaks. A second control routine was based on a modified proportional-integral (PI) controller [19]. Its inputs were the target heart rate (THR) that was entered through the keyboard of the computer and the running filtered HR value. HR was filtered by a causal moving averaging (MA) filter which suppressed heart rate variability disturbances to the 0018-9294/03$17.00 © 2003 IEEE