Relationship Between Spectral Components of Cardiovascular Variabilities and Direct Measures of Muscle Sympathetic Nerve Activity in Humans
Background Spectral analysis of RR interval and systolic arterial pressure variabilities may provide indirect markers of the balance between sympathetic and vagal cardiovascular control.
Methods and Results We examined the relationship between power spectral measurements of variabilities in RR interval, systolic arterial pressure, and muscle sympathetic nerve activity (MSNA) obtained by microneurography over a range of blood pressures. In eight healthy human volunteers, MSNA, RR interval, intra-arterial pressure, and respiration were measured during blood pressure reductions induced by nitroprusside and during blood pressure increases induced by phenylephrine. Both low-frequency (LF; 0.10±0.01 Hz) and high-frequency (HF; 0.23±0.01 Hz) components were detected in MSNA variability. Increasing levels of MSNA were associated with a shift of the spectral power toward its LF component. Decreasing levels of MSNA were associated with a shift of MSNA spectral power toward the HF component. Over the range of pressure changes, the LF component of MSNA variability was positively and tightly correlated with LF components of RR interval (in normalized units; P<10−6) and of systolic arterial pressure variability (both in millimeters of mercury squared and normalized units; P<5×10−5 and P<5×10−6, respectively). The HF component of MSNA variability was positively and tightly correlated with the HF component (in normalized units) of RR-interval variability (P<3×10−4) and of systolic arterial pressure variability (P<.01).
Conclusions During sympathetic activation in normal humans, there is a predominance in the LF oscillation of blood pressure, RR interval, and sympathetic nerve activity. During sympathetic inhibition, the HF component of cardiovascular variability predominates. This relationship is best seen when power spectral components are normalized for total power. Synchronous changes in the LF and HF rhythms of both RR interval and MSNA during different levels of sympathetic drive are suggestive of common central mechanisms governing both parasympathetic and sympathetic cardiovascular modulation.
Cardiovascular neural regulation is the integrated response to a continuous interaction of inhibitory and excitatory reflexes.1 In physiological conditions, there is a dynamic closed-loop interaction of these reflexes with rhythmic hemodynamic oscillations, such as those caused by respiratory and vasomotor activity. Thus, it has been proposed that power-spectrum analysis of the short-term fluctuations affecting heart period (RR interval)2 3 and arterial pressure4 may provide indices of neural regulation and, in particular, of the balance between sympathetic and parasympathetic cardiovascular modulation.3 4 5 6
An LF and an HF component were first reported for RR interval.3 Similar components have also been reported for arterial pressure,4 direct measurements of efferent MSNA in humans,7 8 efferent cardiac sympathetic activity in cats,9 and renal sympathetic nerve activity in rats.10 Accordingly, changes in these oscillations, examined over a range of autonomic drives, would provide important clues to their reliability as markers of autonomic regulation. We therefore obtained simultaneous measurements of MSNA, RR interval, intra-arterial pressure, and respiration during graded changes in arterial pressure in resting, spontaneously breathing human volunteers. The goals of the study were as follows:
1. To determine the effects of sympathetic excitation (by induced graded hypotension) and sympathetic inhibition (by induced graded hypertension) on the traditional indices of MSNA (bursts and amplitude of the discharge) and on the spectral components of MSNA variability;
2. To contrast the relationship between LF and HF components of MSNA variability with similar oscillations present in both RR interval and SAP variability; and, thus,
3. To define and characterize the link between time- and frequency-domain measurements of cardiovascular variabilities and direct measurements of sympathetic efferent activity in normal humans.
We studied eight normal human volunteers (seven male, one female) aged 31±7 years. All were nonsmokers and receiving no medication.
We recorded the ECG, intra-arterial pressure (from a catheter inserted into the brachial artery), CVP (measured in six subjects), respiration (using a pneumatic chest belt), and efferent MSNA measured directly using microneurography. MSNA was recorded from sympathetic nerve fascicles to muscle blood vessels in the peroneal nerve. This technique has been described extensively in previous studies.11 12 In brief, recordings were obtained by percutaneous insertion of tungsten microelectrodes into sympathetic fascicles in the peroneal nerve. The electrodes were connected to a preamplifier, and the nerve signal was fed through a band-pass filter and routed through an amplitude discriminator to a storage oscilloscope and loudspeaker. For recording and analysis, the filtered neural signal was fed through a resistance-capacitance integrating network to obtain a mean voltage display of the neural activity. Data were stored via an FM tape recorder (TEAC).
Subjects underwent measurements of baseline variables for a 10-minute period during which saline was infused at a rate of 0.15 mL/min. Then a stepwise infusion of either sodium nitroprusside or phenylephrine hydrochloride was begun. The order of infusion of drugs was chosen randomly. Nitroprusside was infused at rates of 0.3, 0.6, 1.2, and 1.8 μg·kg−1·min−1, while phenylephrine was infused at rates of 0.5 and 1.5 μg·kg−1·min−1. Each dose was infused for 10 minutes. On completion of infusion of each dose, saline was infused for 10 minutes to allow heart rate, blood pressure, and sympathetic activity to return to their initial resting levels. The aim of this protocol was to obtain changes in pressure of ≈10% of the resting control level and to avoid extremes of either hypotension or hypertension. Data were analyzed during the latter part of the 10-minute infusion period so as to obtain measurements during steady-state conditions.
Data were analyzed off-line with a 486-PC after analog-to-digital conversion at a rate of 600 Hz per channel by use of a 12-bit convertor (Gould).
The methodology and the software for data acquisition and spectral analysis have been described previously.3 4 5 In brief, a derivative-threshold algorithm provided the continuous series of RR intervals (tachogram) from the ECG signal. From the continuous arterial pressure signal, beat-by-beat systolic (systogram) and diastolic (diastogram) values were calculated, and the signal of respiratory activity was sampled once for every cardiac cycle.
Traditional analysis7 11 12 13 of the microneurographic discharge is based on visual computation of the rate (bursts/min) and amplitude of the bursts. In the present study, a digital algorithm was used to automatically detect bursts and compute amplitude. A burst in neural activity was recognized on the basis of a user-defined voltage and time threshold. For each individual sympathetic burst, the computer program provided the time of the occurrence and its amplitude (time×voltage area). Simulation studies showed that this permitted an accurate automatic computation of the average number of bursts in a time unit (minutes) and of their average amplitude (expressed in arbitrary units). In addition, the neurogram was provided through integration of the continuous MSNA signal, according to where each integral was performed over the time window between two consecutive diastolic values (Fig 1⇓) delimiting the i-th cardiac cycle of period t(i). Accordingly, this new series of variability measures of MSNA is synchronous with the other variability signals, such as tachogram and systogram, on a beat-by-beat basis.
All variability series were analyzed by means of autoregressive parametric spectral and cross-spectral algorithms3 14 that can automatically provide the number, center frequency, and associated power of each relevant oscillatory component. The very-low-frequency component (0.00 to 0.03 Hz) requiring specific algorithms and longer data series15 was not addressed in the present study and accordingly was considered as a DC component.4 The power is expressed both in absolute and in normalized units,16 which are obtained by dividing the power of each component by total variance from which the very-low-frequency component had been subtracted, and multiplying this value by 100.4 5 14
Cross-spectral analysis was performed by means of bivariate autoregressive identification14 and was used to compute a squared coherence function (ie, the square cross-spectrum amplitude normalized by the product of the spectra of the two signals). Coherence is a measure of the statistical link between two variability series at any given frequency and is expressed as a number between 1 and 0; only values ≥0.5 were considered significant.
Data are expressed as mean±SEM. Computation of nonparametric Theil regressions between different variables was performed with the use of Institution of Mathematical and Statistical Libraries and ad hoc routines. A value of P<.05 was considered significant.
Average data describing the effects of the graded intravenous infusion of vasoactive agents are depicted in Fig 2⇓. Intravenous nitroprusside produced progressive reductions of SAP (from 114±2 to 107±5 mm Hg), RR interval (from 985±42 to 739±18 ms), and CVP (from 7.2±0.4 to 2.3±0.6 mm Hg) (all P<.01), whereas phenylephrine increased SAP, RR interval, and CVP (to 120±2 mm Hg, 1083±79 ms, and 9.5±0.8 mm Hg, respectively; all P<.01). Graded hypotension (Fig 3⇓) and hypertension were accompanied by progressive increases and decreases in measures of sympathetic nerve traffic, respectively. There was a significant negative correlation between absolute values of SAP and both burst frequency and amplitude of MSNA (Table 1⇓).
Spectral analysis of MSNA variability (Fig 4⇓) demonstrated the presence of two major oscillatory components at ≈0.1±0.01 Hz and at the respiratory frequency (0.23±0.01 Hz) in resting conditions. As expected, two similar components were observed in the autospectra of both RR-interval and SAP variability.
During nitroprusside-induced hypotension, an increased LF component was observed in all the variability signals, ie, MSNA, RR-interval, and SAP variabilities (Fig 4⇑). Conversely, a relative reduction of the LF (and attendant relative increase of the HF) was observed during phenylephrine-induced hypertension. During the experimental protocol, respiratory frequency did not change significantly from the resting value of 0.24±0.01 Hz.
Correlation of Absolute SAP With Spectral Components of RR Interval and MSNA
SAP also showed a negative correlation with LFMSNA in absolute arbitrary units, a tighter negative correlation with LFMSNA in normalized units, and a positive correlation with HFMSNA in normalized units (Fig 5⇓; Table 1⇑). No correlation was present between SAP and absolute measures of HFMSNA. In other words, a decrease in SAP was accompanied by both an increase in total MSNA and a redistribution of MSNA spectral power toward the LF end of the spectrum; this was especially apparent when spectral powers were normalized for total power.
Correlation of MSNA Amplitude and Frequency With Spectral Components of MSNA
Table 2⇓ shows the link between changes in MSNA as bursts/min and amplitude and spectral measurements of MSNA. Amplitude correlated positively with absolute LFMSNA and also with LFMSNA normalized units and LF/HFMSNA ratio, and negatively with HFMSNA normalized units. Measurements of burst frequency were poorly correlated with spectral measures of MSNA variability.
A significant relationship was found between measures of LFRR and LFSAP spectral components and MSNA amplitude (Fig 6⇑; Tables 1⇑ and 3⇓). However, the tightest correlation was observed between LF oscillations of MSNA (in normalized units) and LF oscillations of both SAP and RR-interval variability.
Cross-spectral analysis results (Fig 7⇑) indicated a constant link between LF and HF oscillations of MSNA and corresponding components of either RR interval or SAP variabilities. Lower coherence values were observed only during the infusion of the high dose of phenylephrine because MSNA was markedly or totally suppressed.
Correlations between LF/HFMSNA and measures of RR-interval variability were similar to those observed for LFMSNA in normalized units. The important influence of respiration on circulatory variables was confirmed by the significance of the coherence function (K2>0.5) between respiration and RR interval, arterial pressure, and MSNA variability at HF.
In normal resting volunteers, we increased arterial pressure with phenylephrine so as to elicit vagal activation and sympathetic inhibition. The increased blood pressure resulted in the expected reflex reductions in heart rate and efferent sympathetic activity.17 Conversely, when we decreased blood pressure with nitroprusside, reflex vagal inhibition and sympathetic activation resulted in tachycardia and increased levels of MSNA.
The major new findings of the present study include the following:
1. Increasing levels of MSNA were associated with a shift of the MSNA spectral power distribution toward a predominance of the LF component; decreasing levels of MSNA were associated with decreased spectral power in the LF region and a relative increase in the HF component.
2. Over the range of arterial pressure changes, the LF component of MSNA variability (in absolute and in normalized units) was tightly correlated with LF components of RR-interval (in normalized units) and SAP variability.
3. The HF component of MSNA variability in normalized units was tightly correlated with the HF component (in normalized units) of RR-interval and SAP variability.
4. The use of a normalization procedure is most effective in evaluating physiological changes in MSNA spectral components.
Neural control of the circulation is coded simultaneously in two different modalities, as amplitude (strength of signal or tonic activity) and frequency (oscillatory or phasic activity). Our data indicate that the changes in tonic activity are accompanied by tightly linked modulations in oscillatory characteristics. These findings apply to small perturbations within the physiological range in free-breathing normal humans within the same session. Although there is a correlation between tonic and oscillatory properties, the relationship is complex, and our data do not imply that the frequency composition of an oscillatory signal can be equated with the strength of that signal. In addition, our findings cannot be extrapolated to comparisons between individuals or to extreme conditions such as heavy exercise or severe heart failure.18
Relationship Between MSNA and MSNA Variability
Over the limited range of pressure changes that were explored, the absolute values (in arbitrary units squared) of spectral components of MSNA correlated with measures of nerve activity7 11 12 13 such as burst frequency and amplitude. The strong positive correlation between MSNA and the normalized LF component of MSNA variability indicate that reflexly induced increases in average levels of MSNA are accompanied by a shift of the spectral power of MSNA variability toward the LF region. The HF component of MSNA, expressed in normalized units, was instead negatively correlated with absolute nerve activity, indicating that increases in sympathetic nerve traffic were associated with a decrease in the normalized power of the respiratory modulation of MSNA.
Increased sympathetic activity was associated with a decrease in the normalized power of the HF components of RR interval and SAP as well. The close link between HFMSNA normalized units and HFRR normalized units, a marker of parasympathetic modulation of the sinoatrial node,4 5 is suggestive of common central control mechanisms for respiratory modulation of autonomic activity.19
Relationship Between MSNA and RR-Interval Variabilities
The simultaneous evaluation of spectral measures of MSNA and RR-interval variability over graded autonomic changes allowed us to directly address the issue of which spectral measures of RR-interval variability better reflect changes in autonomic state. Our results show that the use of normalized units or the LF/HF ratio4 of RR-interval variability provides the strongest correlations with attendant changes in MSNA (particularly if assessed by its amplitude or spectral components rather than bursts/min). Blood pressure changes elicit qualitatively similar changes in sympathetic drive to the heart and blood vessels. RR-interval variance (RR ς2) and absolute LFRR, however, were not linked to any measure of MSNA, thus confirming the limited value of these absolute measures of RR-interval variability as indices of autonomic modulation of the sinoatrial node.20 21 In fact, in the case of the absolute values of LFRR, they undergo discordant influences during sympathetic excitation: they tend to be decreased by the reduction of variance, but they also tend to be increased by the greater concentration of residual power in the LF component as reflected by its rise in normalized units. The absolute HF component of RR-interval variability nevertheless correlated with various measures of sympathetic drive, albeit to a lesser degree than its normalized value, suggesting a potential role for sympathetic modulation of the HF component of RR-interval variability.22
Our findings support a prior study that used only measurements of bursts/min of MSNA and showed a weak positive correlation between fractional LF power of RR interval and absolute MSNA during nitroprusside-induced hypotension and not during phenylephrine-induced hypertension.8 The present study differs from the initial report evaluating the relationship between MSNA and RR-interval variability8 because not only changes in burst frequency of MSNA but also total amplitude and MSNA variability were examined in relationship to changes in either RR-interval or SAP variabilities.
Relationship Between MSNA and SAP Variabilities
Despite blood pressure being altered by an exogenous agent, oscillation characteristics of MSNA were closely mirrored by oscillations of SAP. Sympathetic activation was associated with an increase of the LF component of both MSNA and SAP (particularly evident with millimeters of mercury squared), thus providing more direct support for the concept of using changes in the LF of SAP variability as a marker of changes in sympathetic efferent activity to the peripheral vasculature.4 5
Increased blood pressure, with consequent sympathetic inhibition, induced progressive increases in the normalized (but not absolute) HFMSNA, indicating that a predominantly parasympathetic state elicits a preponderance in the normalized HF component of not only RR-interval but also MSNA variability.
Central Regulatory Mechanisms
Distinct HF and LF components have also been detected in the discharge variability of brain stem neurons involved in the regulation of cardiovascular function.23 It is conceivable that two main rhythms, one a marker of excitation and one of inhibition,24 25 26 might be modulated in a reciprocal manner that can be detected from the discharge of central neurons and in peripheral neural outflows and that is reflected by target function variability,27 according to a closed-loop organization. In this way, the relative balance between oscillations over the spectral range may serve as a marker of functional states. Changes in the central excitatory-inhibitory balance might induce reciprocal changes not only in average nerve activity of central vagal and sympathetic motor neurons but also in the balance between LF and HF rhythms that can be observed peripherally. Spectral analysis of MSNA, using normalized units, thus provides a unique window to explore directly the changing dynamics of central regulatory rhythmic activity.
These data also have implications for the understanding of central regulation of both vagal and sympathetic activity. Standish et al,28 using transsynaptic retroviruses injected into the heart, obtained evidence of interconnections between vagal and sympathetic pathways at a supraspinal level. Retroviral labels traveled not only to neurons of the nucleus ambiguus, the dorsal motor nucleus of the vagus, and the nucleus of tractus solitarius but also to many brain stem structures such as the caudal and rostral ventrolateral medulla, which constitute a major source of sympathetic efferent activity. Their studies provided structural evidence for close interactions between vagal and sympathetic nuclei at a central level.
Our results indicate a compelling relationship between changes in autonomic drive and cardiovascular oscillations. The balance between LF and HF components of cardiovascular variability tracks closely the changes in directly measured peripheral sympathetic activity. During stress consistent with parasympathetic activation, the increased HF component and decreased LF component in not only RR interval but also MSNA suggests an intimate interaction between parasympathetic and sympathetic neural oscillatory structures. The coherence between HF and LF fluctuations in MSNA and RR interval and the persistence of this synchrony of rhythms across a range of arterial pressure perturbations provide functional evidence to support the concept of common central mechanisms governing sympathetic and parasympathetic rhythmic activity.
Selected Abbreviations and Acronyms
|CVP||=||central venous pressure|
|MSNA||=||muscle sympathetic nerve activity|
|SAP||=||systolic arterial pressure|
These studies were supported by the National Institutes of Health (NIH-HL 14388), an NIH Sleep Academic Award, the Council for Tobacco Research, and an American Heart Association Grant-in-Aid (Dr Somers). We thank Mary P. Clary for technical assistance and Dr Giuseppe S. Mela for statistical advice.
- Received June 5, 1996.
- Revision received October 23, 1996.
- Accepted November 12, 1996.
- Copyright © 1997 by American Heart Association
Malliani A, Pagani M, Lombardi F. Positive feedback reflexes. In: Zanchetti A, Tarazi RC, eds. Handbook of Hypertension: Pathophysiology of Hypertension. Amsterdam, Netherlands: Elsevier Science Publishing Co Inc; 1986;8:69-81.
Akselrod S, Gordon D, Hubel FA, Shannon DC, Barger AC, Cohen RJ. Power spectrum analysis of heart rate variability: a quantitative probe of beat-to-beat cardiovascular control. Science. 1981;213:220-222.
Brovelli M, Baselli G, Cerutti S, Guzzetti S, Liberati D, Lombardi F, Malliani A, Pagani M, Pizzinelli P. Computerized analysis for an experimental validation of neurophysiological models of heart rate control. Comput Cardiol. 1983;205-208.
Pagani M, Lombardi F, Guzzetti S, Rimoldi O, Furlan R, Pizzinelli P, Sandrone G, Malfatto G, Dell’Orto S, Piccaluga E, Turiel M, Baselli G, Cerutti S, Malliani A. Power spectral analysis of heart rate and arterial pressure variability as a marker of sympatho-vagal interactions in man and conscious dog. Circ Res. 1986;59:178-193.
Malliani A, Pagani M, Lombardi F, Cerutti S. Cardiovascular neural regulation explored in the frequency domain. Circulation. 1991;84:482-492.
Saul JP, Rea RF, Eckberg DL, Berger RD, Cohen RJ. Heart rate and muscle sympathetic nerve variability during reflex changes of autonomic activity. Am J Physiol. 1990;258:H713-H721.
Brown DR, Brown LV, Patwardhan A, Randall DC. Sympathetic activity and blood pressure are tightly coupled at 0.4 Hz in conscious rats. Am J Physiol. 1994;267:R1378-R1384.
Wallin G. Intraneural recording and autonomic function in man. In: Bannister R, ed. Autonomic Failure. London, UK: Oxford University Press; 1983:36-51.
Mark AL, Victor RG, Nerhed C, Wallin BG. Microneurographic studies of the mechanisms of sympathetic nerve responses to static exercise in humans. Circ Res. 1985;57:461-469.
Saul JP, Albrecht P, Berger RD, Cohen RJ. Analysis of long term heart rate variability: methods, 1/f scaling and implications. Comput Cardiol. 1988;419-422.
Task Force of the European Society of Cardiology and the North American Society of Pacing and Electrophysiology. Heart rate variability: standards of measurement, physiological interpretation, and clinical use. Circulation. 1996;93:1043-1065.
Abboud FM, Thames MD. Interaction of cardiovascular reflexes in circulatory control. In: Shepherd JT, Abboud FM, Geiger SR, eds. Handbook of Physiology: The Cardiovascular System III. Bethesda, Md: American Physiological Society; 1983:675-753.
van de Borne P, Montano N, Pagani M, Oren R, Somers VK. Absence of low frequency variability of sympathetic nerve activity in severe heart failure. Circulation. 1997;95:1449-1454.
Koepchen HP. History of studies and concepts of blood pressure waves. In: Miyakawa K, Koepchen KP, Polosa C, eds. Mechanisms of Blood Pressure Waves. Tokyo, Japan/Berlin, Germany: Japan Science Society/Springer-Verlag; 1984:3-23.
Montano N, Gnecchi-Ruscone T, Porta A, Lombardi F, Pagani M, Malliani A. Power spectrum analysis of heart rate variability to assess the changes in sympatho-vagal balance during graded orthostatic tilt. Circulation. 1994;90:1826-1831.
Rimoldi O, Pierini S, Ferrari A, Cerutti S, Pagani M, Malliani A. Analysis of short term oscillations of RR and arterial pressure in conscious dogs. Am J Physiol. 1990;258:H967-H976.
Hess WR. Diencephalon: Autonomic and Extrapyramidal Functions. New York, NY: Grune & Stratton; 1954.
Malliani A. Association of heart rate variability components with physiological regulatory mechanisms. In: Malik M, Camm JA, eds. Heart Rate Variability. Armonk, NY: Futura Publishing Co Inc; 1995:173-188.
Standish A, Enquist LW, Escardo JA, Schwaber JS. Central neuronal circuit innervating the rat heart defined by transneuronal transport of pseudorabies virus. J Neurosci. 1995;15:1998-2012.