© 2001 by European Society of Cardiology
Copyright © 2001, European Society of Cardiology
Autonomic indexes based on the analysis of heart rate variability: a view from the sinus node
aDipartimento di Biotecnologie e Bioscienze, Università di Milano-Bicocca, Piazza della Scienza 2, Room U3-3013, 20126 Milan, Italy
bCardiologia, Dipartimento di Medicina, Chirurgia e Odontoiatria, Ospedale San Paolo, Università di Milano, Milan, Italy
* Corresponding author. Tel.: +39-02-6448-3307; fax: +39-02-6448-3565 antonio.zaza{at}unimib.it
Received 4 October 2000; accepted 2 February 2001
 |
Abstract |
|---|
 Top Abstract 1 Introduction2 Relationship between neural...3 Pathophysiological...4 ConclusionsAppendix AAcknowledgmentsReferences |
|---|
Objective: Clinical indexes of autonomic activity are based on the analysis of sinus cycle length and of its variability. A common assumption underlying this practice is that neural activity and cycle length may be linearly related. Recent experimental evidence suggests that such an assumption may not be correct; indeed, the relation linking autonomic agonist concentration to cycle length was found to be strongly non-linear in single sinoatrial myocytes. In the present work, we (i) test compatibility of non-linearity of neural modulation of cycle length (CL) with previous experimental and clinical observations; (ii) evaluate its implications for what concerns the interpretation of time- and frequency-domain parameters of heart rate variability (HRV) and baroreflex sensitivity (BRS). Conclusions: Non-linearity of neural modulation of CL may result in an intrinsic rate-dependency of autonomic indexes, with the exception of normalised frequency-domain indexes (e.g. the low frequency/high frequency (LF/HF) ratio), which appear to be devoid of intrinsic rate-dependency. This characteristic may not limit the value of HRV indexes and BRS in risk stratification, but has to be taken into account in their pathophysiological interpretation.
KEYWORDS Autonomic nervous system; Baroreflex; Heart rate (variability)
|
1 Introduction |
|---|
|
TopAbstract1 Introduction2 Relationship between neural...3 Pathophysiological...4 ConclusionsAppendix AAcknowledgmentsReferences |
|---|
The concept that sinus node pacemaker activity is under control of the autonomic nervous system has promoted the use of different measurements of sinus node cycle length (CL) to evaluate autonomic modulation of cardiovascular function. Among them, heart rate variability (HRV) parameters and baroreflex sensitivity (BRS) indexes are the most accepted and well-known ones [1,2].
Although dominated by it, sinus CL is a surrogate measurement of autonomic neural activity. Thus, a correct pathophysiological interpretation of HRV and BRS requires knowledge of how neural input to the sinus node is transduced into CL changes. Whereas such a transduction process has been thus far considered as a linear one [3–5], recent findings, summarised in Fig. 1 [6], argue against this view, with potentially broad implications on the interpretation of autonomic indexes currently in use. In isolated pacemaker cells, steady-state superfusion with acetylcholine increased pacemaking cycle length (Fig. 1a) and its variability (Fig. 1b) concurrently. To explain this observation, we developed a numerical model (which we will refer to as ‘the model’) reconstructing changes in CL from the effect of acetylcholine on action potential parameters. By fitting the data with the model (solid lines in each panel of Fig. 1), we showed that changes in CL and in its variability could be simultaneously accounted for by acetylcholine effect on diastolic depolarisation rate (DDR; Fig. 1c). This was essentially due to non-linearity of the relation linking CL to diastolic depolarisation rate (Fig. 1d), which resulted in amplified variability at longer CL. In other words, CL and its variability may be intrinsically interdependent. The model accurately reproduces such a relation, and will be amply exploited in this work to achieve the following aims: (i) to confute the current assumption that CL may be a linear index of neural activity; (ii) to provide examples of how non-linearity of such a relation, even if hypothetical, is compatible with a number of clinical observations and might even explain some apparently puzzling findings; and (iii) to discuss whether and how the interpretation of currently used autonomic indexes may be confused by the assumption of linearity.
|
|
2 Relationship between neural activity and heart period |
|---|
|
TopAbstract1 Introduction2 Relationship between neural...3 Pathophysiological...4 ConclusionsAppendix AAcknowledgmentsReferences |
|---|
The sinus node can be viewed as a signal transducer which converts neural firing rate into changes of pacemaker cycle according to a specific input/output (I/O) relation. The transduction can be schematically divided into two sequential steps:
- 1. conversion of neuronal firing rate into neuromediator concentrations available at the cell membrane;
- 2. conversion of neuromediator concentrations into changes in pacemaker cycle length.
- 2. conversion of neuromediator concentrations into changes in pacemaker cycle length.
To the best of our knowledge, direct information on the relation between neural firing rate and neuromediator concentration at sinoatrial receptors (step 1) is not available and may be extremely difficult to obtain. Indeed, local neuromediator concentrations cannot be reliably measured; moreover, the number of variables involved (synapse recruitment; neuromediator release rate, clearance rate, size of the release store terminal, diffusion to distant receptors, etc.) is so large that heavy and unverifyable assumptions are required in modelling the process.
Rather surprisingly, the relation between neuromediator concentration and pacemaker cycle length (step 2) was, thus far, consistently assumed to be linear [3–5]. This assumption may be unwarranted; indeed, as demonstrated in single sinoatrial myocytes, cycle length is a non-linear function of neuromediator concentration [6]. The question then is whether strong non-linearity of transduction step 2 can be reconciled with previous findings suggesting a substantial linearity of the whole sinus node I/O relation [3,4,7]. To answer this question we tested whether such findings could be reproduced by modelling step 2 according to the non-linear relation detected in sinoatrial myocytes (Appendix A).
Fig. 2a shows the relation between vagal stimulation rate and heart period presented by Warner and Cox [3]. The authors attempted to interpolate the data by a relation in which acetylcholine concentration was a saturating function of vagal stimulation rate (step 1) and CL was assumed to be a linear function of acetylcholine concentration (step 2). The relation proposed by Warner and Cox (Fig. 2a, solid line), though adequate to account for small CL changes, failed to account for large ones. Fig. 2b shows that the full range of CL changes reported by Warner and Cox (dots) might be explained by model simulations considering non-linearity of step 2 (solid line; see figure legend for details). According to the non-linear model, CL dependency on acetylcholine concentration becomes quite steep at long CLs; thus, the relation between rate of vagal stimulation and acetylcholine concentration (step 1) required to fit the data is a saturating one (inset of Fig. 2b). It is fair to stress that the model proposed by Warner and Cox and the non-linear one suggested here are equally arbitrary; nonetheless, this analysis shows that the data presented by these authors can be explained with non-linearity of step 2.
|
Katona et al. recorded simultaneously vagal firing rate and CL in chloralose anesthetised dogs [4]. Fluctuations of CL in the interval between 300 and 500 ms could be reconstructed from the vagal firing rate through a linear function whose parameters were arbitrarily set. Interestingly, the authors also stated that the linear function was inadequate to predict larger CL changes (e.g. to 1 s). Fig. 2c shows modelling of an acetylcholine-induced CL fluctuation between 300 and 500 ms generated either by assuming the sinus node I/O relation to be entirely linear (solid line), or by introducing the non-linear step 2 (dots). The two curves are almost superimposed; this implies that for relatively small CL changes occurring at heart rates similar to those reported by Katona et al., linear dependency of CL on neural firing cannot be discriminated from non-linear one. Nonetheless, the two models would substantially diverge in the case of larger CL fluctuations (Fig. 2d).
To summarize, experimental observations commonly quoted to support the assumption of a linear relation between efferent cardiac neural discharge and heart period appear to be equally compatible with a non-linear one. Thus, even if direct evidence concerning the features of the whole transduction process may be difficult to obtain, the assumption of linearity seems largely unjustified and the potential consequences of a non-linear transduction process should be carefully considered.
In the next paragraphs we will assume that the non-linear relation between neuromediator concentration and CL (step 2) may dominate the whole transduction process. By using model simulations, we will discuss consistency of this assumption with clinical findings and evaluate its potential impact on the pathophysiological interpretation of autonomic indexes.
|
3 Pathophysiological interpretation of autonomic indexes |
|---|
|
TopAbstract1 Introduction2 Relationship between neural...3 Pathophysiological...4 ConclusionsAppendix AAcknowledgmentsReferences |
|---|
3.1 Time domain indexes of HRV
Time domain indexes of HRV are a measure of the dispersion of individual CLs around their mean. The standard deviation of normal (sinus) RR intervals (SDNN) is the simplest time domain index of HRV. Since the fluctuation in the AV conduction time is usually negligible [8], SDNN closely reflects the standard deviation of sinus CL. Other time domain indexes, such as the percentage of cycles exceeding by 50 ms the preceeding one (pNN50), may emphasise particular aspects of HRV, but are substantially similar and strongly correlated to SDNN [1].
A determinant support to the concept that time domain indexes of HRV and in particular SDNN might be considered indexes of autonomic modulation has derived from the interpretation of results observed in numerous clinical studies. In 1987, Kleiger and co-workers reported that a reduced SDNN of a 24-h recording has an independent negative predictive value after myocardial infarction [9]. The reduction of HRV in high risk patients was associated with a relatively faster heart rate with smaller day to night variations. Both findings concurred to generate the hypothesis that a reduced HRV could reflect an abnormal autonomic modulation of sinus node characterised by a sympathetic predominance and by a reduced vagal tone. Several studies subsequently confirmed the predictive value of a reduced HRV, thus supporting the importance of autonomic imbalance in the genesis of cardiovascular events. However, even early studies indicated that HRV was largely correlated with heart rate [10]; such a correlation has been thus far interpreted as a consequence of both indexes being determined by autonomic balance.
The presence of beat-to-beat variations in cardiac cycle even during resting controlled conditions is due to the fact that the sinus node is continuously subjected to time-varying neural and (to a lesser extent) mechanical influences, which may be considered as the primary cause of HRV in vivo. Indeed, spontaneous CL variability of single pacemaker cells can be largely eliminated by a sort of averaging process, due to electrotonic interactions among different cells [11]. Accordingly, changes in HRV are generally assumed to represent either variations in the pattern of discharge of autonomic nerves, or changes in the ability of pacemaker cells to respond to stimulation by neuromediators. The association between increased HRV and vagal predominance has been established by simple correlation. Indeed, large values of SDNN (or other indexes) are usually associated with relatively slow heart rate and other signs of prevailing parasympathetic activity. Nevertheless, the biological mechanism underlying such an association has never been directly demonstrated. Reasonable biological hypotheses may include: (i) acetylcholine mediated effects occur with a faster kinetics than those of catecholamines (faster receptor-effector coupling, faster quenching of synaptic signal, etc.) [12], thus having a larger potential for inducing variability; and (ii) strong sources of variability in neural discharge (e.g. respiratory activity) are mediated by parasympathetic efferent fibers [13]. As an additional factor, non-linearity of the pacemaker response to autonomic agonists may contribute to the correlation between large SDNN values and vagal predominance. Indeed, the latter is generally associated with bradycardia, which can amplify the effects on CL of any source of variability affecting the sinus node [6]. The rate-dependency of time domain indexes such as SDNN, CVNN and pNN50 on heart rate, as predicted by the model, is shown in Fig. 3a–c.
|
Tsuji et al. evaluated the determinants of heart rate variability with multivariant analysis applied to a large cohort of subjects from the Framingham study database [14]. Mean heart rate was found to be one of the two strongest determinants of SDNN and, when reported on a semilog scale (lnSDNN vs. HR), the relation between SDNN and heart rate could be interpolated by a linear function. The regression coefficient of such a function indicated that SDNN may decrease by 0.17 ms per each 10 beats/min increment in mean heart rate. The relation between mean heart rate and SDNN predicted by the model closely reflected that observed by Tsuji et al. and linear interpolation of its semilog plot yields a regression coefficient of –0.16 ms per 10 beats/min [6]. Thus, the correlation between SDNN and mean heart rate observed in man might essentially reflect rate-dependency of SDNN values.
The impact of sinus node transducer properties on the pathophysiological interpretation of SDNN measurements, may be illustrated by discussing the results of a recently published study on the ‘autonomic’ effects of the bradycardic agent zatebradine in conscious rats [15]. Zatebradine is a selective blocker of If [16], a current contributing to diastolic depolarisation in the sinus node [17]. Thus zatebradine is expected to prolong CL, by depressing diastolic depolarisation rate, independently of effects on the autonomic balance. Since SDNN and BRS values were both increased, and in spite of the unmodified low frequency/high frequency (LF/HF) ratio, the authors of this work concluded that zatebradine affects the autonomic balance. Although autonomic effects of zatebradine are theoretically possible, it is difficult to rule out that the observed changes were due to the direct effect of the drug on sinus node pacemaker activity. Also relevant to this point are the effects of dobutamine on autonomic indexes [18]. The drug-induced increase in arterial pressure, although expected to elicit a vagal reflex, was accompanied by a reduction in HRV and BRS. This apparently paradoxical finding might be explained by the drug-induced increase in heart rate, whose effects on autonomic indexes might override those of autonomic changes. Dobutamine also failed to affect the LF/HF ratio, which should be theoretically decreased by baroreflex activation. While this may not be ascribed to rate-dependency of this index (see below), the direct effects of dobutamine on the sinus node might still confound its response to the periodic influence of neural activity.
The examples in which concomitant changes in mean CL might complicate the interpretation of SDNN (or pNN50) changes as ‘autonomic effect’ are many. This applies particularly to studies evaluating drugs which may have direct effects on sinus pacemaking (e.g. Ca2+ or Na+ channel blockers); due to rate-dependency of the indexes, the latter might either mimic or mask changes in autonomic activity. The predicted impact of changes in the activation threshold of pacemaker cells on SDNN, as those expected from Ca2+ channel blockade, are illustrated for example in the inset of Fig. 3a. The interpretation of changes as an ‘autonomic’ effect may be more justified when the LF/HF ratio is also affected by the drug [19]. Nonetheless, it is still possible that such an effect may represent a change in the cellular response to autonomic agonists (at receptor or post-receptor levels), rather than a different pattern of neural activity.
Vagal predominance does usually produce concomitant decrease in mean heart rate and increase in variability; however, unless disproportionately large, the change in SDNN might be considered as an epiphenomenon of the change in mean heart rate. In this case, measurement of time domain indexes of HRV would not add to the information provided by the simpler assessment of mean heart rate. The need of correcting for rate-dependency of HRV indexes has been appreciated in some studies, in which HRV has been expressed as the coefficient of variation of RR intervals (CVNN). Unfortunately, CVNN may retain a significant rate-dependency (Fig. 3b) [6]; thus, it may also prove inadequate as an independent autonomic index.
It is fair to stress that what has been discussed thus far does not imply that changes in time domain HRV indexes are exclusively due to their intrinsic rate-dependency; indeed they can occur independently of heart rate. A striking example of this may be extreme depression of HRV observed in advanced stages of heart failure, far too large to be justified by the heart rate levels usually observed in this condition [1,20].
Irrespective of their pathophysiological interpretation, time-domain indexes of HRV have been indisputably shown to predict risk in a variety of cardiac diseases, including myocardial ischemia and heart failure. As mentioned above, this has been assumed to indicate the involvement of autonomic imbalance (with sympathetic predominance) in the pathogenesis of life threatening events, or in the progression of organ damage. Since sinus node pacemaker function may not be directly altered by the disease, such an assumption appears to be largely justified, and untouched by the above considerations.
3.2 Frequency-domain indexes of HRV
Neural activity causing HRV is periodic, with sympathetic and parasympathetic components oscillating at different frequencies. The purpose of the frequency domain analysis of HRV (spectral analysis) is to dissect HRV in its specific frequency components. This analysis defines the variance, often referred to as ‘power’, in absolute (ms2) or normalised units (power of individual spectral components divided by total power minus aperiodic components), contained in each frequency component. When considering short-term recordings obtained in stationary conditions, the HRV spectrum is characterised by three major components at high (HF), low (LF) and very low (VLF) frequency [1]. Extensive experimental and clinical evidence has shown that: (i) HF is synchronous with the respiratory cycle (respiratory sinus arrhythmia) and it is mainly supported by vagal activity; (ii) LF is largely correlated with sympathetic efferent activity; and (iii) VLF may sometimes be referred to an abnormal respiratory pattern, but more often, cannot be clearly attributed to a specific neural component [1]. When considering 24-h recordings, most of the power (up to 90%) is contained within the ultra- and very low-frequency range. Under this condition, LF and HF components account for less than 10% of total power, thus making their appraisal more difficult and unsuitable to assess autonomic modulation of sinus node.
Though limited information is available, it is interesting to consider how the periodicity detected in HRV may be related to periodicity in autonomic neural discharge. Neural recordings in cats [21] show that, when the frequency component synchronous with each cardiac cycle was filtered off, the two principal rhythmical components of HRV were also detectable in the pattern of discharge of sympathetic and vagal fibres. In addition, a reflex increase in sympathetic activity was associated with an increase not only in the absolute number of impulses per unit time, but also in the rhythmical component corresponding to LF. Similarly, a hemodynamic manoeuvre associated with reflex vagal activation determined not only an increase in the number of impulses, but also in the amplitude of the oscillations in neural discharge synchronous with HF [21]. These observations suggest that, in the case of frequency content, the transduction of neural activity into HRV may be largely linear. To test this hypothesis, the sum of two periodic (sin) components at frequencies corresponding to LF and HF was introduced, as an additional source of variability, in the model. Contamination by non-periodic components (i.e. deterministic variability lacking periodicity within the time interval considered), present in most of patient's spectra and referred to as VLF by some investigators [1], was not included in the simulation. Series of CL values, generated by the model at different mean heart rates, were analysed with autoregressive spectral techniques [22]. Examples of the spectra obtained at three heart rates are shown in Fig. 4. Total and single component powers, expressed in ms2, were steeply proportional to HR, whereas the LF/HF ratio, though oscillating in a narrow range, was not systematically affected by mean HR (Fig. 5). Thus, consistently with the observations discussed above, the frequency contents of neural input to the node and of CL series are expected to largely coincide, and the LF/HF ratio to be a rate-independent index of neural modulation.
|
2), and the LF/HF ratio are also reported.
|
It is important to stress that, as spectral power expressed in absolute units is steeply dependent on mean CL, the interpretation of its changes may suffer from the same ambiguities discussed for time-domain HRV indexes [23]. Thus, the fundamental parameters that can be derived from spectral analysis of RR interval series, although all related to autonomic control mechanisms, may carry different information. Normalised indexes, with particular reference to the LF/HF ratio, being independent of mean heart rate, appear to be a better index of sympatho-vagal balance. Their use appears to be essential if changes in autonomic balance are to be discriminated from variations in sinus node intrinsic pacemaker function (e.g. bradycardic agents, etc.). However, as all normalised indexes, LF/HF would not detect a loss of nodal responsiveness to neural inputs or a reduced variability of neural input to the node, as it may occur in end stage heart failure, diabetes and CNS disorders, etc. [1,20,23]. Total power (variance) may instead be highly valuable for evaluating such aspects, but dependency on mean heart rate must be taken into account when interpreting its changes.
3.3 Baroreflex sensitivity
Baroreflex sensitivity (BRS) is commonly measured as the extent of CL prolongation induced by pharmacological or mechanical modulation of carotid arterial pressure. Thus, also in the case of BRS measurements, CL is assumed to reflect changes in neural activity induced by baroreceptor activation, which we will refer to as ‘neural’ baroreceptor response. Animal studies [24] have led to a widespread use of BRS as an index of autonomic balance in the experimental and clinical settings [25,26]. This conception provided grounds for interpreting the increase in BRS induced by zatebradine as a shift toward prevailing parasympathetic activity [15]. Although such an interpretation cannot be conclusively ruled out, the lack of concomitant changes in the LF/HF ratio is certainly not consistent with it. Thus, the alternative hypothesis that the observed increase in BRS might simply be secondary to CL prolongation should be considered.
To test the extent to which BRS measurements might be affected by mean heart rate, the model was used to calculate the response of CL to a ‘neural’ baroreceptor response curve. The simulations were developed by considering the relation between carotid blood pressure and efferent vagal discharge rate (‘neural’ baroreceptor curve) as sigmoidal [27] and the one between DDR of sinus cells and cleft acetylcholine concentration as linear [6]. Since no information was available, we assumed efferent vagal discharge to be, at least in the interval of interest, linearly related to acetylcholine concentration at the effector synapse. Fig. 6 shows curves constructed from the same ‘neural’ baroreceptor curve, but starting from different mean heart rates. At a basal heart rate of 90 beats/min, the slope of the linear portion of the curve (BRS) was significantly smaller than at 70 beats/min. Thus, the same ‘neural’ baroreceptor response might produce different BRS values (estimated from CL changes) as a function of basal heart rate. This would clearly affect the interpretation of BRS changes whenever these are associated with changes in mean heart rate.
|
|
4 Conclusions |
|---|
|
TopAbstract1 Introduction2 Relationship between neural...3 Pathophysiological...4 ConclusionsAppendix AAcknowledgmentsReferences |
|---|
Due to its relation with the action potential parameters affected by acetylcholine and catecholamines, sinus CL (or RR interval) might depend on neural activity in a non-linear fashion. This would imply an intrinsic rate-dependency of time-domain HRV indexes and BRS, which would confuse their pathophysiological interpretation. Critical reappraisal of previous experimental and clinical observations is compatible with this hypothesis; moreover, the latter provides an interpretation for apparent inconsistencies in the behavior of autonomic indexes under specific conditions. Thus, the concepts proposed here should be carefully considered whenever sinus CL changes are used as a surrogate measurement of autonomic neural activity. The appreciation of the rate-dependency of BRS and time-domain HRV indexes does not detract from their value in risk stratification in the clinical setting, but it might suggest their potential redundancy with mean heart rate measurements. These limitations do not apply to LF/HF ratio, which may not be intrinsically dependent on heart rate.
It should be stressed that further complexity in the sinus node transduction process may arise from sensitivity of pacemaker discharge to the timing of pulsatile neural activity [28] and from functional inhomogeneity within the sinus node tissue [29]. Nonetheless, it seems unlikely that such complexity may obscure the influence of the pacemaker cell transducer properties, whose implications have been discussed here.
Time for primary review 21 days.
|
Appendix A |
|---|
|
TopAbstract1 Introduction2 Relationship between neural...3 Pathophysiological...4 ConclusionsAppendix AAcknowledgmentsReferences |
|---|
Numerical model of CL variability
Except for new details concerning simulations of periodic components, the following formulation of the model is identical to that described and validated in Ref. [6] and is reported here for practicality.
Series of n=1–512 consecutive cycles were constructed by the equation:
 |
APDn (action potential duration) was calculated from the equation describing dependency of APD on the duration of the preceeding diastolic interval (DI): APDn=0.112+0.0713*DIn–1. This equation was obtained by linear fitting of experimental data points (R=0.995, P<0.05).
Vth (the difference between maximum diastolic and threshold potentials) was assigned a value (–13 mV, except for Fig. 3a inset) constant through all cycles (independent of n). Thus, Vth did not contribute to random variability of CL.
DDRn (diastolic depolarisation rate) was obtained by adding a ±random value, different for each cycle, to a constant (‘DDR level’). Random values were fractions of ‘DDR level’ (DDR level times a random fractional number), as required by the observation that the coefficient of variation of DDR was unchanged by ACh [6]. In the simulations concerning the frequency-domain analysis (power spectra) diastolic depolarisation rate (DDR) was made to oscillate according to the sum of two (in phase) sin functions: LF=0.1 Hz (amplitude=1), HF=0.25 Hz (amplitude=0.5). Such oscillators were superimposed to random DDR variability.
To summarize, CL variability was determined in the model by (i) random+periodic beat-to-beat variations of DDR; (ii) beat-to-beat variations of APD, secondary to the changes in diastolic interval.
|
Acknowledgments |
|---|
|
TopAbstract1 Introduction2 Relationship between neural...3 Pathophysiological...4 ConclusionsAppendix AAcknowledgmentsReferences |
|---|
The authors are grateful to Drs Gabriella Malfatto and Marcella Rocchetti for providing insightful criticism and suggestions. This study has been partially supported by grant MURST COFIN 1997.
|
References |
|---|
|
TopAbstract1 Introduction2 Relationship between neural...3 Pathophysiological...4 ConclusionsAppendix AAcknowledgmentsReferences |
|---|
- 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.
[Free Full Text] - Smyth H.S., Sleight P., Pickering G.W. Reflex regulation of arterial pressure during sleep in man. A quantitative method of assessing baroreflex sensitivity. Circ Res (1969) 24:109–121.
[Abstract/Free Full Text] - Warner H.R., Cox A. A mathematical model of heart rate control by sympathetic and vagus efferent information. J Appl Physiol (1962) 17:349–355.
[Free Full Text] - Katona P.G., Poitras J.W., Barnett G.O., Terry B.S. Cardiac vagal efferent activity and heart period in the carotid sinus reflex. Am J Physiol (1970) 218:1030–1037.
[Free Full Text] - Ursino M. Interaction between carotid baroregulation and the pulsating heart: a mathematical model. Am J Physiol (1998) 275:H1733–H1747.[Web of Science][Medline]
- Rocchetti M., Malfatto G., Lombardi F., Zaza A. Role of the input/output relation of sinoatrial myocytes in cholinergic modulation of heart rate variability. J Cardiovasc Electrophys (2000) 11:522–530.[CrossRef][Web of Science][Medline]
- Eckberg D.L. Non-linearities of the human carotid baroreceptor-cardiac reflex. Circ Res (1980) 47:208–216.
[Abstract/Free Full Text] - Nollo G., Del G.M., Ravelli F., Disertori M. Evidence of low- and high-frequency oscillations in human AV interval variability: evaluation with spectral analysis. Am J Physiol (1994) 267:H1410–H1418.[Web of Science][Medline]
- Kleiger R.E., Miller J.P., Bigger J.T., Moss A.J. Multicenter Post-Infarction Research Group. Decreased heart rate variability and its association with increased mortality after acute myocardial infarction. Am J Cardiol (1987) 59:256–262.[CrossRef][Web of Science][Medline]
- Lombardi F. Heart rate variability: a contribution to a better understanding of the clinical role of heart rate. Eur Heart J (1999) 20:H44–H51.
- Jalife J. Mutual+entrainment and electrical coupling as mechanisms for synchronous firing of rabbit sino-atrial pace-maker cells. J Physiol (1984) 356:221–243.
[Abstract/Free Full Text] - Downing S.E. Handbook of physiology. Berne R.M., Sperelakis N., Geiger J.D., eds. (1979) Bethesda, MD: American Physiological Society. 621–652.
- Davidson N.S., Goldner S., McCloskey D.I. Respiratory modulation of baroreceptor and chemoreceptor reflexes affecting heart rate and cardiac vagal efferent nerve activity. J Physiol (1976) 259:523–530.
[Abstract/Free Full Text] - Tsuji H., Venditti F.J.J., Manders E.S., et al. Determinants of heart rate variability. J Am Coll Cardiol (1996) 28:1539–1546.[Abstract]
- Kruger C., Landerer V., Zugck C., Ehmke H., Kubler W., Haass M. The bradycardic agent zatebradine enhances baroreflex sensitivity and heart rate variability in rats early after myocardial infarction. Cardiovasc Res (2000) 45:900–912.
[Abstract/Free Full Text] - Goethals M., Raes A., van Bogaert P.-P. Use-dependent block of the pacemaker current If in rabbit sinoatrial node cells by zatebradine (UL-FS 49). On the mode of action of sinus node inhibitors. Circulation (1993) 88:2389–2401.
[Abstract/Free Full Text] - Zaza A., Micheletti M., Brioschi A., Rocchetti M. Ionic currents during sustained pacemaker activity in rabbit sino-atrial myocytes. J Physiol (1997) 505:677–688.
[Abstract/Free Full Text] - van Bogaert P.-P., Heron S., Nguyen H., et al. Arterial baroreflex control of the sinus node during dobutamine exercise stress testing. Hypertension (1999) 33:987–991.
[Abstract/Free Full Text] - Bonaduce D., Petretta M., Ianniciello A., Apicella C., Cavallaro V., Marciano F. Comparison of verapamil versus felodipine on heart rate variability after acute myocardial infarction. Am J Cardiol (1997) 79:564–569.[CrossRef][Web of Science][Medline]
- Lombardi F. Chaos theory, heart rate variability, and arrhythmic mortality. Circulation (2000) 101:8–10.
[Free Full Text] - Montano N., Lombardi F., Gnecchi R.T., et al. Spectral analysis of sympathetic discharge. R-R interval and systolic arterial pressure in decerebrate cats. J Auton Nerv Syst (1992) 40:21–31.[CrossRef][Web of Science][Medline]
- Baselli G., Cerutti S., Civardi S., et al. Heart rate variability signal processing: a quantitative approach as an aid to diagnosis in cardiovascular pathologies. Int J Bio-Med Comput (1987) 20:51–70.[CrossRef][Medline]
- Malliani A., Pagani M., Lombardi F., Cerutti S. Cardiovascular neural regulation explored in the frequency domain. Circulation (1991) 84:482–492.
[Abstract/Free Full Text] - Billman G.E., Schwartz P.J., Stone H.L. Baroreceptor reflex control of heart rate: a predictor of sudden cardiac death. Circulation (1982) 66:874–880.
[Free Full Text] - La Rovere M.T., Bigger J.T., Marcus F.I., Mortara A., Schwartz P.J. Baroreflex sensitivity and heart-rate variability in prediction of total cardiac mortality after myocardial infarction. Lancet (1998) 351:478–484.[CrossRef][Web of Science][Medline]
- Schwartz P.J. The autonomic nervous system and sudden death. Eur Heart J (1998) 19:F72–F80.[Web of Science][Medline]
- Pelletier C.L., Clement D.L., Shepherd J.T. Comparison of afferent activity of canine aortic and sinus nerves. Circ Res (1972) 31:557–568.
[Abstract/Free Full Text] - Brown G.L., Eccles J.C. The action of a single vagal volley on the rhythm of heart beat. J Physiol (1934) 82:242–257.
[Free Full Text] - MacKaay A.J.C., Opthof T., Bleeker W.K., Jongsma H.J., Bouman L.N. Interaction of adrenaline and acetylcholine on cardiac pacemaker function. J Pharmacol Exp Ther (1980) 214:417–422.
[Abstract/Free Full Text]
CiteULike 
Connotea 
Del.icio.us What's this?
This article has been cited by other articles:
|
|
 |
|
  A. Bauer, M. Malik, G. Schmidt, P. Barthel, H. Bonnemeier, I. Cygankiewicz, P. Guzik, F. Lombardi, A. Muller, A. Oto, et al. Heart Rate Turbulence: Standards of Measurement, Physiological Interpretation, and Clinical Use: International Society for Holter and Noninvasive Electrophysiology Consensus J. Am. Coll. Cardiol., October 21, 2008; 52(17): 1353 - 1365. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 G. Malfatto, G. Branzi, B. Riva, L. Sala, G. Leonetti, and M. Facchini Recovery of cardiac autonomic responsiveness with low-intensity physical training in patients with chronic heart failure Eur J Heart Fail, March 1, 2002; 4(2): 159 - 166. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||