Skip main navigation

Role of Respiratory Motor Output in Within-Breath Modulation of Muscle Sympathetic Nerve Activity in Humans

Originally published Research. 1999;85:457–469


    Abstract—We measured muscle sympathetic nerve activity (MSNA, peroneal microneurography) in 5 healthy humans under conditions of matched tidal volume, breathing frequency, and end-tidal CO2, but varying respiratory motor output as follows: (1) passive positive pressure mechanical ventilation, (2) voluntary hyperventilation, (3) assisted mechanical ventilation that required the subject to generate –2.5 cm H2O to trigger each positive pressure breath, and (4) added inspiratory resistance. Spectral analyses showed marked respiratory periodicities in MSNA; however, the amplitude of the peak power was not changed with changing inspiratory effort. Time domain analyses showed that maximum MSNA always occurred at end expiration (25% to 30% of total activity) and minimum activity at end inspiration (2% to 3% of total activity), and the amplitude of the variation was not different among conditions despite marked changes in respiratory motor output. Furthermore, qualitative changes in intrathoracic pressure were without influence on the respiratory modulation of MSNA. In all conditions, within-breath changes in MSNA were inversely related to small changes in diastolic pressure (1 to 3 mm Hg), suggesting that respiratory rhythmicity in MSNA was secondary to loading/unloading of carotid sinus baroreceptors. Furthermore, at any given diastolic pressure, within-breath MSNA varied inversely with lung volume, demonstrating an additional influence of lung inflation feedback on sympathetic discharge. Our data provide evidence against a significant effect of respiratory motor output on the within-breath modulation of MSNA and suggest that feedback from baroreceptors and pulmonary stretch receptors are the dominant determinants of the respiratory modulation of MSNA in the intact human.

    The independent influence of the respiratory rhythm generator on the timing of sympathetic outflow has been studied using reduced preparations in which peripheral reflex mechanisms were eliminated by vagotomy and/or sinoaortic denervation. In these deafferented animal preparations, sympathetic neurons fire mainly during inspiration (ie, in synchrony with phrenic discharge), with their minimum activity occurring during expiration.12 The close temporal relationship between phrenic discharge and sympathetic activity recorded in anesthetized, paralyzed, and vagotomized animals has led to the hypothesis that the 2 neural outputs either arise from the same brain stem neurons or are driven by a common oscillator.13 However, in the intact human, sympathetic outflow to skeletal muscle (muscle sympathetic nerve activity; MSNA) declines during inspiration, reaching its nadir at end inspiration/early expiration, and then rises, reaching its peak at end expiration.45 The influence of central respiratory motor output on the within-breath modulation of MSNA has not been studied in humans.

    The purpose of the present study was to examine the relative contribution of central respiratory drive in generating the respiratory modulation of MSNA in intact human subjects. We varied respiratory motor output using passive positive pressure mechanical hyperventilation at high tidal volume (Vt) to eliminate respiratory motor output, and voluntary elevation of Vt and inspiratory flow rate, with and without added inspiratory resistance, to increase respiratory motor output. We also determined the effects of intrathoracic pressure on the modulation of MSNA and cardiac frequency and controlled for the effects of lung inflation and chemoreceptor reflexes.

    Materials and Methods

    General Procedures

    Six men ages 27 to 59 years served as subjects after providing written, informed consent. All subjects were normotensive and free from cardiopulmonary disease. All experimental procedures and protocols were approved by the University of Wisconsin Center for Health Sciences and the Middleton Memorial Veterans Hospital Human Research Review Committees. Airflow rates, Vt, mask pressure (Pm), and end-tidal partial pressure of CO2 (PETco2) were measured using equipment and techniques described previously.6 A single-lead ECG was continuously recorded. Diaphragmatic electromyogram (EMGdi) was obtained from surface electrodes placed over the sixth and seventh intercostal spaces in the anterior axillary line. The raw EMGdi signal was amplified and band-pass filtered from 30 to 1000 Hz (model P511, Grass Instruments). Arterial pressure was measured beat by beat using the finger photoplethysmography technique (Finapres model 2300, Ohmeda) and with an automated sphygmomanometer (Dinamap, Critikon) at 1-minute intervals.

    Recordings of Sympathetic Nerve Activity

    Multiunit recordings of postganglionic MSNA were obtained from the peroneal nerve of the right leg as described previously.78 The neural signals were passed to a differential preamplifier, an amplifier, a band-pass filter (700 to 2000 Hz), and an integrator (time constant=100 ms; total gain=100 000). When acceptable MSNA recordings (spontaneous pulse synchronous activity with signal-to-noise ratio >3:1) were obtained, the subject was instructed to maintain the leg in a relaxed position for the duration of the study. Segments of the neural recording that showed evidence of mechanoreceptor or α-motoneuron activity were excluded from the analysis. MSNA data from one subject were discarded because of inadequate neural recordings.

    Experimental Protocols

    Subjects were studied supine in the postabsorptive state. All protocols were performed during the application of a nonhypotensive level (applied pressure, <20 mm Hg) of lower body negative pressure (LBNP) to augment basal MSNA,9 thereby facilitating the study of respiratory modulation. In a single subject, who was studied both with and without lower body suction, LBNP increased MSNA frequency compared with normal resting conditions (35.4±1.5 versus 26.4±2.6 bursts per minute). However, in agreement with previous reports,5 we saw that the respiratory modulation of MSNA under all experimental conditions was unaffected by LBNP.

    The experimental protocols used are summarized in Table 1. These different protocols were designed to study the effects of increasing or decreasing central respiratory motor output on the within-breath modulation of MSNA and to control for the effects of changing lung volume, intrathoracic pressure, and systemic arterial pressure.

    Voluntary Hyperventilation

    The subject voluntarily maintained Vt at twice the eupneic level, using visual feedback from an oscilloscope. The subject maintained a constant respiratory frequency (f) and a duty cycle [inspiratory time (Ti)/total time (TTOT)] of 50% using auditory feedback with distinct inspiratory and expiratory tones. Vt, f, and Ti/TTOT were held at the same level for each of the subsequent experimental protocols.

    Passive Positive Pressure Mechanical Ventilation (Passive PPV)

    PPV was applied until the respiratory muscles were inhibited and was then continued for a minimum of 5 minutes during which blood pressure, ECG, and nerve activity were continuously recorded. Respiratory muscle inhibition was determined from specific criteria, as previously outlined,10 including the absence of EMGdi, stabilization of the Pm waveform, constancy of peak positive end-inspiratory Pm, and a significant (>10-second) prolongation of expiratory time on cessation of mechanical ventilation (Figure 1). Subjects visited the laboratory on a separate occasion for a familiarization session so that they were able to relax during the passive mechanical ventilator trials on the test day.

    Assisted Positive Pressure Mechanical Ventilation (Assisted PPV)

    Each subject had to generate a threshold Pm of –2.5 cm H2O to trigger each positive pressure mechanical breath. The subjects were asked to maintain the set f by following computer-generated audio signals. The aim of this protocol was to produce positive intrathoracic pressure changes during inspiration and at the same time require a high amount of respiratory muscle activation, as determined by the EMGdi.

    Inspiratory Resistor

    In 3 subjects, the voluntary hyperventilation trials were repeated with the addition of an inspiratory resistive load.

    Before each series of voluntary hyperventilation, passive PPV, and assisted PPV protocols, baseline measurements of ventilatory and cardiovascular variable data were made during 5 minutes of spontaneous breathing. These measurements of MSNA made during spontaneous breathing (22.7±4.7 and 22.6±3.1 bursts per minute, respectively; P>0.05) showed that baseline nerve traffic did not change over time. The voluntary hyperventilation trial served as a “control” protocol, during which all potential mechanical and neural influences on the respiratory modulation of MSNA were present. The sequence of the trials was as follows: (1) passive PPV, (2) voluntary hyperventilation, (3) assisted PPV, (4) repeat passive PPV, (5) repeat voluntary hyperventilation, and (6) inspiratory resistor. This sequence ensured that each experimental protocol was bracketed by a voluntary hyperventilation trial with which any within-breath and per-minute changes in MSNA were compared. This design ensured that potential baseline shifts and/or time-dependent changes in MSNA did not affect our results.

    Data Analysis

    Respiratory and cardiovascular parameters were computed as described previously.11 Bursts of MSNA were identified by visual inspection of the integrated neurogram by one investigator (C.M.S.). The amplitude of each burst of MSNA and the beat-to-beat levels of arterial blood pressure were determined by computer. MSNA was expressed both as burst frequency and total activity (calculated as the product of burst frequency and relative burst amplitude and presented in arbitrary units).

    Time Domain Analyses

    The blood pressure signal was advanced by 0.2 seconds to correct for the propagation time from the central circulation to the finger on the basis of previous studies (B. Morgan, A. Xie, unpublished data, 1998) of the delay between the peak of the R-wave from the ECG signal and the corresponding peak in the blood pressure waveform. Bursts of MSNA were also advanced in time to account for nerve conduction delays using the subject’s height and an estimate of conduction velocity in the peroneal nerve of 1.11 m/sec.12 The exact locations and amplitudes of the bursts of MSNA and the arterial pressure waves within the breath cycle were determined for each breath in each protocol both as a function of time and breath volume. Both burst frequency and total MSNA (frequency×relative amplitude) in each 12.5% time interval from the onset of inspiration to end expiration were then normalized (as a percentage of the number of bursts and total MSNA for each breath and averaged over the entire condition). Arterial pressure was expressed as the change from the last quartile of expiration (end expiration). To statistically test for differences in the within-breath patterns of modulation of MSNA among the 4 conditions (spontaneous breathing, voluntary hyperventilation, passive PPV, and assisted PPV), principal component analysis13 on the covariance matrix of the time series data were used to identify independent patterns that contributed to the composite pattern represented by the raw data points. A principal component was considered statistically significant if it explained >5% of the total observed variance. This technique computes both the common shape and the individual relative importance of each significant pattern in the MSNA in each subject, under each condition.

    Total minute values for MSNA, heart rate, and blood pressure were calculated during all protocols and compared using 1-way ANOVA with repeated measures. Post hoc analyses were performed if differences were detected. Significance was set at P<0.05.

    Frequency Domain Analyses

    Power spectral densities were obtained from analyses of 16-second epochs of arterial pressure, respiration (airflow), and sympathetic nerve activity. The power spectra for the R-R interval data were obtained using previously described methods.11


    Differences in Respiratory Motor Output

    Figure 1 and Table 2 show the different conditions across which respiratory motor output was varied. Note that with voluntary hyperventilation, the 2-fold increase in Vt above spontaneous eupnea was accompanied by increased EMGdi, more negative swings in Pm, and reduced PETco2. This is contrasted with passive mechanical hyperventilation with similar Vt, duty cycle, and PETco2, but with positive Pm and no discernable EMGdi during mechanical ventilation and with a postventilator apnea. The assisted PPV condition matched the Vt and positive Pm during the passive mechanical ventilation condition, only at high levels of respiratory motor output, as evidenced by the EMGdi. Finally, the increased inspiratory resistance trial matched the Vt and breathing pattern of voluntary hyperventilation, only at exaggerated levels of respiratory motor output (see EMGdi) and negative Pm. Average heart rate, systolic pressure, and diastolic pressure over the entire condition were not different among any of the conditions studied (Table 2). However, there was a small but significant (P<0.05) increase in mean MSNA (burst frequency×burst amplitude) for the 5 subjects during passive PPV (range, 19 to 38 arbitrary units/min) when compared with spontaneous breathing (range, 19 to 32 arbitrary units/min), voluntary hyperventilation (range, 19 to 33 arbitrary units/min), or assisted PPV (range, 18 to 33 arbitrary units/min). Nevertheless, there was no significant difference (P>0.05) in burst frequency among any of the 4 conditions (Table 2).

    Within-Breath Modulation of Arterial Pressure and MSNA: Frequency Domain Analyses

    During spontaneous eupneic breathing, power spectral analyses of MSNA and diastolic pressure identified 2 distinct peaks in power at frequencies corresponding to the respiration and heart period in all subjects. Representative spectra from 1 subject are presented in Figure 2. In all 5 subjects, the amplitude of the peak power in MSNA at the respiratory frequency was increased during voluntary hyperventilation (Vt=1.2 L) when compared with eupneic breathing (Vt=0.6 L). However, there were no further changes in the amplitude of the peak power of MSNA at either the respiratory or cardiac frequency, as respiratory motor output was varied from 0 (passive PPV) to very high (added inspiratory resistance), and Pm was changed from negative to positive, at the same Vt and frequency.

    Within-Breath Modulation of Arterial Pressure and MSNA: Time Domain Analyses

    Data for MSNA are presented only in terms of total activity (burst frequency×relative burst amplitude), because the pattern and amplitude of the within-breath variations in burst frequency and total activity were the same under all conditions. During spontaneous breathing at low Vt, when respiratory motor output was low, a within-breath variation of MSNA was evident in all subjects (Figure 3A), with peak activity occurring at end expiration (mean, 22.7±7.5% of total within-breath activity) and minimum activity (mean, 4.2±3.8% of total activity) at end inspiration. The amplitude of this respiratory modulation of MSNA was accentuated during voluntary hyperventilation, when both Vt and respiratory motor output were increased and Pm was more negative (Figure 3B). In all 5 subjects, MSNA fell progressively from onset to late inspiration, with the nadir at end inspiration (mean, 2.2±1.6% of total activity), and then rose sharply to a peak at end expiration (mean, 25.5±6.8% of total activity). There was some variability among subjects, with 2 of the 5 showing peak activity slightly earlier in expiration.

    Respiratory motor output was eliminated during passive PPV (Figure 3C), but the within-breath pattern of MSNA was qualitatively similar to that observed during active ventilation at the same Vt. However, during passive PPV, the maximum activity (31.4±10.2% of total activity) occurred at a slightly earlier quartile in expiration, rather than end expiration. The variability in MSNA among subjects was also more pronounced during the ventilator trials, with one subject showing a pattern that was markedly different from the remaining 4 during both passive and assisted mechanical ventilation. Assisted mechanical ventilation with positive pressure during inspiration and high respiratory motor output showed a similar within-breath modulation of MSNA, as did passive mechanical ventilation at equal Vt and positive Pm, but no respiratory motor output.

    Three subjects completed the resistor trials. Vt, f, Ti/TTOT, and PETco2 were the same as during the voluntary hyperventilation trials, but inspiratory effort was greatly augmented, as shown by the 7-fold decrease in the nadir for inspiratory Pm (Table 2) and the much larger EMGdi (Figure 1). The respiratory modulation of MSNA was qualitatively the same during both voluntary hyperventilation (Figure 4A) and voluntary breathing, with increased inspiratory resistance (Figure 4B). In all 3 subjects, MSNA decreased from onset to late inspiration and increased during expiration, reaching a peak at end expiration/early inspiration.

    Changes in diastolic pressure during inspiration and expiration for spontaneous breathing, voluntary hyperventilation, passive PPV, and assisted mechanical ventilation are shown in Figure 3. The within-breath pattern of change in systolic, mean arterial, and pulse pressures (not shown) were the same as those shown for diastolic pressure under all conditions. On average, during inspiration, arterial pressure increased 1 to 2 mm Hg (relative to end expiration) and decreased during expiration under all 4 conditions, with considerable variability among subjects. The within-breath pattern of change in diastolic pressure shown during voluntary hyperventilation (Figure 4A) was more pronounced during resistor breathing (Figure 4B), with diastolic pressure increasing by >4 mm Hg during inspiration in 2 of the 3 subjects. The third subject showed a continual increase in diastolic pressure from early inspiration to late expiration.

    Principal Component Analysis

    Principal component analysis of both MSNA and diastolic pressure for each of the 4 conditions revealed 2 distinct patterns, as follows: (1) a basic underlying pattern that was similar to the mean responses pictured in Figure 3, and (2) an asymmetry component reflecting the variability among subjects. When combined, these 2 components explained >90% of the total variance in each variable, for each condition. Marked differences in respiratory motor output among the 4 conditions (spontaneous eupnea, voluntary hyperventilation, passive PPV, and assisted PPV) had no effect on either the shape or the relative importance of the individual significant patterns. For MSNA, the basic pattern accounted for 70% to 88% (asymmetry, 10% to 29%) of the total variance. The asymmetry component was relatively more important in describing the average arterial pressure pattern, reflecting the greater variability among subjects in this variable, explaining 28% to 39% of the total variance, with the basic pattern explaining the remaining 47% to 61%. Neither of the 2 principal components describing the within-breath pattern of change in either MSNA, nor in diastolic pressure, were significantly different among the 4 conditions (P>0.05). Cross-correlation of the principal components describing the within-breath patterns of variation in MSNA and diastolic pressure revealed a significant inverse relationship between diastolic pressure and MSNA under all conditions with correlation coefficients of –0.75 (spontaneous), –0.89 (voluntary hyperventilation), –0.87 (passive PPV), and –0.72 (assisted PPV).

    Effects of Arterial Pressure Versus Lung Volume on Within-Breath Modulation of MSNA

    To “control for” within-breath changes in diastolic pressure and to examine the effects of changing lung volume on the respiratory modulation of MSNA, we compared the MSNA at any given change in diastolic pressure that occurred at higher lung volumes (the latter 50% of inspiration plus the first 50% of expiration) versus the (remaining) lower lung volumes (Figure 5). The mean level of MSNA at any given change in diastolic pressure was significantly higher at low versus higher lung volumes (P<0.05). This effect of lung volume on MSNA held for all conditions of active and passive ventilation. In addition, the MSNA at any given lung volume showed a significant negative correlation with the change in diastolic pressure during both voluntary hyperventilation and passive PPV. The slope of this relationship was significantly higher (P<0.05) at the lower lung volumes.

    Posthyperventilation Apnea

    The apneas that followed passive PPV showed a dissociation between central respiratory motor output and MSNA (Figure 6). In the 3 instances in which the duration of the apnea was >30 seconds, we observed the following: (a) in the initial seconds of the postventilator apnea (when PETco2 was still low and arterial pressure stable), MSNA burst frequency and amplitude remained unchanged relative to voluntary hyperventilation or the mechanical ventilation period; (b) as the apnea proceeded, and CO2 was rising, MSNA burst frequency and amplitude increased before any appearance of significant respiratory motor output, as noted by the unchanging Pm or Vt; and (c) at apnea termination, MSNA remained high coincident with the appearance of significant respiratory motor output.

    Respiratory Sinus Arrhythmia (RSA): Frequency Domain Analyses

    During spontaneous eupneic breathing, power spectral analyses of the R-R interval data in each subject revealed a small peak in power at the respiratory frequency (see representative subject in Figure 2). The amplitude of the peak in power was markedly increased during voluntary hyperventilation at high Vt. In all subjects, assisted PPV at the same Vt markedly reduced the amplitude of the power at the respiratory frequency compared with voluntary hyperventilation. The amplitude of RSA was further reduced with passive mechanical ventilation.

    RSA: Time Domain Analyses

    All subjects showed a small RSA during spontaneous breathing that was characterized by a decrease in R-R interval throughout inspiration and an increase during expiration (Figure 3A). The RSA was significantly accentuated with increased Vt during voluntary hyperventilation (Figure 3B). When respiratory motor output was eliminated during passive PPV at the same elevated Vt, RSA was eliminated in 4 of the 5 subjects and markedly reduced in the fifth subject (Figure 3C). The RSA was also considerably reduced during assisted mechanical ventilation when respiratory motor output was high, but positive Pm was equal to that during passive mechanical ventilation (Figure 3D). During the resistor trials, when inspiratory effort was greatly increased and Pm markedly decreased, the RSA amplitude was enhanced (Figure 4).


    Summary of Findings

    We examined the breathing-induced variations in MSNA under conditions in which feedback influences, such as within-breath changes in systemic arterial pressure, lung volume, and intrathoracic pressure, either remained constant or were controlled, and the magnitude of respiratory motor output was varied from undetectable to severalfold greater than normal. We found that maximum MSNA always occurred at end expiration and minimum MSNA occurred at end inspiration, and the amplitude of this modulation was not different among widely varying amounts of respiratory motor output. Similarly, variations in intrathoracic pressure, per se, were also without systematic, independent effects on MSNA; this was demonstrated by comparing negative pressure (voluntary) with positive pressure ventilation at comparable levels of respiratory motor output, intravascular pressure, and lung volume. We propose that central respiratory motor output provides a relatively minor, if any, contribution to the respiratory modulation of MSNA in intact humans. Rather, our findings support the postulate that peripheral reflexes elicited by changes in systemic arterial pressure and lung volume are the dominant influences on the within-breath modulation of MSNA.

    Influence of Central Respiratory Motor Output on MSNA

    Our data provide evidence against a significant independent effect of central respiratory motor output on the within-breath modulation of MSNA in the intact human. These findings were consistent across several types of change in respiratory motor output, including (1) voluntary increases in respiratory motor output, (2) hypocapnia-induced elimination of respiratory motor output, and (3) chemoreceptor stimulation during the latter stages of posthyperventilation apnea.

    These findings differ from the strong central respiratory component inferred from studies in anesthetized animals deprived of vagal and baroreceptor feedback.12 Why did our findings not reveal this strong contribution from central respiratory motor output? We hypothesize that feedback mechanisms from baroreceptors and pulmonary stretch receptors are the dominant determinants of the respiratory modulation of MSNA in the intact state. This idea is confirmed in part by studies in intact, anesthetized cats, which showed a pattern of inspiratory inhibition and expiratory activation of MSNA during eupnea that was similar to the pattern observed in intact humans.2 However, when respiratory motor output was increased via CO2 inhalation in this animal model, the MSNA modulation showed inspiratory excitation and expiratory inhibition.2 This pattern implies that the role of central respiratory motor output is important even in the intact cat and remains quite different from that in the human, in whom inspiratory inhibition and expiratory excitation of MSNA occurs even when respiratory motor output is increased via high CO2 or high voluntary drive to breathe (Figure 8, Seals et al5 ).

    Our findings showing a dissociation between respiratory motor output and MSNA during and after hypocapnia-induced apnea imply quite different chemoreceptor thresholds for activation of the respiratory and sympathetic outputs (see Figure 6). This concept is consistent with 2 sets of findings in anesthetized cats. Huang et al14 showed that transient carotid chemoreceptor stimulation (via sodium cyanide) during hypocapnia-induced apnea stimulated cervical sympathetic outflow in the absence of phrenic nerve activity. Trzebski and Kubin15 demonstrated that the Paco2 threshold for sympathetic activation during posthyperventilation apnea was 36 mm Hg, whereas the threshold for resumption of phrenic activity was 44 mm Hg.

    Our conclusions concerning the absence of an effect of respiratory motor output on MSNA are predicated on the assumption that passive PPV eliminated central respiratory motor output. Truly passive mechanical ventilation markedly reduces phasic output from medullary respiratory neurons.1617 However, it may not inhibit all medullary respiratory neuronal activity; in fact, tonic expiratory nerve activity16 and oscillatory glossopharyngeal nerve activity18 have been shown to persist during phrenic apneas secondary to mechanical hyperventilation. The significant prolongation of expiratory time after each passive trial indicated that any residual component of central respiratory drive that remained during the mechanical ventilation was not sufficient to initiate a breath once the ventilator was turned off. So, phasic respiratory motor output, if not completely eliminated, must have been markedly inhibited. Furthermore, despite this greatly reduced central respiratory motor output during passive PPV, the amplitude of the within-breath variation in MSNA was observed to be equal to that shown during active voluntary hyperventilation against a resistive load, when a very large amount of inspiratory effort was required to generate the same Vt.

    Influence of Arterial Pressure Changes

    In agreement with most previous studies in humans45 and intact cats,2 we showed that MSNA changed reciprocally with changes in arterial pressure. This correlation between the within-breath pattern of change in arterial pressure and the respiratory modulation of MSNA was unaffected by PPV in a background of either very high (assisted) or 0 (passive) respiratory motor output, or by dramatically increasing inspiratory effort by voluntary efforts during resistor breathing. Whereas the within-breath changes in diastolic pressure averaged <3 mm Hg under all conditions, small changes in baroreceptor input have been shown to trigger large changes in MSNA both in humans and intact animals.1920 A strong influence of breathing-related arterial baroreceptor activity on sympathetic outflow was also demonstrated in anesthetized cats. Respiratory modulation of MSNA (peroneal nerve) during hyperventilation-induced phrenic silence was unaffected by vagotomy but was eliminated by bilateral carotid occlusion.2

    Our findings suggest that during normal negative-pressure breathing, sympathetic inhibition during inspiration is caused, at least in part, by activation of carotid and aortic baroreceptors resulting from a rise in intravascular pressure and decline in intrathoracic pressure.21 Interestingly, during PPV, the aortic arch baroreceptors were most likely deactivated during inspiration at the same time that the carotid sinus baroreceptors were activated; ie, the 2 sets of receptors provided conflicting information to the central nervous system. In this situation, we also observed sympathetic inhibition during inspiration similar to that observed during normal negative pressure ventilation. Does this mean that carotid sinus baroreceptors predominate over aortic arch baroreceptors in the respiratory modulation of MSNA? Other investigators have advanced the hypothesis that neither set of receptors is predominant under all conditions.22 Instead, when inputs from aortic arch and carotid sinus baroreceptors conflict, the net sympathetic response is determined by the input from the set of receptors that is activated. In the present study, the sympathoinhibition observed during inspiration with positive pressure ventilation was most likely the result of activation of carotid sinus baroreceptors caused by a rise in intravascular pressure, suggesting that the carotid reflex was dominant in that instance.

    Influence of Lung Stretch

    Activation of pulmonary stretch receptors also has a direct inhibitory action on sympathetic activity in anesthetized cats and rats.2324 However, data from patients without intact pulmonary innervation below the carina (as a result of orthotopic heart-lung transplantation) suggest that intact lung inflation reflexes are not obligatory for the respiratory oscillation in MSNA during eupneic breathing, but that vagally mediated lung inflation feedback is the primary mechanism through which the within-breath variation in muscle sympathetic discharge is augmented at high Vt in the human.5 We showed that the level of MSNA, at any given change in diastolic pressure, was higher at low versus higher lung volume phases of the breath cycle, which is suggestive of an independent effect of lung inflation on sympathetic outflow. However, the greatest influence of afferent input from pulmonary stretch receptors may be in modulating sympathetic responsiveness to baroreceptor influences. In fact, our data showed that the slope of the relationship between changes in diastolic pressure and MSNA was greater at lower lung volumes (see Figure 5). These correlative findings are consistent with reports showing that the sensitivity of the sympathetic nervous system to baroreceptor influence fluctuates during breathing, with the greatest responsiveness occurring at low lung volumes when spontaneous MSNA is highest.425

    Influence of Intrathoracic Pressure

    The use of positive intrathoracic pressure to produce passive ventilation introduced additional confounding influences on venous return, cardiac filling, and aortic transmural pressure. However, we did not find any effect of positive pressure, per se, on the respiratory modulation of MSNA as shown by the similarities in the patterns of MSNA modulation observed during voluntary hyperventilation (negative pressure) and assisted mechanical ventilation (positive pressure) at similar high levels of central respiratory motor output and Vt. Macefield and Wallin26 also showed no significant effect of PPV on respiratory modulation of MSNA in humans; however, they also reported that respiration-associated changes in arterial pressure were in the opposite direction during positive versus negative pressure ventilation.26


    Strong evidence in support of a dominant role for central respiratory motor output in causing RSA was provided by the persistence of heart rate modulation, occurring synchronously with phrenic nerve activity, reported in the anesthetized dog during constant flow mechanical ventilation, which eliminated phasic inputs related to respiration.27 Similarly, our findings also point to an effect of central respiratory motor output as indicated by the reduction in RSA amplitude during passive versus active mechanical ventilation (see Figure 3C versus 3D) and also during voluntary increases in Vt at high versus very high levels of respiratory motor output (see Figure 4A versus 4B). However, we also observed that RSA was greatly attenuated during assisted PPV, even when respiratory motor output was very high (see Figure 3B versus 3D). These findings are consistent with the previously reported effects of PPV in greatly attenuating RSA compared with the persistent RSA observed during negative pressure mechanical ventilation.28 Such findings point strongly to atrial stretch and cardiac vagal afferent activity as important modifiers of RSA. Secondly, intact neural feedback from pulmonary stretch receptors also seems to be obligatory to RSA as shown by the absence of neurally mediated RSA in lung transplant patients (with intact hearts)–even in the presence of large superimposed respiration synchronous swings in intrathoracic pressure and also in central respiratory motor output.11 In turn, as with the respiratory modulation of MSNA (see previous section), feedback from lung inflation may be important to RSA because of its ability to limit accessibility of medullary vagal neurons to sensory input from systemic baroreceptors.25 So, to date, two apparently obligatory feedback mechanisms for RSA have been identified in the human, namely those related to intrathoracic pressure and those related to lung inflation. These mechanisms are clearly not redundant, because neither one alone will produce RSA; rather, they may be mutually dependent.

    In summary, it is clear that cardiovascular and pulmonary feedback mechanisms have different relative influences on the within-breath changes in MSNA as compared with the respiratory modulation of heart rate. First, passive PPV abolished RSA but had no effect on the respiratory modulation of MSNA. Second, the within-breath variation in heart rate was critically dependent on input from pulmonary stretch receptors,11 whereas lung denervation did not alter breathing-related oscillations in MSNA during eupneic breathing.5 Finally, our present data using passive versus active mechanical ventilation also showed that central respiratory motor output, per se, had a negligible independent role in the respiratory modulation of MSNA in the intact human but did contribute to the amplitude of RSA.

          Figure 1.

    Figure 1. Raw data from a representative subject during eupneic breathing and under conditions of matched Vt (2 times eupnea), breathing frequency (f), and PETco2, but varying respiratory motor output as evidenced by marked changes in EMGdi, as follows: voluntary hyperventilation, passive positive pressure ventilation, assisted positive pressure ventilation, and added inspiratory resistance (augmented hyperventilation). Note that arterial pressure and MSNA have not yet been corrected for conduction delays.

          Figure 2.

    Figure 2. Power spectral densities for a representative subject under conditions of varying respiratory motor output. Arrows indicate respiratory (Resp) and cardiac (Card) frequencies.

          Figure 3.
          Figure 3.

    Figure 3. Individual and mean (n=5) data for the within-breath changes in MSNA, R-R interval, and diastolic pressure (DP) during spontaneous breathing (A), voluntary hyperventilation (B), passive PPV (C), and assisted PPV (D).

          Figure 4.

    Figure 4. Individual and mean (n=3) data for the within-breath changes in MSNA, R-R interval, and diastolic pressure (DP) in 3 subjects during voluntary hyperventilation (A) and added inspiratory resistance (B).

          Figure 5.

    Figure 5. Effect of changing lung volume on the respiratory modulation of MSNA during spontaneous breathing, voluntary hyperventilation, passive PPV, and assisted PPV. Data are presented as high lung volume points (last 50% of inspiration plus first 50% of expiration) or low lung volume points (last 50% of expiration plus first 50% of inspiration). NS indicates nonsignificant slope (P>0.05).

          Figure 6.

    Figure 6. Raw data from a representative subject during voluntary hyperventilation, passive mechanical ventilation, and the postventilator apneic period. The apneic period shows that in the initial seconds of the apnea when respiratory motor output was absent (a), MSNA remained unchanged relative to voluntary hyperventilation or the preceding period of mechanical ventilation; late in the apneic period (b), MSNA was increased, but respiratory motor output was undetectable; and at apnea termination (c), both MSNA and respiratory motor output were increased. Note that arterial pressure and MSNA have not yet been corrected for conduction delays.

    Table 1. Experimental Protocols Used and the Effects of These Interventions on the Proposed Central and Peripheral Influences on Respiratory Modulation of Muscle Sympathetic Nerve Activity

    ProtocolCentral Inspiratory Motor OutputTidal VolumeIntrathoracic Pressure (Peak Inspiration, cm H2O)PETco2, mm HgSystemic Arterial Pressure (Inspiration)
    Spontaneous eupneaLow0.6 L−237Increasing
    Voluntary hyperventilationHigh2× eupnea−327Increasing
    Passive PPV02× eupnea+2125Increasing
    Assisted PPVHigh2× eupnea+1825Increasing
    Voluntary hyperventilation+inspiratory resistorVery high2× eupnea−1826Increasing

    Table 2. Average (n=5) Per-Minute Measurements of Respiratory and Cardiovascular Variables for Each Experimental Condition

    VariableSpontaneousVoluntary HyperventilationPassive PPVAssisted PPVInspiratory Resistor (n=3)
    Vt, L0.62±0.1111.20±0.111.15 ±0.131.22±0.121.02±0.06
    f, breaths per minute12.8±2.613.3±1.313.3±1.313.1±1.314.8 ±1.04
    PETco2, mm Hg37.1±4.7126.8±3.224.5±2.324.6±2.826.0 ±4.5
    Ti/TTOT0.41 ±0.0520.48±0.020.43±0.0320.45±0.040.50±0.01
    Pm, cm H2O3−1.8±0.5−2.6 ±0.3+20.6±0.9+18.0±3.7−18.2±1.8
    Heart rate, bpm56.0±8.556.8±9.558.3±9.756.9±8.557.7 ±12.2
    MSNA, burst frequency, bursts per minute24.9 ±4.323.1±5.126.3±6.423.2±4.823.1±11.6
    MSNA, total activity424.2±4.422.9±3.329.1 ±6.4122.6±3.222.8±9.2
    Systolic BP, mm Hg114.6±12.6110.4±10.8110.5±10.7107.6 ±11.6108.2±20.3
    Diastolic BP, mm Hg68.2±9.764.1±11.066.1±12.664.3 ±12.259.1±16.3

    BP indicates blood pressure.

    1Significantly different from other conditions.

    2Significantly different from voluntary hyperventilation.

    3Peak inspiratory mask pressure. During assisted PPV, subjects generated −3.5±1.0 cm H2O to trigger each positive pressure, ventilator breath.

    4Total activity refers to the product of burst frequency and peak amplitude and is expressed in arbitrary units per minute.

    This work was supported by the National Heart, Lung, and Blood Institute (Grants R01-15469 and R29-57401), the Veterans Affairs Medical Research Service, and (in part) by a research fellowship from the American Heart Association of Wisconsin and the Hazel Mae Mayer Foundation (to C.M.S.). We thank D. Puleo, D.F. Pegelow, and A. Jacques for technical support, and Dr B. Goodman for his assistance with the statistical analyses.


    Correspondence to Jerome A. Dempsey, PhD, Department of Preventive Medicine, University of Wisconsin-Madison, 504 N Walnut St, Madison, WI 53705. E-mail


    • 1 Bachoo M, Polosa C. Properties of the inspiration-related activity of sympathetic preganglionic neurones of the cervical trunk in the cat. J Physiol (Lond).1987; 385:545–564.CrossrefGoogle Scholar
    • 2 Boczek-Funcke A, Habler H-J, Janig W, Michealis M. Respiratory modulation of the activity in sympathetic neurones supplying muscle, skin and pelvic organs in the cat. J Physiol (Lond).1992; 449:333–361.CrossrefGoogle Scholar
    • 3 Richter DW, Spyer KM. Cardiorespiratory control. In: Loewy AD, Spyer KM, eds. Central Regulation of Autonomic Functions. New York, NY: Oxford University Press; 1990:189–207.Google Scholar
    • 4 Eckberg DL, Nerhed C, Wallin BG. Respiratory modulation of muscle sympathetic and vagal cardiac outflow in man. J Physiol (Lond).1985; 365:181–196.CrossrefGoogle Scholar
    • 5 Seals DR, Suwarno NO, Joyner MJ, Iber C, Copeland JG, Dempsey JA. Respiratory modulation of muscle sympathetic nerve activity in intact and lung denervated humans. Circ Res.1993; 72:440–454.CrossrefMedlineGoogle Scholar
    • 6 Wilson CR, Manchanda S, Crabtree D, Skatrud JB, Dempsey JA. An induced blood pressure rise does not alter upper airway resistance in sleeping humans. J Appl Physiol.1998; 84:269–276.CrossrefMedlineGoogle Scholar
    • 7 Vallbo A, Hagbarth K-E, Torebjork HE, Wallin BG. Somatosensory, proprioceptive, and sympathetic activity in human peripheral nerves. Physiol Rev.1979; 59:919–957.CrossrefMedlineGoogle Scholar
    • 8 Morgan BJ, Denahan T, Ebert TJ. Neurocirculatory consequences of negative intrathoracic pressure versus asphyxia during voluntary apnea. J Appl Physiol.1993; 74:2969–2975.CrossrefMedlineGoogle Scholar
    • 9 Victor RG, Leimbach WN. Effects of lower body negative pressure on sympathetic discharge to leg muscles in humans. J Appl Physiol.1987; 63:2558–2562.CrossrefMedlineGoogle Scholar
    • 10 Leevers AM, Simon PM, Dempsey JA. Apnea after normocapnic mechanical ventilation during NREM sleep. J Appl Physiol.1994; 77:2079–2085.CrossrefMedlineGoogle Scholar
    • 11 Taha BH, Simon PM, Dempsey JA, Skatrud JB, Iber C. Respiratory sinus arrhythmia in humans: an obligatory role for vagal feedback from the lungs. J Appl Physiol.1995; 78:638–645.CrossrefMedlineGoogle Scholar
    • 12 Fagius J, Wallin BG. Sympathetic reflex latencies and conduction velocities in normal man. J Neurol Sci.1980; 47:433–448.CrossrefMedlineGoogle Scholar
    • 13 Johnson R, Wichern D. Applied Multivariate Statistical Methods. 2nd ed. Englewood Cliffs, NJ: Prentice-Hall; 1988.Google Scholar
    • 14 Huang W, Lahiri S, Mokashi A, Sherpa AK. Relationship between sympathetic and phrenic nerve responses to peripheral chemoreflex in the cat. J Auton Nerv Syst.1988; 25:95–105.CrossrefMedlineGoogle Scholar
    • 15 Trzebski A, Kubin L. Is the central inspiratory activity responsible for Pco2-dependent drive of the sympathetic discharge? J Auton Nerv Syst.1981; 3:401–420.CrossrefMedlineGoogle Scholar
    • 16 Batsel HL. Activity of bulbar respiratory neurons during passive hyperventilation. Exp Neurol.1967; 19:357–374.CrossrefMedlineGoogle Scholar
    • 17 Orem J, Vidruk EH. Activity of medullary respiratory neurons during ventilator-induced apnea in sleep and wakefulness. J Appl Physiol.1998; 84:922–932.CrossrefMedlineGoogle Scholar
    • 18 Fukuda Y. Difference in glossopharyngeal and phrenic inspiratory activities of rats during hypocapnia and hypoxia. Neurosci Lett.1992; 137:261–264.CrossrefMedlineGoogle Scholar
    • 19 Rea RF, Eckberg DL. Carotid baroreceptor-muscle sympathetic relation in humans. Am J Physiol.1987; 253:R929–R934.MedlineGoogle Scholar
    • 20 Tang PC, Maire FW, Amassian VE. Respiratory influence on the vasomotor center. Am J Physiol.1957; 191:218–224.CrossrefMedlineGoogle Scholar
    • 21 Angell-James JE. The effects of changes in extramural, “intra-thoracic,” pressure on aortic arch baroreceptors. J Physiol (Lond).1971; 214:89–103.CrossrefGoogle Scholar
    • 22 Smith ML, Beightol LA, Fritsch-Yelle JM, Ellenbogen KA, Porter TR, Eckberg DL. Valsalva’s maneuver revisited: a quantitative method yielding insights into human autonomic control. Am J Physiol.1996; 27:H1240–H1249.Google Scholar
    • 23 Gerber U, Polosa C. Effects of pulmonary stretch receptors afferent stimulation on sympathetic preganglionic neuron firing. Can J Physiol Pharmacol.1978; 56:191–198.CrossrefMedlineGoogle Scholar
    • 24 Yu J, Roberts AM, Joshua IG. Lung inflation evokes reflex dilation of microvessels in rat skeletal muscle. Am J Physiol.1990; 258:H939–H945.MedlineGoogle Scholar
    • 25 Eckberg DL, Kifle YT, Roberts VL. Phase relationship between normal human respiration and baroreflex responsiveness. J Physiol.1980; 304:489–502.CrossrefMedlineGoogle Scholar
    • 26 Macefield VG, Wallin BG. Modulation of muscle sympathetic activity during spontaneous and artificial ventilation and apnoea in humans. J Auton Nerv Syst.1995; 53:137–147.CrossrefMedlineGoogle Scholar
    • 27 Shykoff BE, Naqvi SJ, Menon AS, Slutsky AS. Respiratory sinus arrhythmia in dogs. J Clin Invest.1991; 87:1621–1627.CrossrefMedlineGoogle Scholar
    • 28 Freychuss U, Melcher A. Sinus arrhythmia in man: influence of tidal volume and oesophageal pressure. Scand J Clin Lab Invest.1975; 35:487–496.CrossrefMedlineGoogle Scholar


    eLetters should relate to an article recently published in the journal and are not a forum for providing unpublished data. Comments are reviewed for appropriate use of tone and language. Comments are not peer-reviewed. Acceptable comments are posted to the journal website only. Comments are not published in an issue and are not indexed in PubMed. Comments should be no longer than 500 words and will only be posted online. References are limited to 10. Authors of the article cited in the comment will be invited to reply, as appropriate.

    Comments and feedback on AHA/ASA Scientific Statements and Guidelines should be directed to the AHA/ASA Manuscript Oversight Committee via its Correspondence page.