- Email: [email protected]
Dynamic changes in baroreceptor-sympathetic coupling during the respiratory cycle
Brain Research 1046 (2005) 216 – 223 www.elsevier.com/locate/brainres
Research report
Dynamic changes in baroreceptor-sympathetic coupling during the respiratory cycle Gerard L. Gebber*, Mahasweta Das, Susan M. Barman Department of Pharmacology and Toxicology, Michigan State University, East Lansing, MI 48824-1317, USA Accepted 1 April 2005 Available online 24 May 2005
Abstract In urethane-anesthetized, paralyzed, and artificially ventilated cats, we observed an unusual form of ‘‘phase walk’’ of the cardiac-related burst of inferior cardiac postganglionic sympathetic nerve discharge (SND) relative to the systolic phase of the arterial pulse (AP) and thus pulse-synchronous baroreceptor nerve activity. Unlike classic phase walk ascribable to weakened coupling (desynchronization) of two oscillators, AP-SND phase walk was characterized by epochs of progressive, bidirectional changes in the angle of strong coupling (AP-SND coherence values, >0.9) of these signals that recurred on the time scale of the respiratory cycle and whose range was approximately one third of the period of the heart beat. AP-SND phase walk was linked to two respiratory variables (central respiratory drive and vagal lung inflation afferent activity) as demonstrated by the following observations. First, in five normocapnic cats (end-tidal CO2, 4.3 T 0.2%) with intact vagus nerves and three vagotomized cats, AP-SND phase walk was characterized by a progressive heart-beat-to-heart-beat decrease in the lag of SND relative to the AP during the inspiratory phase of phrenic nerve activity and an increase in the lag during the expiratory phase. Second, in three cats with intact vagus nerves that were hyperventilated (end tidal CO2, 1.6 T 0.4%) to phrenic nerve quiescence, the lag of the cardiac-related burst of SND relative to the AP increased during lung inflation and decreased during lung deflation. Additional experimentation revealed that AP-SND phase walk is attributable to respiratory-induced changes in the frequency of the centrally generated sympathetic nerve rhythm rather than heart rate. Moreover, the data demonstrate that the frequency and amplitude of the sympathetic oscillation are independently controlled by the above mentioned respiratory parameters. D 2005 Elsevier B.V. All rights reserved. Theme: Endocrine and autonomic regulation Topic: Cardiovascular regulation Keywords: Cardiorespiratory coordination; Frequency and amplitude modulation; Phase walk; Resetting of a rhythm
1. Introduction In earlier reports from our laboratory [14,15], two modes of coordination of cardiac-related sympathetic nerve discharge (SND) to pulse-synchronous baroreceptor nerve activity were described in urethane-anesthetized cats. In these experiments, the peak of the systolic phase of the arterial pulse (AP) was used to reflect the timing of pulsesynchronous baroreceptor nerve activity. One mode of
* Corresponding author. Fax: +1 517 353 8915. E-mail address: [email protected] (G.L. Gebber). 0006-8993/$ - see front matter D 2005 Elsevier B.V. All rights reserved. doi:10.1016/j.brainres.2005.04.009
baroreceptor-sympathetic coordination was characterized by phase-locking (i.e., nearly constant delay of SND relative to peak systole). The second mode was characterized by respiratory-related epochs of progressive and bidirectional heart-beat-to-heart-beat changes in the delay of SND relative to peak systole (expressed as a phase angle). We refer to this mode as an unusual type of ‘‘phase walk,’’ a similar form of which has been observed between the 10-Hz rhythmic discharges of sympathetic nerve pairs recorded in baroreceptor-denervated and vagotomized cats [9]. The current study was initiated with three purposes in mind. Our first objective was to define the respiratory variables that account for AP-SND phase walk which we
G.L. Gebber et al. / Brain Research 1046 (2005) 216 – 223
consider to reflect dynamic changes in the coupling of a central sympathetic generator to the cardiac cycle by pulsesynchronous baroreceptor discharge [14,15]. Specifically, we investigated the roles played by central respiratory drive and phasic vagal lung inflation afferent activity in producing AP-SND phase walk. The second objective was to define the mechanism of AP-SND phase walk. In particular, we investigated whether changes in AP-SND phase angle were more closely linked to respiratory-related changes in heart rate, and thus the timing of pulse-synchronous baroreceptor nerve activity or the frequency at which central circuits generate rhythmic bursts of SND. The third objective was to determine whether the respiratory-related changes in the AP-SND phase angle are linked to those of the amplitude of cardiac-related bursts of SND. Regarding this issue, respiratory modulation of sympathetic nerve burst amplitude is commonly viewed as an electrophysiological correlate of cardiorespiratory coordination [1,4,10].
217
cardiac nerve recordings were made with the Grass Model P511 preamplifier bandpass set at 1– 1000 Hz so that bursts of multiunit spikes appeared as slow waves [4]. Phrenic nerve activity (PNA) was bandpass filtered (1 –1000 Hz) after which the signal was passed through a moving averager (CWE, Inc., Model RA-821 RSP) with a time constant of 100 ms. PNA was used as an index of the central respiratory cycle. 2.3. Frequency analysis Fast Fourier transform was used to construct autospectra of SND and the AP and coherence functions (normalized cross-spectra) relating these signals [9,14]. The sampling rate of 200 Hz yielded a bin-resolution of 0.2 Hz. The spectra were averages of 32 5-s data windows with 50% overlap. Coherence values are presented on a scale of 0 – 1.0 with a value of 1.0 reflecting an ideal relationship and values 0.1, reflecting a significant relationship between the AP and SND at the frequency of the heart beat [2].
2. Methods 2.4. Time series 2.1. Anesthesia and general procedures The experimental protocols were approved by the AllUniversity Committee on Animal Use and Care of Michigan State University. The experiments were performed on five baroreceptor-intact cats anesthetized with an intravenous injection of urethane (1.4 –1.8 g/kg). With this dose, the cat is maintained in a state of surgical anesthesia for a period (8– 10 h [7]) that exceeded the duration of our experiments, as indicated by the failure of noxious stimuli (pinch, cauterization of skin and muscle) to desynchronize spindles and delta-slow wave activity in the frontal – parietal electroencephalogram [14,15,17]. The animals were artificially ventilated and paralyzed with gallamine triethiodide (4 mg/kg iv, initial dose); bilateral pneumothoracotomy was performed to minimize ventilator pump-related changes in blood pressure. Blood pressure was measured from a catheter inserted into a femoral artery, and drugs were administered into a femoral vein. End-tidal CO2 was measured by using a Traverse Medical Monitors Capnometer (model 2200). Body temperature was kept near 37 -C with a heat lamp. Intratracheal pressure (ITP) was monitored by using a Grass Instruments volumetric pressure transducer (Model PT5). Upward deflections of ITP denote inflation of the lungs by the respirator pump (Harvard, Inspira ASV). In three experiments, both vagus nerves were sectioned in the neck. 2.2. Nerve recordings As described in an earlier report [9], potentials were recorded monophasically with bipolar platinum electrodes from the central end of the left inferior cardiac postganglionic sympathetic nerve and right phrenic nerve. Inferior
As a prelude to time series analysis, the cardiac-related component in the original wide bandpass recording of SND was extracted by digital filtering. The software for the digital filter (symmetric, nonrecursive type with a Lanczos smoothing function) was obtained from RC Electronics (Santa Barbara, CA). The width of the bandpass digital filter was set at 4 Hz with the center frequency matched to that of the sharp peak at the cardiac frequency in the autospectrum of SND. The digital filter had a roll-off slope of 39%/Hz outside of the bandpass and produced minimal phase and amplitude distortion as ascertained by comparing the filtered records with the originals [9,15]. The filtered records were smoother than the originals, thereby improving the accuracy of detection of the peaks and troughs of the cardiac-related sympathetic nerve slow waves. Software written in our laboratory [14,15] was used to construct time series showing heart-beat-to-heart-beat measurements of (1) systolic blood pressure, (2) intervals between heart beats (heart period, ms), (3) peak-to-trough sympathetic nerve slow wave amplitude normalized on a scale of 0 –1.0, (4) intervals (ms) between the peaks of slow waves, and (5) phase angle between the peak of the AP and the peak of the corresponding slow wave in SND, expressed in degrees. AP-SND phase angle was measured with a resolution of 5.4-/bin (sampling period was 5 ms) when the heart period was 333 ms (heart rate, 3 Hz). Positive and negative values of phase angle refer to instances when the peak of the cardiac-related sympathetic nerve slow wave lagged or led peak systole, respectively. Whereas, for example, a lag of 240- is mathematically equivalent to a lead of 120-, values of AP-SND phase angle were plotted so that epochs of respiratory-related phase walk appear as a continuous rather than as a discontinuous curve.
218
G.L. Gebber et al. / Brain Research 1046 (2005) 216 – 223
2.5. Statistical analysis Values in the text are means T SEM. The Pearson product-moment correlation coefficient (r) was used to determine whether changes in sympathetic nerve slow wave amplitude were correlated to changes in AP-SND phase angle. Colton’s [5] rule of thumb was used to interpret the strength of correlation of these two parameters. An r value from 0 to 0.25 (or 0.25) indicates little or no relationship; one from 0.25 to 0.50 (or 0.25 to 0.50) a weak relationship; one from 0.50 to 0.75 (or 0.50 to 0.75) a moderate to good relationship; and one >0.75 (or < 0.75) indicates a very good to excellent relationship.
3. Results 3.1. Normocapnic cats with intact vagus nerves Fig. 1 shows the changes in AP-SND phase angle during one respiratory cycle (as reflected by PNA; top trace) in a normocapnic cat with intact vagus nerves. Using the vertical lines as guides, one can note that the peak of the cardiacrelated slow wave in SND (bottom trace) lagged peak systole (middle trace) by 94 ms (approximately one-third of the heart period) during expiration. The lag was progressively shortened so that peak systole and the peak of the sympathetic nerve slow wave occurred almost simultane-
Fig. 2. SND is strongly correlated to AP during respiratory-related phase walk (same experiment as in Fig. 1). Traces are (top to bottom) autospectrum (AS) of AP, AS of SND, and coherence function relating AP and SND. The spectra are averages of thirty-two 5-s windows with 50% overlap and have a resolution of 0.2 Hz/bin.
Fig. 1. Respiratory-related phase walk of cardiac-related slow wave of inferior cardiac postganglionic sympathetic nerve discharge (SND) relative to arterial pulse (AP). Oscilloscopic traces are (top to bottom) phrenic nerve activity (PNA; inspiration is upward), AP, and SND. Time scale is 0.5 s/division. Vertical lines drawn through peak systolic phase of AP illustrate the bidirectional AP-SND phase walk during one respiratory cycle.
ously (in-phase) in mid-inspiration. The lag of the slow wave relative to peak systole then progressively returned to its original value during late inspiration and the next expiratory period. In this example, changes in the amplitude of the sympathetic nerve slow wave were inversely related to the changes in AP-SND phase angle, and blood pressure did not change during the respiratory cycle. Autospectra of the AP and SND and coherence function relating the two signals during an 80-s epoch of recurring respiratory-related phase walk are shown in Fig. 2. The traces in Fig. 1 formed part of this data block. Both autospectra and the coherence function contain a sharp peak at the frequency of the heart beat. The coherence value relating SND to the AP was 0.85. Thus, despite phase walk over a considerable range (see Fig. 1), AP and SND were strongly coupled. Time series constructed from a 15-s data block collected later in the same experiment are shown in Fig. 3. At this time, respiratory-related oscillations in systolic blood pressure (SBP, top trace) had developed. As is usually the case in
G.L. Gebber et al. / Brain Research 1046 (2005) 216 – 223
Fig. 3. Respiratory-related changes in AP-SND phase angle and sympathetic nerve slow wave amplitude in a normocapnic cat with intact vagus nerves. Traces are (top to bottom) heart-beat-to-heart-beat readings of peak systolic blood pressure (SBP; mm Hg), and heart period (HP; ms), oscilloscopic traces of PNA and intratracheal pressure (ITP; inflation is upwards), and heart-beat-to-heart-beat readings of AP-SND phase angle (in degrees with SND lagging AP), and sympathetic nerve slow wave amplitude (normalized on a scale of 0 to1.0). The r value relating AP-SND phase angle to sympathetic slow wave amplitude is 0.70. Time base is 1 s/division.
artificially ventilated cats [3,20], the inspiratory phase of the central respiratory cycle (PNA, 3rd trace from top) was out of phase with lung inflation (ITP, 4th trace). The respiratoryrelated epochs of AP-SND phase walk (5th trace) were characterized by a progressive heart-beat-to-heart-beat decrease in the lag of SND relative to the AP during inspiration (which coincided with deflation of the lungs) and an increase in the lag during expiration (coinciding with lung inflation). Note that the range of the bidirectional phase walk was ¨100- (about 28% of the heart period). Moreover, AP-SND phase walk occurred in the absence of respiratoryrelated changes in heart period (2nd trace). The bottom trace in Fig. 3 shows that sympathetic nerve slow wave amplitude
219
increased during the inspiratory phase of PNA and decreased during expiration. As such, the changes in slow wave amplitude were out of phase with those in AP-SND phase angle. In this case, the r value relating slow wave amplitude to the phase angle was 0.70 ( P < 0.0001; a good relationship according to Colton’s rule of thumb [5]). We analyzed two or three 15-s episodes of recurring respiratory-related AP-SND phase walk in each of five normocapnic cats with intact vagus nerves. The total number of episodes analyzed was 14. Mean central respiratory rate was 20.2 T 1.6/min, and the central respiratory cycle was locked 1:1 to the artificial ventilation cycle. Mean blood pressure was 142 T 6 mm Hg, mean heart rate was 3.4 T 0.1 Hz, and end-tidal CO2 averaged 4.3 T 0.2%. The peak coherence value relating SND to the AP at the frequency of the heart beat averaged 0.92 T 0.01, and the range of the phase walk was 126 T 10- (approximately onethird of the heart period). In all cases, the AP-SND phase angle decreased during the inspiratory phase of PNA and increased during expiration. According to Colton [5], the strength of the inverse relationship between AP-SND phase angle and sympathetic nerve slow wave amplitude was excellent in one case (r = 0.76), good in six cases (r value range: 0.58 to 0.70), and weak in another six cases (r value range: 0.26 to 0.47). There was no correlation (r = 0.15) in the remaining case. The case in which no correlation existed is shown in Fig. 4. Note the prominent AP-SND phase walk with the peak of the sympathetic nerve slow wave lagging peak systole by as much as 40- in expiration and leading by as much as 85- in inspiration. In contrast to the case in Fig. 3, however, there were little or no respiratory-related changes in sympathetic nerve slow wave amplitude. In this experiment, respiratoryrelated changes in SBP were of small amplitude (¨4 mm Hg) and there were no respiratory-related changes in heart rate. 3.2. Normocapnic, vagotomized cats Fig. 5 shows data from a vagotomized cat. As expected in vagotomized cats, the central respiratory cycle (as reflected by PNA) was not coupled to the artificial ventilation cycle (ITP). Under these conditions, AP-SND phase walk was coupled to the central respiratory cycle, but not ITP. Note that the cardiac-related sympathetic nerve slow waves lagged peak systole by ¨70- during early expiration and led by as much as 100- during the inspiratory phase of PNA. In this case, AP-SND phase walk occurred in the absence of central respiratory-related changes in SBP and heart period, and respiratory modulation of sympathetic nerve slow wave amplitude was weak. The r value relating AP-SND phase angle and slow wave amplitude was 0.32. We analyzed two or three 15-s episodes (n = 7) of recurring AP-SND phase walk coupled to the central respiratory cycle in each of three vagotomized cats. Mean central respiratory rate was 15.3 T 2.0/min. Mean blood pressure was 138 T 3 mm Hg, and mean heart rate was 3.3 T
220
G.L. Gebber et al. / Brain Research 1046 (2005) 216 – 223
PNA although one ‘‘gasp-like’’ episode appears near the middle of the 10-s data block. The low CO2 level was induced by raising the volume of inflation from 35 to 55 ml and increasing the rate of artificial ventilation from 20 to 30/ min. Under these conditions, the cardiac-related slow wave in SND lagged peak systole by ¨270- during inflation of the lungs. The lag was decreased to a value close to 90during lung deflation. Note also that sympathetic nerve slow wave amplitude decreased during inflation and increased during deflation. In this case, the r value relating AP-SND phase angle to sympathetic nerve slow wave amplitude was 0.72, reflecting a good inverse relationship. Although ventilator pump-related changes in SBP (¨4 mm Hg) occurred in this experiment, there were essentially no respiratory-related changes in heart period. We analyzed one 10- or 15-s episode of AP-SND phase walk coupled to the artificial ventilation cycle in each of three hyperventilated cats with intact vagus nerves. Mean
Fig. 4. Respiratory-related AP-SND phase walk unrelated to changes in sympathetic nerve slow wave amplitude. The r value relating AP-SND phase angle to sympathetic slow wave amplitude is 0.15 in this normocapnic cat with intact vagus nerves. Positive and negative phase angles reflect lags and leads, respectively, of the peak of the sympathetic nerve slow wave relative to peak systole. Sequence of traces, abbreviations, and time base are same as in Fig. 3.
0.3 Hz. End-tidal CO2 averaged 4.3 T 0.2%. The peak coherence value relating SND to the AP at the frequency of the heart beat averaged 0.94 T 0.02, and the range of the respiratory-related AP-SND phase walk was 143 T 19-. In all cases, AP-SND phase angle decreased during the inspiratory phase of PNA and increased during the expiratory phase. The strength of the inverse relationship between sympathetic nerve slow wave amplitude and AP-SND phase angle was good in one case (r = 0.55) and weak in four cases (r value range: 0.25 to 0.35). These parameters were uncorrelated in the remaining two cases (r = 0.15 and 0.20). 3.3. Hyperventilated cats with intact vagus nerves Fig. 6 shows a case of AP-SND phase walk that recurred on the time scale of artificial ventilation in a hyperventilated cat (end-tidal CO2 = 1.8%). Note the absence of rhythmic
Fig. 5. Respiratory-related AP-SND phase walk in a normocapnic, vagotomized cat. The r value relating AP-SND phase angle to sympathetic slow wave amplitude is 0.32. Sequence of traces, abbreviations, and time base are same as in Fig. 3.
G.L. Gebber et al. / Brain Research 1046 (2005) 216 – 223
221
SND phase walk. This was the case independent of whether the vagus nerves were intact or not. A typical example in a cat with intact vagus nerves is shown in Fig. 7. Three respiratory-related epochs of AP-SND phase walk appear in the 3rd panel from the top. The peak of the sympathetic nerve slow wave lagged the AP by nearly 50- during the expiratory phase of PNA and led by nearly 55- during inspiration. The cycle-by-cycle changes in the interval between the peaks of adjacent sympathetic slow waves (4th panel) occurred in parallel with those in AP-SND phase angle. The range (319 –420 ms) of inter-sympathetic slow wave intervals during the respiratory cycle was equivalent to a change in the frequency of sympathetic slow wave generation from 3.1 to 2.4 Hz. In contrast to the intersympathetic slow wave interval, heart period remained essentially constant throughout the respiratory cycle (see bottom panel in Fig. 7 and additional cases in Figs. 3– 5). The absence or minimal change in heart period likely is
Fig. 6. AP-SND phase walk and changes in sympathetic nerve slow wave amplitude related to artificial ventilation cycle in a hyperventilated cat with intact vagus nerves. Rhythmic PNA was absent when end-tidal CO2 was lowered to 1.8%. The r value relating AP-SND phase angle to sympathetic slow wave amplitude is 0.72. Sequence of traces, abbreviations, and time base are same as in Fig. 3.
ventilatory rate was 33 T 2 breaths/min. Mean blood pressure was 126 T 4 mm Hg, and mean heart rate was 3.0 T 0.1 Hz. End-tidal CO2 averaged 1.6 T 0.4%. The peak coherence value relating SND to the AP at the frequency of the heart beat averaged 0.91 T 0.01, and the range of the APSND phase walk was 127 T 33-. In all cases, the lag of the cardiac-related sympathetic nerve wave increased during inflation and decreased during deflation. The r values relating sympathetic nerve slow wave amplitude to AP-SND phase angle were 0.72, 0.35, and 0.09. 3.4. Changes in heart period and inter-sympathetic nerve slow wave interval during AP-SND phase walk Respiratory-related changes in the interval between the peaks of successive cardiac-related sympathetic nerve slow waves invariably exceeded those in heart period during AP-
Fig. 7. Changes in inter-sympathetic nerve slow wave interval and heart period during respiratory-related AP-SND phase walk in a normocapnic cat with intact vagus nerves. Traces are (top to bottom) oscilloscopic traces of PNA and ITP, heart-beat-to-heart-beat readings of AP-SND phase angle (in degrees), interval (in ms) between peaks of sympathetic nerve slow waves, and heart period (in ms). Time base is 1 s/division. Abbreviations are same as in Fig. 3.
222
G.L. Gebber et al. / Brain Research 1046 (2005) 216 – 223
attributable to the vagolytic action of the neuromuscular blocking agent, gallamine.
4. Discussion Traditionally, phase walk has been attributed to weakened coupling of two oscillators or the loss of entrainment of an oscillator to a periodic forcing input [6,11 – 13,18]. Such is thought to occur when the intrinsic frequencies of the oscillators become so disparate that the coupling is no longer able to enforce synchrony. As a result, the phase angle between the oscillators cycles repeatedly in one direction through a full 360-. The phase walk of cardiacrelated SND relative to the AP described in the current study did not fit this pattern. First, the coherence value relating SND to the AP approached 1.0 during phase walk. Thus, SND was strongly coupled to pulse-synchronous baroreceptor nerve activity despite the progressive heart-beat-toheart-beat changes in the phase angle between these signals. We attribute the high coherence to the similarity of recurring epochs of the respiratory-related AP-SND phase walk. Second, respiratory-related phase walk was bidirectional rather than unidirectional. Third, rather than progressing through a full 360- (heart period), the range of the phase walk was restricted to approximately one-third of the period of the heart beat. These characteristics argue against the possibility that AP-SND phase walk arose from weakened coupling of cardiac-related SND to pulse-synchronous baroreceptor nerve activity. Rather, we propose that APSND phase walk reflected heart-beat-to-heart-beat resetting of the angle of strong coupling of a centrally generated sympathetic rhythm to forcing pulse-synchronous baroreceptor input. Because the phase walk was respiratory related, we attribute the progressive resetting of the angle of coupling to the continuously changing level of respiratory inputs to circuits responsible for baroreceptor-sympathetic coordination. We have identified two respiratory variables (central respiratory drive and vagal lung inflation afferent activity) involved in AP-SND phase walk. AP-SND phase walk was linked to each of these factors in the absence of the other. First, phase walk recurred on the time scale of the central respiratory cycle in vagotomized cats with the AP-SND phase angle decreasing during the inspiratory phase of PNA and increasing during expiration. Second, phase walk recurred on the time scale of the artificial ventilation cycle in cats with intact vagus nerves when hyperventilation eliminated rhythmic PNA. Under these conditions, the APSND phase angle increased during lung inflation and decreased during deflation. There were clear-cut cases (see Figs. 1 and 5) when AP-SND phase walk occurred in the absence of respiratory-related changes in blood pressure. Thus, AP-SND phase walk could occur independently of cyclic changes in the strength of pulse-synchronous baroreceptor nerve activity linked to respiratory-related
oscillations in blood pressure. Nonetheless, it is possible that changes in baroreceptor nerve activity linked to respiratory-related blood pressure oscillations (see Figs. 3, 4, and 6) contributed to AP-SND phase walk in many cases. Time series analysis revealed progressive shortening of the inter-sympathetic slow wave interval during the inspiratory phase of PNA and progressive lengthening during expiration (see Fig. 7). Thus, the frequency of the sympathetic nerve rhythm increased during inspiration and decreased during expiration. In most cases, such changes occurred in the absence of respiratory-related fluctuations in heart period. These data suggest that the mechanism of APSND phase walk involved respiratory-related changes in the excitability of the central sympathetic rhythm generator. Increased excitability during the inspiratory phase of the central respiratory cycle would shorten the inter-sympathetic nerve burst interval, thereby leading to a progressive heartbeat-to-heart-beat advancement of the timing of the peak of cardiac-related sympathetic activity relative to pulse-synchronous baroreceptor nerve activity. Under the conditions of a constant heart period, AP-SND phase angle would decrease progressively. This follows from the work of Wever [19] on the entrainment of circadian rhythms to the light-dark cycle. Wever [19] demonstrated that the higher the intrinsic frequency (shorter period) of the central rhythm, the more it leads (or the less it lags) an entraining light pulse. Changes in phase angle occur in the opposite direction (increased lag or decreased lead) when the intrinsic frequency of the central rhythm is lowered (increased period). Thus, as might be expected, we observed a progressive increase in AP-SND phase angle when the inter-sympathetic nerve slow wave interval increased from heart-beat-to-heart-beat during the expiratory phase of the central respiratory cycle. Vagal lung inflation afferent activity appeared to exert an effect on central sympathetic excitability opposite to that of central inspiratory drive. During inflation in normocapnic cats (Figs. 3 and 4), AP-SND phase angle increased progressively. AP-SND phase angle also increased during inflation in hyperventilated cats when rhythmic PNA was lost (Fig. 6). The increase in phase angle is presumed to reflect decreased excitability of the sympathetic generator leading to an increase in the period of the rhythm (i.e., lowered frequency) and thus lengthening of the AP-SND phase angle. It is unlikely that AP-SND phase angle served as an important determinant of sympathetic nerve slow wave amplitude under the conditions of our experiments. Rather, during phase walk, it appears that these parameters were independently controlled, perhaps by the actions of different groups of respiratory neurons mediated at different levels (oscillator versus follower circuits) of the networks controlling SND. Regarding this point, the inverse relationship between AP-SND phase angle and sympathetic nerve slow wave amplitude was weak (r value between 0.25 and 0.50) or absent (r value between 0 and 0.25) in the
G.L. Gebber et al. / Brain Research 1046 (2005) 216 – 223
majority of the cases. examined in cats with intact or sectioned vagus nerves. In some cases, prominent respiratory-related AP-SND phase walk was observed with little or no sign of amplitude modulation on the same time scale (Fig. 5). In view of the proposed mechanism for the respiratory-related changes in AP-SND phase angle (see above), it follows that, at least to some extent, the frequency of sympathetic nerve bursts can be regulated independently of burst amplitude. Quite obviously, changes in either of these parameters will influence end organ function. Additional examples of differential or independent control of sympathetic nerve burst frequency and amplitude have been cited by Malpas [16]. Whereas the physiological significance of heart-beat-toheart-beat changes in the angle of strong coupling of SND to pulse-synchronous baroreceptor activity during the respiratory cycle remains obscure, the phenomenon of bidirectional phase walk supports the view that the cardiac-related rhythm in SND reflects the entrainment of a central oscillator by an extrinsic, periodic forcing input rather than the periodic inhibition of randomly generated central activity [8,14,15]. If the cardiac-related rhythm in SND was the simple consequence of the waxing and waning of constant latency central inhibition locked to pulse-synchronous baroreceptor activity, little or no change in the AP-SND phase angle would be expected during the respiratory cycle. Regarding this point, Lewis et al. [15] have demonstrated that the extent of AP-SND phase walk, which was equivalent to between one-third and one-half of the heart period, is too large to be explained by the range of axonal conduction velocities of bulbospinal and preganglionic neurons that provide driving input to the inferior cardiac postganglionic nerve. Perhaps the most convincing evidence for the entrainment of a central rhythm generator by pulsesynchronous baroreceptor activity is provided by epochs of recurring respiratory-related phase walk in which the lag of the peak of the sympathetic nerve slow wave relative to peak systole was reversed to a lead (see Figs. 4, 5, and 7). In these cases, the period of reduced SND (start of falling phase of slow wave) could actually precede the onset of systole. This occurred at points in the respiratory cycle when the peak of the slow wave led peak systole by approximately one-fourth of the heart period. During these heart beats, the period of reduced SND began before pulsesynchronous baroreceptor nerve activity was initiated. In summary, the current study revealed new information on the neural mechanisms responsible for cardiorespiratory coordination. First, we demonstrated that the dynamic changes in baroreceptor-sympathetic coupling (AP-SND phase walk) occurring during the respiratory cycle are attributable to progressive changes in the excitability of a central sympathetic rhythm generator rather than the timing of pulse-synchronous baroreceptor nerve activity. Specifically, we propose that AP-SND phase walk reflected heartbeat-to-heart-beat resetting of the angle of strong coupling of a centrally generated sympathetic rhythm to forcing
223
pulse-synchronous baroreceptor input by continuously changing levels of central respiratory drive and vagal lung inflation afferent activity. Second, we demonstrated independent control of the frequency and amplitude cardiacrelated sympathetic nerve bursts by each of the two above mentioned respiratory variables.
Acknowledgment This study was supported by National Heart, Lung, and Blood Institute Grant HL-33266.
References [1] E.D. Adrian, D.W. Bronk, G. Phillips, Discharges in mammalian sympathetic nerves, J. Physiol. (Lond.) 74 (1932) 115 – 133. [2] V.A. Benignus, Correction to ‘‘Estimation of the coherence spectrum and its confidence interval using the Fast Fourier Transform’’, IEEE Trans. Audio Electroacoust AU-18 (1970) 320. [3] M.I. Cohen, Neurogenesis of respiratory rhythm in the mammal, Physiol. Rev. 59 (1979) 1105 – 1173. [4] M.I. Cohen, P.M. Gootman, Periodicities in efferent discharges of splanchnic nerve of the cat, Am. J. Physiol. 218 (1970) 1092 – 1101. [5] T. Colton, Statistics in Medicine, Little, Brown, Boston, 1974 (274 pp.). [6] G.B. Ermentrout, J. Rinzel, Beyond a pacemaker’s entrainment limit: phase walk-through, Am. J. Physiol. 246 (1984) R102 – R106. [7] P.A. Flecknell, Laboratory Animal Anaesthesia, An Introduction for Research Workers and Technicians, Academic, London, 1987 (156 pp.). [8] G.L. Gebber, Basis for phase relations between baroreceptor and sympathetic nervous discharge, Am. J. Physiol. 230 (1976) 263 – 270. [9] G.L. Gebber, M. Das, S.M. Barman, An unusual form of phase walk in a system of coupled oscillators, J. Biol. Rhythms 19 (2004) 542 – 550. [10] H.-J. Ha¨bler, W. Ja¨nig, M. Michaelis, Respiratory modulation in the activity of sympathetic neurones, Prog. Neurobiol. 43 (1994) 567 – 606. [11] F.E. Hanson, Comparative studies of firefly pacemakers, Fed. Proc. 37 (1978) 2158 – 2164. [12] J.A.S. Kelso, Dynamic Patterns, The Self-Organization of Brain and Behavior, MIT Press, Cambridge, MA, 1995 (334 pp.). [13] R.E. Kronauer, C.A. Czeisler, S.F. Pilato, M.C. Moore-Ede, E.D. Weitzman, Mathematical model of the human circadian system with two interacting oscillators, Am. J. Physiol. 242 (1982) R3 – R17. [14] P.D. Larsen, C.D. Lewis, G.L. Gebber, S. Zhong, Partial spectral analysis of cardiac-related sympathetic nerve discharge, J. Neurophysiol. 84 (2000) 1168 – 1179. [15] C.D. Lewis, G.L. Gebber, S. Zhong, P.D. Larsen, S.M. Barman, Modes of baroreceptor-sympathetic coordination, J. Neurophysiol. 84 (2000) 1157 – 1167. [16] S.C. Malpas, The rhythmicity of sympathetic nerve activity, Prog. Neurobiol. 56 (1998) 65 – 96. [17] M. Steriade, R.R. Llinas, The functional states of the thalamus and the associated neuronal interplay, Physiol. Rev. 68 (1988) 649 – 742. [18] E. von Holst, Relative coordination as a phenomenon and as a method of analysis of central nervous functions, in: R. Martin (Ed.), The Collected Papers of Erich von Holst, Univ. of Miami Press, Coral Gables, FL, 1973, pp. 33 – 135. [19] R. Wever, A mathematical model for circadian rhythms, in: J. Aschoff (Ed.), Circadian Clocks, North-Holland, Amsterdam, 1965, pp. 46 – 63. [20] S.-Y. Zhou, G.L. Gebber, S. Zhong, S.M. Barman, Pathways involved in synchronization of sympathetic nerve discharge to lung inflation, Brain. Res. 931 (2002) 107 – 116.
Research report
Dynamic changes in baroreceptor-sympathetic coupling during the respiratory cycle Gerard L. Gebber*, Mahasweta Das, Susan M. Barman Department of Pharmacology and Toxicology, Michigan State University, East Lansing, MI 48824-1317, USA Accepted 1 April 2005 Available online 24 May 2005
Abstract In urethane-anesthetized, paralyzed, and artificially ventilated cats, we observed an unusual form of ‘‘phase walk’’ of the cardiac-related burst of inferior cardiac postganglionic sympathetic nerve discharge (SND) relative to the systolic phase of the arterial pulse (AP) and thus pulse-synchronous baroreceptor nerve activity. Unlike classic phase walk ascribable to weakened coupling (desynchronization) of two oscillators, AP-SND phase walk was characterized by epochs of progressive, bidirectional changes in the angle of strong coupling (AP-SND coherence values, >0.9) of these signals that recurred on the time scale of the respiratory cycle and whose range was approximately one third of the period of the heart beat. AP-SND phase walk was linked to two respiratory variables (central respiratory drive and vagal lung inflation afferent activity) as demonstrated by the following observations. First, in five normocapnic cats (end-tidal CO2, 4.3 T 0.2%) with intact vagus nerves and three vagotomized cats, AP-SND phase walk was characterized by a progressive heart-beat-to-heart-beat decrease in the lag of SND relative to the AP during the inspiratory phase of phrenic nerve activity and an increase in the lag during the expiratory phase. Second, in three cats with intact vagus nerves that were hyperventilated (end tidal CO2, 1.6 T 0.4%) to phrenic nerve quiescence, the lag of the cardiac-related burst of SND relative to the AP increased during lung inflation and decreased during lung deflation. Additional experimentation revealed that AP-SND phase walk is attributable to respiratory-induced changes in the frequency of the centrally generated sympathetic nerve rhythm rather than heart rate. Moreover, the data demonstrate that the frequency and amplitude of the sympathetic oscillation are independently controlled by the above mentioned respiratory parameters. D 2005 Elsevier B.V. All rights reserved. Theme: Endocrine and autonomic regulation Topic: Cardiovascular regulation Keywords: Cardiorespiratory coordination; Frequency and amplitude modulation; Phase walk; Resetting of a rhythm
1. Introduction In earlier reports from our laboratory [14,15], two modes of coordination of cardiac-related sympathetic nerve discharge (SND) to pulse-synchronous baroreceptor nerve activity were described in urethane-anesthetized cats. In these experiments, the peak of the systolic phase of the arterial pulse (AP) was used to reflect the timing of pulsesynchronous baroreceptor nerve activity. One mode of
* Corresponding author. Fax: +1 517 353 8915. E-mail address: [email protected] (G.L. Gebber). 0006-8993/$ - see front matter D 2005 Elsevier B.V. All rights reserved. doi:10.1016/j.brainres.2005.04.009
baroreceptor-sympathetic coordination was characterized by phase-locking (i.e., nearly constant delay of SND relative to peak systole). The second mode was characterized by respiratory-related epochs of progressive and bidirectional heart-beat-to-heart-beat changes in the delay of SND relative to peak systole (expressed as a phase angle). We refer to this mode as an unusual type of ‘‘phase walk,’’ a similar form of which has been observed between the 10-Hz rhythmic discharges of sympathetic nerve pairs recorded in baroreceptor-denervated and vagotomized cats [9]. The current study was initiated with three purposes in mind. Our first objective was to define the respiratory variables that account for AP-SND phase walk which we
G.L. Gebber et al. / Brain Research 1046 (2005) 216 – 223
consider to reflect dynamic changes in the coupling of a central sympathetic generator to the cardiac cycle by pulsesynchronous baroreceptor discharge [14,15]. Specifically, we investigated the roles played by central respiratory drive and phasic vagal lung inflation afferent activity in producing AP-SND phase walk. The second objective was to define the mechanism of AP-SND phase walk. In particular, we investigated whether changes in AP-SND phase angle were more closely linked to respiratory-related changes in heart rate, and thus the timing of pulse-synchronous baroreceptor nerve activity or the frequency at which central circuits generate rhythmic bursts of SND. The third objective was to determine whether the respiratory-related changes in the AP-SND phase angle are linked to those of the amplitude of cardiac-related bursts of SND. Regarding this issue, respiratory modulation of sympathetic nerve burst amplitude is commonly viewed as an electrophysiological correlate of cardiorespiratory coordination [1,4,10].
217
cardiac nerve recordings were made with the Grass Model P511 preamplifier bandpass set at 1– 1000 Hz so that bursts of multiunit spikes appeared as slow waves [4]. Phrenic nerve activity (PNA) was bandpass filtered (1 –1000 Hz) after which the signal was passed through a moving averager (CWE, Inc., Model RA-821 RSP) with a time constant of 100 ms. PNA was used as an index of the central respiratory cycle. 2.3. Frequency analysis Fast Fourier transform was used to construct autospectra of SND and the AP and coherence functions (normalized cross-spectra) relating these signals [9,14]. The sampling rate of 200 Hz yielded a bin-resolution of 0.2 Hz. The spectra were averages of 32 5-s data windows with 50% overlap. Coherence values are presented on a scale of 0 – 1.0 with a value of 1.0 reflecting an ideal relationship and values 0.1, reflecting a significant relationship between the AP and SND at the frequency of the heart beat [2].
2. Methods 2.4. Time series 2.1. Anesthesia and general procedures The experimental protocols were approved by the AllUniversity Committee on Animal Use and Care of Michigan State University. The experiments were performed on five baroreceptor-intact cats anesthetized with an intravenous injection of urethane (1.4 –1.8 g/kg). With this dose, the cat is maintained in a state of surgical anesthesia for a period (8– 10 h [7]) that exceeded the duration of our experiments, as indicated by the failure of noxious stimuli (pinch, cauterization of skin and muscle) to desynchronize spindles and delta-slow wave activity in the frontal – parietal electroencephalogram [14,15,17]. The animals were artificially ventilated and paralyzed with gallamine triethiodide (4 mg/kg iv, initial dose); bilateral pneumothoracotomy was performed to minimize ventilator pump-related changes in blood pressure. Blood pressure was measured from a catheter inserted into a femoral artery, and drugs were administered into a femoral vein. End-tidal CO2 was measured by using a Traverse Medical Monitors Capnometer (model 2200). Body temperature was kept near 37 -C with a heat lamp. Intratracheal pressure (ITP) was monitored by using a Grass Instruments volumetric pressure transducer (Model PT5). Upward deflections of ITP denote inflation of the lungs by the respirator pump (Harvard, Inspira ASV). In three experiments, both vagus nerves were sectioned in the neck. 2.2. Nerve recordings As described in an earlier report [9], potentials were recorded monophasically with bipolar platinum electrodes from the central end of the left inferior cardiac postganglionic sympathetic nerve and right phrenic nerve. Inferior
As a prelude to time series analysis, the cardiac-related component in the original wide bandpass recording of SND was extracted by digital filtering. The software for the digital filter (symmetric, nonrecursive type with a Lanczos smoothing function) was obtained from RC Electronics (Santa Barbara, CA). The width of the bandpass digital filter was set at 4 Hz with the center frequency matched to that of the sharp peak at the cardiac frequency in the autospectrum of SND. The digital filter had a roll-off slope of 39%/Hz outside of the bandpass and produced minimal phase and amplitude distortion as ascertained by comparing the filtered records with the originals [9,15]. The filtered records were smoother than the originals, thereby improving the accuracy of detection of the peaks and troughs of the cardiac-related sympathetic nerve slow waves. Software written in our laboratory [14,15] was used to construct time series showing heart-beat-to-heart-beat measurements of (1) systolic blood pressure, (2) intervals between heart beats (heart period, ms), (3) peak-to-trough sympathetic nerve slow wave amplitude normalized on a scale of 0 –1.0, (4) intervals (ms) between the peaks of slow waves, and (5) phase angle between the peak of the AP and the peak of the corresponding slow wave in SND, expressed in degrees. AP-SND phase angle was measured with a resolution of 5.4-/bin (sampling period was 5 ms) when the heart period was 333 ms (heart rate, 3 Hz). Positive and negative values of phase angle refer to instances when the peak of the cardiac-related sympathetic nerve slow wave lagged or led peak systole, respectively. Whereas, for example, a lag of 240- is mathematically equivalent to a lead of 120-, values of AP-SND phase angle were plotted so that epochs of respiratory-related phase walk appear as a continuous rather than as a discontinuous curve.
218
G.L. Gebber et al. / Brain Research 1046 (2005) 216 – 223
2.5. Statistical analysis Values in the text are means T SEM. The Pearson product-moment correlation coefficient (r) was used to determine whether changes in sympathetic nerve slow wave amplitude were correlated to changes in AP-SND phase angle. Colton’s [5] rule of thumb was used to interpret the strength of correlation of these two parameters. An r value from 0 to 0.25 (or 0.25) indicates little or no relationship; one from 0.25 to 0.50 (or 0.25 to 0.50) a weak relationship; one from 0.50 to 0.75 (or 0.50 to 0.75) a moderate to good relationship; and one >0.75 (or < 0.75) indicates a very good to excellent relationship.
3. Results 3.1. Normocapnic cats with intact vagus nerves Fig. 1 shows the changes in AP-SND phase angle during one respiratory cycle (as reflected by PNA; top trace) in a normocapnic cat with intact vagus nerves. Using the vertical lines as guides, one can note that the peak of the cardiacrelated slow wave in SND (bottom trace) lagged peak systole (middle trace) by 94 ms (approximately one-third of the heart period) during expiration. The lag was progressively shortened so that peak systole and the peak of the sympathetic nerve slow wave occurred almost simultane-
Fig. 2. SND is strongly correlated to AP during respiratory-related phase walk (same experiment as in Fig. 1). Traces are (top to bottom) autospectrum (AS) of AP, AS of SND, and coherence function relating AP and SND. The spectra are averages of thirty-two 5-s windows with 50% overlap and have a resolution of 0.2 Hz/bin.
Fig. 1. Respiratory-related phase walk of cardiac-related slow wave of inferior cardiac postganglionic sympathetic nerve discharge (SND) relative to arterial pulse (AP). Oscilloscopic traces are (top to bottom) phrenic nerve activity (PNA; inspiration is upward), AP, and SND. Time scale is 0.5 s/division. Vertical lines drawn through peak systolic phase of AP illustrate the bidirectional AP-SND phase walk during one respiratory cycle.
ously (in-phase) in mid-inspiration. The lag of the slow wave relative to peak systole then progressively returned to its original value during late inspiration and the next expiratory period. In this example, changes in the amplitude of the sympathetic nerve slow wave were inversely related to the changes in AP-SND phase angle, and blood pressure did not change during the respiratory cycle. Autospectra of the AP and SND and coherence function relating the two signals during an 80-s epoch of recurring respiratory-related phase walk are shown in Fig. 2. The traces in Fig. 1 formed part of this data block. Both autospectra and the coherence function contain a sharp peak at the frequency of the heart beat. The coherence value relating SND to the AP was 0.85. Thus, despite phase walk over a considerable range (see Fig. 1), AP and SND were strongly coupled. Time series constructed from a 15-s data block collected later in the same experiment are shown in Fig. 3. At this time, respiratory-related oscillations in systolic blood pressure (SBP, top trace) had developed. As is usually the case in
G.L. Gebber et al. / Brain Research 1046 (2005) 216 – 223
Fig. 3. Respiratory-related changes in AP-SND phase angle and sympathetic nerve slow wave amplitude in a normocapnic cat with intact vagus nerves. Traces are (top to bottom) heart-beat-to-heart-beat readings of peak systolic blood pressure (SBP; mm Hg), and heart period (HP; ms), oscilloscopic traces of PNA and intratracheal pressure (ITP; inflation is upwards), and heart-beat-to-heart-beat readings of AP-SND phase angle (in degrees with SND lagging AP), and sympathetic nerve slow wave amplitude (normalized on a scale of 0 to1.0). The r value relating AP-SND phase angle to sympathetic slow wave amplitude is 0.70. Time base is 1 s/division.
artificially ventilated cats [3,20], the inspiratory phase of the central respiratory cycle (PNA, 3rd trace from top) was out of phase with lung inflation (ITP, 4th trace). The respiratoryrelated epochs of AP-SND phase walk (5th trace) were characterized by a progressive heart-beat-to-heart-beat decrease in the lag of SND relative to the AP during inspiration (which coincided with deflation of the lungs) and an increase in the lag during expiration (coinciding with lung inflation). Note that the range of the bidirectional phase walk was ¨100- (about 28% of the heart period). Moreover, AP-SND phase walk occurred in the absence of respiratoryrelated changes in heart period (2nd trace). The bottom trace in Fig. 3 shows that sympathetic nerve slow wave amplitude
219
increased during the inspiratory phase of PNA and decreased during expiration. As such, the changes in slow wave amplitude were out of phase with those in AP-SND phase angle. In this case, the r value relating slow wave amplitude to the phase angle was 0.70 ( P < 0.0001; a good relationship according to Colton’s rule of thumb [5]). We analyzed two or three 15-s episodes of recurring respiratory-related AP-SND phase walk in each of five normocapnic cats with intact vagus nerves. The total number of episodes analyzed was 14. Mean central respiratory rate was 20.2 T 1.6/min, and the central respiratory cycle was locked 1:1 to the artificial ventilation cycle. Mean blood pressure was 142 T 6 mm Hg, mean heart rate was 3.4 T 0.1 Hz, and end-tidal CO2 averaged 4.3 T 0.2%. The peak coherence value relating SND to the AP at the frequency of the heart beat averaged 0.92 T 0.01, and the range of the phase walk was 126 T 10- (approximately onethird of the heart period). In all cases, the AP-SND phase angle decreased during the inspiratory phase of PNA and increased during expiration. According to Colton [5], the strength of the inverse relationship between AP-SND phase angle and sympathetic nerve slow wave amplitude was excellent in one case (r = 0.76), good in six cases (r value range: 0.58 to 0.70), and weak in another six cases (r value range: 0.26 to 0.47). There was no correlation (r = 0.15) in the remaining case. The case in which no correlation existed is shown in Fig. 4. Note the prominent AP-SND phase walk with the peak of the sympathetic nerve slow wave lagging peak systole by as much as 40- in expiration and leading by as much as 85- in inspiration. In contrast to the case in Fig. 3, however, there were little or no respiratory-related changes in sympathetic nerve slow wave amplitude. In this experiment, respiratoryrelated changes in SBP were of small amplitude (¨4 mm Hg) and there were no respiratory-related changes in heart rate. 3.2. Normocapnic, vagotomized cats Fig. 5 shows data from a vagotomized cat. As expected in vagotomized cats, the central respiratory cycle (as reflected by PNA) was not coupled to the artificial ventilation cycle (ITP). Under these conditions, AP-SND phase walk was coupled to the central respiratory cycle, but not ITP. Note that the cardiac-related sympathetic nerve slow waves lagged peak systole by ¨70- during early expiration and led by as much as 100- during the inspiratory phase of PNA. In this case, AP-SND phase walk occurred in the absence of central respiratory-related changes in SBP and heart period, and respiratory modulation of sympathetic nerve slow wave amplitude was weak. The r value relating AP-SND phase angle and slow wave amplitude was 0.32. We analyzed two or three 15-s episodes (n = 7) of recurring AP-SND phase walk coupled to the central respiratory cycle in each of three vagotomized cats. Mean central respiratory rate was 15.3 T 2.0/min. Mean blood pressure was 138 T 3 mm Hg, and mean heart rate was 3.3 T
220
G.L. Gebber et al. / Brain Research 1046 (2005) 216 – 223
PNA although one ‘‘gasp-like’’ episode appears near the middle of the 10-s data block. The low CO2 level was induced by raising the volume of inflation from 35 to 55 ml and increasing the rate of artificial ventilation from 20 to 30/ min. Under these conditions, the cardiac-related slow wave in SND lagged peak systole by ¨270- during inflation of the lungs. The lag was decreased to a value close to 90during lung deflation. Note also that sympathetic nerve slow wave amplitude decreased during inflation and increased during deflation. In this case, the r value relating AP-SND phase angle to sympathetic nerve slow wave amplitude was 0.72, reflecting a good inverse relationship. Although ventilator pump-related changes in SBP (¨4 mm Hg) occurred in this experiment, there were essentially no respiratory-related changes in heart period. We analyzed one 10- or 15-s episode of AP-SND phase walk coupled to the artificial ventilation cycle in each of three hyperventilated cats with intact vagus nerves. Mean
Fig. 4. Respiratory-related AP-SND phase walk unrelated to changes in sympathetic nerve slow wave amplitude. The r value relating AP-SND phase angle to sympathetic slow wave amplitude is 0.15 in this normocapnic cat with intact vagus nerves. Positive and negative phase angles reflect lags and leads, respectively, of the peak of the sympathetic nerve slow wave relative to peak systole. Sequence of traces, abbreviations, and time base are same as in Fig. 3.
0.3 Hz. End-tidal CO2 averaged 4.3 T 0.2%. The peak coherence value relating SND to the AP at the frequency of the heart beat averaged 0.94 T 0.02, and the range of the respiratory-related AP-SND phase walk was 143 T 19-. In all cases, AP-SND phase angle decreased during the inspiratory phase of PNA and increased during the expiratory phase. The strength of the inverse relationship between sympathetic nerve slow wave amplitude and AP-SND phase angle was good in one case (r = 0.55) and weak in four cases (r value range: 0.25 to 0.35). These parameters were uncorrelated in the remaining two cases (r = 0.15 and 0.20). 3.3. Hyperventilated cats with intact vagus nerves Fig. 6 shows a case of AP-SND phase walk that recurred on the time scale of artificial ventilation in a hyperventilated cat (end-tidal CO2 = 1.8%). Note the absence of rhythmic
Fig. 5. Respiratory-related AP-SND phase walk in a normocapnic, vagotomized cat. The r value relating AP-SND phase angle to sympathetic slow wave amplitude is 0.32. Sequence of traces, abbreviations, and time base are same as in Fig. 3.
G.L. Gebber et al. / Brain Research 1046 (2005) 216 – 223
221
SND phase walk. This was the case independent of whether the vagus nerves were intact or not. A typical example in a cat with intact vagus nerves is shown in Fig. 7. Three respiratory-related epochs of AP-SND phase walk appear in the 3rd panel from the top. The peak of the sympathetic nerve slow wave lagged the AP by nearly 50- during the expiratory phase of PNA and led by nearly 55- during inspiration. The cycle-by-cycle changes in the interval between the peaks of adjacent sympathetic slow waves (4th panel) occurred in parallel with those in AP-SND phase angle. The range (319 –420 ms) of inter-sympathetic slow wave intervals during the respiratory cycle was equivalent to a change in the frequency of sympathetic slow wave generation from 3.1 to 2.4 Hz. In contrast to the intersympathetic slow wave interval, heart period remained essentially constant throughout the respiratory cycle (see bottom panel in Fig. 7 and additional cases in Figs. 3– 5). The absence or minimal change in heart period likely is
Fig. 6. AP-SND phase walk and changes in sympathetic nerve slow wave amplitude related to artificial ventilation cycle in a hyperventilated cat with intact vagus nerves. Rhythmic PNA was absent when end-tidal CO2 was lowered to 1.8%. The r value relating AP-SND phase angle to sympathetic slow wave amplitude is 0.72. Sequence of traces, abbreviations, and time base are same as in Fig. 3.
ventilatory rate was 33 T 2 breaths/min. Mean blood pressure was 126 T 4 mm Hg, and mean heart rate was 3.0 T 0.1 Hz. End-tidal CO2 averaged 1.6 T 0.4%. The peak coherence value relating SND to the AP at the frequency of the heart beat averaged 0.91 T 0.01, and the range of the APSND phase walk was 127 T 33-. In all cases, the lag of the cardiac-related sympathetic nerve wave increased during inflation and decreased during deflation. The r values relating sympathetic nerve slow wave amplitude to AP-SND phase angle were 0.72, 0.35, and 0.09. 3.4. Changes in heart period and inter-sympathetic nerve slow wave interval during AP-SND phase walk Respiratory-related changes in the interval between the peaks of successive cardiac-related sympathetic nerve slow waves invariably exceeded those in heart period during AP-
Fig. 7. Changes in inter-sympathetic nerve slow wave interval and heart period during respiratory-related AP-SND phase walk in a normocapnic cat with intact vagus nerves. Traces are (top to bottom) oscilloscopic traces of PNA and ITP, heart-beat-to-heart-beat readings of AP-SND phase angle (in degrees), interval (in ms) between peaks of sympathetic nerve slow waves, and heart period (in ms). Time base is 1 s/division. Abbreviations are same as in Fig. 3.
222
G.L. Gebber et al. / Brain Research 1046 (2005) 216 – 223
attributable to the vagolytic action of the neuromuscular blocking agent, gallamine.
4. Discussion Traditionally, phase walk has been attributed to weakened coupling of two oscillators or the loss of entrainment of an oscillator to a periodic forcing input [6,11 – 13,18]. Such is thought to occur when the intrinsic frequencies of the oscillators become so disparate that the coupling is no longer able to enforce synchrony. As a result, the phase angle between the oscillators cycles repeatedly in one direction through a full 360-. The phase walk of cardiacrelated SND relative to the AP described in the current study did not fit this pattern. First, the coherence value relating SND to the AP approached 1.0 during phase walk. Thus, SND was strongly coupled to pulse-synchronous baroreceptor nerve activity despite the progressive heart-beat-toheart-beat changes in the phase angle between these signals. We attribute the high coherence to the similarity of recurring epochs of the respiratory-related AP-SND phase walk. Second, respiratory-related phase walk was bidirectional rather than unidirectional. Third, rather than progressing through a full 360- (heart period), the range of the phase walk was restricted to approximately one-third of the period of the heart beat. These characteristics argue against the possibility that AP-SND phase walk arose from weakened coupling of cardiac-related SND to pulse-synchronous baroreceptor nerve activity. Rather, we propose that APSND phase walk reflected heart-beat-to-heart-beat resetting of the angle of strong coupling of a centrally generated sympathetic rhythm to forcing pulse-synchronous baroreceptor input. Because the phase walk was respiratory related, we attribute the progressive resetting of the angle of coupling to the continuously changing level of respiratory inputs to circuits responsible for baroreceptor-sympathetic coordination. We have identified two respiratory variables (central respiratory drive and vagal lung inflation afferent activity) involved in AP-SND phase walk. AP-SND phase walk was linked to each of these factors in the absence of the other. First, phase walk recurred on the time scale of the central respiratory cycle in vagotomized cats with the AP-SND phase angle decreasing during the inspiratory phase of PNA and increasing during expiration. Second, phase walk recurred on the time scale of the artificial ventilation cycle in cats with intact vagus nerves when hyperventilation eliminated rhythmic PNA. Under these conditions, the APSND phase angle increased during lung inflation and decreased during deflation. There were clear-cut cases (see Figs. 1 and 5) when AP-SND phase walk occurred in the absence of respiratory-related changes in blood pressure. Thus, AP-SND phase walk could occur independently of cyclic changes in the strength of pulse-synchronous baroreceptor nerve activity linked to respiratory-related
oscillations in blood pressure. Nonetheless, it is possible that changes in baroreceptor nerve activity linked to respiratory-related blood pressure oscillations (see Figs. 3, 4, and 6) contributed to AP-SND phase walk in many cases. Time series analysis revealed progressive shortening of the inter-sympathetic slow wave interval during the inspiratory phase of PNA and progressive lengthening during expiration (see Fig. 7). Thus, the frequency of the sympathetic nerve rhythm increased during inspiration and decreased during expiration. In most cases, such changes occurred in the absence of respiratory-related fluctuations in heart period. These data suggest that the mechanism of APSND phase walk involved respiratory-related changes in the excitability of the central sympathetic rhythm generator. Increased excitability during the inspiratory phase of the central respiratory cycle would shorten the inter-sympathetic nerve burst interval, thereby leading to a progressive heartbeat-to-heart-beat advancement of the timing of the peak of cardiac-related sympathetic activity relative to pulse-synchronous baroreceptor nerve activity. Under the conditions of a constant heart period, AP-SND phase angle would decrease progressively. This follows from the work of Wever [19] on the entrainment of circadian rhythms to the light-dark cycle. Wever [19] demonstrated that the higher the intrinsic frequency (shorter period) of the central rhythm, the more it leads (or the less it lags) an entraining light pulse. Changes in phase angle occur in the opposite direction (increased lag or decreased lead) when the intrinsic frequency of the central rhythm is lowered (increased period). Thus, as might be expected, we observed a progressive increase in AP-SND phase angle when the inter-sympathetic nerve slow wave interval increased from heart-beat-to-heart-beat during the expiratory phase of the central respiratory cycle. Vagal lung inflation afferent activity appeared to exert an effect on central sympathetic excitability opposite to that of central inspiratory drive. During inflation in normocapnic cats (Figs. 3 and 4), AP-SND phase angle increased progressively. AP-SND phase angle also increased during inflation in hyperventilated cats when rhythmic PNA was lost (Fig. 6). The increase in phase angle is presumed to reflect decreased excitability of the sympathetic generator leading to an increase in the period of the rhythm (i.e., lowered frequency) and thus lengthening of the AP-SND phase angle. It is unlikely that AP-SND phase angle served as an important determinant of sympathetic nerve slow wave amplitude under the conditions of our experiments. Rather, during phase walk, it appears that these parameters were independently controlled, perhaps by the actions of different groups of respiratory neurons mediated at different levels (oscillator versus follower circuits) of the networks controlling SND. Regarding this point, the inverse relationship between AP-SND phase angle and sympathetic nerve slow wave amplitude was weak (r value between 0.25 and 0.50) or absent (r value between 0 and 0.25) in the
G.L. Gebber et al. / Brain Research 1046 (2005) 216 – 223
majority of the cases. examined in cats with intact or sectioned vagus nerves. In some cases, prominent respiratory-related AP-SND phase walk was observed with little or no sign of amplitude modulation on the same time scale (Fig. 5). In view of the proposed mechanism for the respiratory-related changes in AP-SND phase angle (see above), it follows that, at least to some extent, the frequency of sympathetic nerve bursts can be regulated independently of burst amplitude. Quite obviously, changes in either of these parameters will influence end organ function. Additional examples of differential or independent control of sympathetic nerve burst frequency and amplitude have been cited by Malpas [16]. Whereas the physiological significance of heart-beat-toheart-beat changes in the angle of strong coupling of SND to pulse-synchronous baroreceptor activity during the respiratory cycle remains obscure, the phenomenon of bidirectional phase walk supports the view that the cardiac-related rhythm in SND reflects the entrainment of a central oscillator by an extrinsic, periodic forcing input rather than the periodic inhibition of randomly generated central activity [8,14,15]. If the cardiac-related rhythm in SND was the simple consequence of the waxing and waning of constant latency central inhibition locked to pulse-synchronous baroreceptor activity, little or no change in the AP-SND phase angle would be expected during the respiratory cycle. Regarding this point, Lewis et al. [15] have demonstrated that the extent of AP-SND phase walk, which was equivalent to between one-third and one-half of the heart period, is too large to be explained by the range of axonal conduction velocities of bulbospinal and preganglionic neurons that provide driving input to the inferior cardiac postganglionic nerve. Perhaps the most convincing evidence for the entrainment of a central rhythm generator by pulsesynchronous baroreceptor activity is provided by epochs of recurring respiratory-related phase walk in which the lag of the peak of the sympathetic nerve slow wave relative to peak systole was reversed to a lead (see Figs. 4, 5, and 7). In these cases, the period of reduced SND (start of falling phase of slow wave) could actually precede the onset of systole. This occurred at points in the respiratory cycle when the peak of the slow wave led peak systole by approximately one-fourth of the heart period. During these heart beats, the period of reduced SND began before pulsesynchronous baroreceptor nerve activity was initiated. In summary, the current study revealed new information on the neural mechanisms responsible for cardiorespiratory coordination. First, we demonstrated that the dynamic changes in baroreceptor-sympathetic coupling (AP-SND phase walk) occurring during the respiratory cycle are attributable to progressive changes in the excitability of a central sympathetic rhythm generator rather than the timing of pulse-synchronous baroreceptor nerve activity. Specifically, we propose that AP-SND phase walk reflected heartbeat-to-heart-beat resetting of the angle of strong coupling of a centrally generated sympathetic rhythm to forcing
223
pulse-synchronous baroreceptor input by continuously changing levels of central respiratory drive and vagal lung inflation afferent activity. Second, we demonstrated independent control of the frequency and amplitude cardiacrelated sympathetic nerve bursts by each of the two above mentioned respiratory variables.
Acknowledgment This study was supported by National Heart, Lung, and Blood Institute Grant HL-33266.
References [1] E.D. Adrian, D.W. Bronk, G. Phillips, Discharges in mammalian sympathetic nerves, J. Physiol. (Lond.) 74 (1932) 115 – 133. [2] V.A. Benignus, Correction to ‘‘Estimation of the coherence spectrum and its confidence interval using the Fast Fourier Transform’’, IEEE Trans. Audio Electroacoust AU-18 (1970) 320. [3] M.I. Cohen, Neurogenesis of respiratory rhythm in the mammal, Physiol. Rev. 59 (1979) 1105 – 1173. [4] M.I. Cohen, P.M. Gootman, Periodicities in efferent discharges of splanchnic nerve of the cat, Am. J. Physiol. 218 (1970) 1092 – 1101. [5] T. Colton, Statistics in Medicine, Little, Brown, Boston, 1974 (274 pp.). [6] G.B. Ermentrout, J. Rinzel, Beyond a pacemaker’s entrainment limit: phase walk-through, Am. J. Physiol. 246 (1984) R102 – R106. [7] P.A. Flecknell, Laboratory Animal Anaesthesia, An Introduction for Research Workers and Technicians, Academic, London, 1987 (156 pp.). [8] G.L. Gebber, Basis for phase relations between baroreceptor and sympathetic nervous discharge, Am. J. Physiol. 230 (1976) 263 – 270. [9] G.L. Gebber, M. Das, S.M. Barman, An unusual form of phase walk in a system of coupled oscillators, J. Biol. Rhythms 19 (2004) 542 – 550. [10] H.-J. Ha¨bler, W. Ja¨nig, M. Michaelis, Respiratory modulation in the activity of sympathetic neurones, Prog. Neurobiol. 43 (1994) 567 – 606. [11] F.E. Hanson, Comparative studies of firefly pacemakers, Fed. Proc. 37 (1978) 2158 – 2164. [12] J.A.S. Kelso, Dynamic Patterns, The Self-Organization of Brain and Behavior, MIT Press, Cambridge, MA, 1995 (334 pp.). [13] R.E. Kronauer, C.A. Czeisler, S.F. Pilato, M.C. Moore-Ede, E.D. Weitzman, Mathematical model of the human circadian system with two interacting oscillators, Am. J. Physiol. 242 (1982) R3 – R17. [14] P.D. Larsen, C.D. Lewis, G.L. Gebber, S. Zhong, Partial spectral analysis of cardiac-related sympathetic nerve discharge, J. Neurophysiol. 84 (2000) 1168 – 1179. [15] C.D. Lewis, G.L. Gebber, S. Zhong, P.D. Larsen, S.M. Barman, Modes of baroreceptor-sympathetic coordination, J. Neurophysiol. 84 (2000) 1157 – 1167. [16] S.C. Malpas, The rhythmicity of sympathetic nerve activity, Prog. Neurobiol. 56 (1998) 65 – 96. [17] M. Steriade, R.R. Llinas, The functional states of the thalamus and the associated neuronal interplay, Physiol. Rev. 68 (1988) 649 – 742. [18] E. von Holst, Relative coordination as a phenomenon and as a method of analysis of central nervous functions, in: R. Martin (Ed.), The Collected Papers of Erich von Holst, Univ. of Miami Press, Coral Gables, FL, 1973, pp. 33 – 135. [19] R. Wever, A mathematical model for circadian rhythms, in: J. Aschoff (Ed.), Circadian Clocks, North-Holland, Amsterdam, 1965, pp. 46 – 63. [20] S.-Y. Zhou, G.L. Gebber, S. Zhong, S.M. Barman, Pathways involved in synchronization of sympathetic nerve discharge to lung inflation, Brain. Res. 931 (2002) 107 – 116.