Transfer function characteristics of the neural and peripheral arterial baroreflex arcs at rest and during postexercise muscle ischemia in humans

Abstract

Previous studies have demonstrated an increase in the arterial baroreflex (ABR) control of muscle sympathetic nerve activity (MSNA) during isolated activation of the muscle metaboreflex with postexercise muscle ischemia (PEMI). However, the increased ABR-MSNA control does not appear to manifest in an enhancement in the ABR control of arterial blood pressure (BP), suggesting alterations in the transduction of MSNA into a peripheral vascular response and a subsequent ABR-mediated change in BP. Thus we examined the operating gains of the neural and peripheral arcs of the ABR and their interactive relationship at rest and during muscle metaboreflex activation. In nine healthy subjects, graded isolation of the muscle metaboreflex was achieved by PEMI following isometric handgrip performed at 15% and 30% maximal voluntary contraction (MVC). To obtain the sensitivities of the ABR neural and peripheral arcs, the transfer function gain from BP to MSNA and MSNA to femoral vascular conductance, respectively, was analyzed. No changes from rest were observed in the ABR neural or peripheral arcs during PEMI after 15% MVC handgrip. However, PEMI following 30% MVC handgrip increased the low frequency (LF) transfer function gain between BP and MSNA (ABR neural arc; +58 ± 28%, P = 0.036), whereas the LF gain between MSNA and femoral vascular conductance (ABR peripheral arc) was decreased from rest (−36 ± 8%, P = 0.017). These findings suggest that during high-intensity muscle metaboreflex activation an increased ABR gain of the neural arc appears to offset an attenuation of the peripheral arc gain to help maintain the overall ABR control of systemic BP.

the arterial baroreflex (ABR) comprises mechanoreceptors that respond to beat-to-beat changes in arterial blood pressure (BP) by reflexively altering autonomic neural outflow to adjust cardiac output and total vascular conductance (26). ABR function during exercise has been extensively investigated, and it is well established that the ABR control of BP is reset without any changes in reflex sensitivity (i.e., gain) during both dynamic (7, 25, 27–29, 31, 32) and isometric (6, 9, 30, 33) exercise. However, Scherrer et al. (34) demonstrated that the ABR more effectively buffers reflex increases in muscle sympathetic nerve activity (MSNA) during isometric handgrip exercise (IHG) compared with rest. In addition, several studies have reported that ABR-MSNA sensitivity is increased during both IHG and postexercise muscle ischemia (PEMI), suggesting that activation of the muscle metaboreflex contributes to an increased sensitivity of the ABR control of MSNA (4, 13, 15). Importantly, given that ABR control of BP has been documented to reset without any changes in sensitivity during IHG and PEMI (6, 9, 10, 36), it appears that the increased ABR-MSNA sensitivity does not manifest in an enhancement in the ABR control of BP under these conditions.

Several studies have demonstrated that the capacity of the ABR to regulate BP depends critically on its ability to alter vascular conductance via the sympathetic nervous system both at rest and during exercise (19, 20, 27, 40). Indeed, changes in total vascular conductance account for 80–90% of the ABR-mediated BP response (27). Thus, although the ABR control of MSNA represents a critical component of the ABR loop, it is clear that transduction of sympathetic outflow to a vascular smooth muscle response is paramount for the regulation of BP.

In this regard, the ABR can be divided into two principal arcs: a neural and a peripheral arc (14, 17; Fig. 1). Recent animal investigations have demonstrated that transfer function analyses for the closed-loop identification of these baroreflex arcs can estimate the open-loop baroreflex characteristics (14, 17). Thus the operating gain of the neural arc can be evaluated from the transfer function gain of the relationship between arterial BP and efferent sympathetic nerve activity (i.e., MSNA), whereas the peripheral arc gain can be evaluated from the transfer function gain of the relationship between efferent sympathetic nerve activity and the nerve-effector junction, which for the control of BP is primarily the transduction of MSNA to a change in vascular smooth muscle tone [i.e., vascular conductance (22, 39)]. However, to date, no studies have attempted to identify the interactions between the neural and peripheral baroreflex arcs in humans and how these may be modified by exercise.

Fig. 1.

Schematic representation of the arterial baroreflex (ABR) neural and peripheral arcs. The ABR system is an integrated negative-feedback system that stabilizes systemic arterial blood pressure against pressure perturbations primarily through reflex changes in the sympathetic nervous system. The ABR loop consists of a fast neural arc and a slow mechanical arc (i.e., peripheral arc). The neural arc consists of the sensing of arterial blood pressure by the baroreceptors (input) to an efferent sympathetic nerve response (output), whereas the peripheral arc consists of the transduction of sympathetic efferent nerve activity (input) to a change in vascular smooth muscle tone and, thus, femoral blood velocity (FBV; output).

With this background in mind, the purpose of the present study was to examine the operating gains of the neural and peripheral arcs of the ABR and their relationships to the overall control of BP in humans at rest and during muscle metaboreflex activation. Because the sensitivity for the ABR control of MSNA has been shown to be increased during isometric exercise (13, 15), while the ABR control of BP was well maintained (6, 9, 30, 33), we reasoned that the elevated ABR-MSNA sensitivity (i.e., neural baroreflex arc) may be compensating for an attenuation in peripheral vascular responses (i.e., peripheral baroreflex arc) to maintain systemic arterial BP regulation during isometric exercise. Since the muscle metaboreflex has been identified as the primary mechanism for increasing ABR-MSNA sensitivity during isometric exercise (4, 13, 15), we examined the sensitivities of the ABR neural and peripheral arcs during periods of PEMI to isolate the muscle metaboreflex from the potential influence of both central command and the muscle mechanoreflex (23, 24). Importantly, this also affords the advantage of applying transfer function analysis under relatively steady-state arterial BP and MSNA conditions compared with the progressive increases observed during isometric exercise. We tested the hypothesis that activation of the muscle metaboreflex during PEMI attenuates the sensitivity of the peripheral baroreflex arc in parallel with an increase in the sensitivity of the neural baroreflex arc.

METHODS

Nine healthy subjects with a mean age of 25 ± 1 yr, height of 180 ± 3 cm, and weight of 78 ± 2 kg (means ± SE) were recruited for voluntary participation in the present study. All subjects were free of any known cardiovascular or respiratory diseases and were currently not taking prescribed or over-the-counter medications. Each subject received a verbal and written explanation of the study objectives, measurement techniques, and the risks and benefits associated with the investigation and provided written informed consent as approved by the Institutional Review Board at the University of Missouri and the Research and Development committee at the Harry S. Truman Memorial Veterans′ Hospital. All experiments were performed in accordance with the Declaration of Helsinki. Prior to the actual experimental day, each subject was familiarized with the equipment and the study protocol. The subjects were requested to abstain from caffeinated beverages for 12 h and strenuous physical activity and alcohol intake for at least 24 h before any testing.

Experimental measurements.

All studies were performed at a constant room temperature between 23° and 24°C with external stimuli minimized. Heart rate (HR) was monitored using a lead II electrocardiogram (ECG). Beat-to-beat arterial BP was monitored noninvasively by a servocontrolled finger photoplethysmograph (Finometer; Finapres Medical Systems, Amsterdam, The Netherlands) placed on the middle finger of the left hand. In addition, an automated sphygmomanometer (SunTech; Medical Instruments, Raleigh, NC) was used to validate the Finometer measurements. Respiratory movements were monitored using a strain-gauge pneumograph placed in a stable position over the abdomen (Pneumotrace; UFI, Morro Bay, CA). Ratings of perceived exertion were obtained using the standard 6–20 Borg scale (2).

Multiunit postganglionic MSNA was recorded with standard microneurographic techniques, as described in detail previously (7, 29, 38). Briefly, a unipolar tungsten microelectrode was inserted into muscle fascicles of the peroneal nerve near the fibular head of the left leg. Neural signals were amplified, filtered (bandwidth, 700–2,000 Hz), rectified, and integrated (time constant, 0.1 s) to obtain mean voltage neurograms. MSNA recordings were identified by their pulse-synchronous burst pattern and increased burst frequency with end-expiratory breath holds and Valsalva maneuvers without any responses to arousal or skin stroking.

Leg blood flow velocity of the right femoral artery was obtained using a Doppler ultrasound unit (model MD6; D. E. Hokanson, Bellevue, WA) equipped with a bidirectional probe operating at a frequency of 5 MHz and calculated using the formula V = f_a/(64.9cosØ), where f_a is the audio frequency, Ø is the angle of insonation, and V is the blood velocity in centimeters per second. The Doppler probe was placed on the skin over the left common femoral artery distal to the inguinal ligament. During the experimental sessions we spent considerable time to determine the probe placement for each subject that provided the best signal quality, which typically occurs at an angle that approximates 60 degrees. However, the exact angle was not able to be determined with the Doppler unit used. To keep the probe position at a constant insonation angle for extended periods, a custom-built holder was used. This holder secured the probe to the subject's leg and was fixed in place and fastened to the medical examination table on which the subject was sitting. Stability of the probe placement was made easier with the legs being inactive for all protocols. In pilot studies, ultrasound imaging of femoral artery diameter was also measured using a 2.5-MHz probe (model RT 6800; GE). Average femoral artery diameter was determined in three subjects at rest (0.72 ± 0.01 cm), during handgrip (0.72 ± 0.02 cm), and during PEMI (0.72 ± 0.02 cm). These findings are in agreement with previous studies demonstrating that common femoral artery diameter did not change significantly in response to IHG (35). Thus femoral blood flow was not directly calculated, and transfer function analysis was performed using the femoral blood velocity (FBV) signal. In addition, a femoral vascular conductance index (FVCi) was calculated by the following equation: FVCi = FBV/mean BP. All data signals were connected to Powerlab (ADInstruments, Bella Vista, NSW, Australia) interfaced with a personal computer and stored for offline analysis.

Experimental protocols.

On the experimental day, the subjects arrived at the laboratory at least 2 h after a light meal. Subjects were seated in a semirecumbent position on a medical examination table with a handgrip dynamometer (model 78010; Lafayette Instrument, Lafayette, IN) held in the right hand while the limb was supported on an adjustable bedside table. Maximal voluntary contraction (MVC) was determined as the highest force produced during three to five maximal efforts, each separated by 1 min. For the experimental protocol, subjects performed IHG at 15% and 30% MVC followed by 3 min of PEMI. Handgrip trials were performed in random order and separated by at least 15 min to ensure re-establishment of baseline mean BP, HR, and MSNA before commencing the subsequent trial. Each bout of handgrip was preceded by a 10-min baseline period in which the subjected rested quietly, after which subjects were instructed to start the IHG and maintain the desired force for 3 min. Visual feedback of the force exerted by the subject, expressed as a percentage of maximum, was displayed on a computer screen positioned in front of the subject at eye level. Five seconds before the cessation of IHG, an occlusion cuff placed over the upper arm was inflated to suprasystolic pressure (>220 mmHg) and remained inflated for 3 min and 15 s. The additional 15 s was used to ensure 3 min of steady-state data for the PEMI period because cessation of IHG causes a robust and transient decrease in BP and MSNA. After PEMI, data was collected for 3 min of recovery.

Data analysis.

Because BP and MSNA progressively increase throughout a bout of IHG exercise (13, 23), all dynamic analyses (i.e., transfer function analysis) used to determine the sensitivities of the ABR neural and peripheral arcs as well as the spontaneous ABR-MSNA sensitivity measures were performed during the relatively steady state of BP and MSNA during PEMI. In addition, since the muscle metaboreflex has been identified as the primary mechanism for increasing ABR-MSNA sensitivity during IHG (13, 23), the period of PEMI allows for the isolation of the influence of the muscle metaboreflex independent of neural inputs from central command and skeletal muscle mechanoreceptors (23, 24).

ABR neural and peripheral arcs analyses.

The ECG signal, arterial BP, MSNA, and FBV waveforms were originally sampled at 1,000 Hz. Three-minute data segments of the BP, MSNA, and FBV waveform signals and calculated FVCi at rest and during PEMI were resampled at 5 Hz and submitted to fast Fourier transformation (FFT) and transfer function analyses to identify the operating gains of the ABR neural and peripheral arcs under each condition. The mean beat-to-beat values for each variable were also derived and processed in a similar manner, and FFT and transfer function analysis was then performed. No major differences were found in the LF transfer function information derived from the mean data compared with the waveform analyses and, therefore, our conclusions would be the same with either analysis. Thus, for brevity reasons, only the analysis with the raw data is presented. Figure 1 presents a schematic outlining the components of the neural and peripheral arcs. To obtain the operating gain of the neural arc, the transfer function from BP as an input (x) to MSNA as an output variable (y) was analyzed, whereas the gain of the peripheral arc was evaluated by the transfer function from MSNA as an input (x) to FBV or FVCi as an output variable (y). The transfer function H(f) between the two signals was calculated as: H(f) = S_xy (f)/S_xx(f), where S_xx (f) is the autospectrum of input signal and S_xy(f) is the cross-spectrum between the two signals. The transfer function magnitude |H(f)| was obtained from the real part H_R (f) and imaginary part H_I (f) of the complex transfer function: |H(f)| = {[H_R (f)]² + [H_I (f)]²}. The squared coherence function (MSC) (f) was estimated as MSC (f) = |S_xy (f)|²/[S_xx(f)S_yy(f)], where S_yy(f) is the autospectrum of output signal. The transfer function gain reflects the relative amplitude relationship between the changes in two variables over a specified frequency range. The squared coherence function reflects the fraction of output power that can be linearly related to the input power at each frequency. Similar to a correlation coefficient, it varies between 0 and 1. Spectral power, mean value of transfer function gain, phase and coherence were calculated in the very low frequency (VLF; 0.02 to 0.07 Hz), LF (0.07 to 0.20 Hz), and high frequency (HF; 0.20 to 0.30 Hz) ranges to reflect different patterns of the dynamic neural-cardiovascular relationships. However, we primarily relied on the LF range since this range has previously been reported to primarily reflect ABR mechanisms (28). In contrast, the VLF range for both blood flow and BP variability appears to reflect multiple physiological mechanisms that confound interpretation and the BP fluctuations in the HF range appear to be mainly induced by respiration (3, 5). Indeed, the average respiratory frequency of the subjects in the present study was 0.25 ± 0.02 Hz.

MSNA analysis.

MSNA bursts were identified from the mean voltage neurogram using a customized computer program employing fixed criteria, which accounted for the latency from the R wave of the ECG to the sympathetic burst (8) and incorporated a signal-to-noise ratio of at least 3:1. Computer-identified bursts were subsequently evaluated and confirmed by an experienced investigator. The burst with the largest amplitude during the rest period was allocated a value of 100 (arbitrary units), and then all bursts within a trial were normalized with respect to this standard in each subject. MSNA burst frequency (bursts/min) and total activity (i.e., burst frequency × mean burst amplitude) were calculated. In addition, as described in detail below, MSNA measurements of burst incidence (bursts/100 heartbeats) and total MSNA were used for the assessment of spontaneous ABR control of MSNA.

ABR-MSNA sensitivity analysis.

ABR control of MSNA was also evaluated by analyzing the relationship between spontaneously occurring variations in diastolic blood pressure (DBP) and MSNA (13, 15, 18, 21, 29). As described in detail previously, these analyses can allow for the assessment of ABR-MSNA sensitivity around the operating point representing the neural arc component of the ABR. Due to the close correlation between changes in MSNA and DBP, DBP was used for this analysis (37). Briefly, the diastolic pressures for each cardiac cycle during an experimental phase (i.e., rest and PEMI) were grouped into 1-mmHg intervals (i.e., bins). Importantly, the number of pressure bins used was consistent among the conditions studied, with 16 ± 1 bins at rest and 15 ± 2 and 15 ± 1 bins during PEMI following IHG at 15% and 30% MVC, respectively. The burst incidence within each DBP bin was calculated by determining the percentage of diastoles that was associated with an MSNA burst and expressed as bursts per 100 heartbeats. The total MSNA was determined for each diastolic pressure bin by calculating the average MSNA associated with each bin (i.e., total burst area of all cardiac cycles within a given diastolic pressure bin divided by the number of cardiac cycles that occurred within that bin) and expressed as total MSNA per beat. In this way, the averaged MSNA for analyzing total MSNA included heartbeats in which a sympathetic burst did not occur.

The calculated burst incidence and total MSNA obtained within each diastolic pressure bin were plotted against the corresponding DBP, and a linear regression was applied. The resulting slope of this relationship was used to provide a measure of the sensitivity of the ABR control of each variable. The data were weighted for linear regression analysis to account for the number of cardiac cycles within each diastolic pressure bin, thus removing bias due to bins containing very few cardiac cycles, which are typically on the ends of the data segment (i.e., high-pressure bins with no sympathetic bursts) (18, 21, 29). In this regard, DBP bins containing zero MSNA were included in the linear regression analysis to maintain consistency among conditions and to avoid introducing subjectivity to the analyses. Importantly, with the use of weighting of the data, zeroes did not affect the overall slope as much as would be the case if the regression was run on the mean data points for each bin. Since poor fits could increase the chance of committing a type II error when comparing slopes at rest with those obtained during PEMI, a minimum r value of 0.5 was used as the criterion for accepting slopes. For total MSNA, r values were 0.61 ± 0.09 at rest and 0.63 ± 0.07 and 0.69 ± 0.06 during PEMI following 15% and 30% MVC IHG, respectively. Results were similar for burst incidence.

Statistical analysis.

Statistical comparisons of all neural-cardiovascular parameters, the ABR neural and peripheral arcs transfer parameters, and the spontaneous ABR-MSNA sensitivity measures were made utilizing two-way repeated-measures ANOVA. After ANOVA analyses, a Student-Newman-Keuls test was employed post hoc to identify significant differences between each condition. Statistical significance was set at P < 0.05, and results are presented as means ± SE. Analyses were conducted using SigmaStat (Jandel Scientific Software; SPSS, Chicago, IL).

RESULTS

Cardiovascular responses to IHG and PEMI.

Tables 1 and 2 show the cardiovascular, hemodynamic, and sympathetic nerve activity variables obtained at rest, during IHG at 15% and 30% MVC, and at subsequent periods of PEMI. HR, FBV, and MSNA did not significantly increase during IHG at 15% MVC; however, systolic BP (+10 ± 4 mmHg), DBP (+9 ± 2 mmHg), and mean BP (+12 ± 2 mmHg) were all significantly increased and remained elevated during PEMI. In contrast, compared with rest, IHG at 30% MVC significantly increased HR, FBV, and MSNA, as well as systolic BP, DBP, and mean BP. During the period of PEMI following IHG at 30% MVC, HR returned to resting values (P = 0.419), whereas all other variables remained significantly elevated from rest (systolic BP, +29 ± 4 mmHg; DBP, +20 ± 2 mmHg; mean BP, +27 ± 2 mmHg; FBV, +2.18 ± 1.14 cm/s; MSNA burst frequency, +19 ± 3 bursts/min; and MSNA burst incidence, +28 ± 6 burst/100 heartbeats). In addition, these variables were all significantly greater than the 15% MVC condition.

Table 1.

Cardiovascular variables at rest, during IHG at 15% and 30% MVC, and subsequent PEMI

	15% MVC			30% MVC
	Rest	IHG	PEMI	Rest	IHG	PEMI
HR, beats/min	68±3	70±3	69±3	69±3	85±4*†	71±3‡
SBP, mmHg	116±3	125±2†	124±2†	118±3	146±3*†	147±3*†
DBP, mmHg	63±1	72±2†	72±2†	65±1	86±1*†	85±2*†
MBP, mmHg	81±1	92±2†	92±2†	83±1	110±2*†	110±3*†
RPE		11±0.4			16±0.3*

Values are means ± SE. IHG, isometric handgrip; PEMI, postexercise muscle ischemia; MVC, maximal voluntary contraction; HR, heart rate; SBP, systolic blood pressure; DBP, diastolic blood pressure; MBP, mean blood pressure; RPE, rating of perceived exertion.

^†P < 0.05 compared with rest.

^*P < 0.05 compared with 15% MVC.

^‡P < 0.05 compared with IHG.

Table 2.

Hemodynamic and sympathetic nerve activity measurements at rest, during IHG at 15% and 30% MVC, and subsequent PEMI

	15% MVC			30% MVC
	Rest	IHG	PEMI	Rest	IHG	PEMI
FBV, cm/s	11.3±1.9	13.2±2.0	13.6±2.5	12.1±1.6	15.9±2.2*†	14.3±2.4†
FVCi, cm·s−1·mmHg−1	0.14±0.02	0.14±0.02	0.15±0.03	0.15±0.02	0.15±0.02	0.13±0.02
MSNA burst frequency, bursts/min	23.6±4.0	25.6±4.1	28.6±3.5	21.0±2.8	42.2±5.1*†	40.1±4.0*†
MSNA burst incidence, bursts/100 heartbeats	39.4±5.9	36.2±6.1	42.5±5.2	29.1±3.3	48.7±5.0*†	57.2±6.3*†‡
Mean burst strength, au	16.7±1.0	16.7±0.7	17.5±1.2	15.6±0.7	19.3±1.1†	21.3±1.4*†
Total Activity	365±67	428±72	515±77	333±48	820±117*†	872±113*†

Values are means ± SE. FBV, femoral blood velocity; FVCi, femoral vascular conductance index; MSNA, muscle sympathetic nerve activity; au, arbitrary units.

^†P < 0.05 compared with rest;

^*P < 0.05 compared with 15% MVC;

^‡P < 0.05 compared with IHG.

ABR neural arc.

Figure 2 summarizes the sensitivity (i.e., operating gain) and coherence values for the ABR neural arc derived from the transfer function analysis between BP and MSNA at rest and during PEMI following 15% and 30% MVC IHG. The operating gain of the ABR neural arc in all frequency ranges was similar to rest during PEMI after IHG at 15% MVC. However, PEMI following IHG at the 30% MVC significantly increased the transfer function gain of the neural arc across all frequencies examined (VLF, +73 ± 32%; LF, +58 ± 28%; and HF, +93 ± 39%; all P < 0.05; Fig. 2, A and B). Coherence values did not change from rest to PEMI following either 15% or 30% MVC IHG (Fig. 2C). Importantly, the LF coherence between BP and MSNA remained >0.5 during both periods of PEMI in all subjects (range from 0.53 to 0.75). In contrast, coherence in the HF range (0.37 ± 0.03) was consistently lower than the VLF or LF ranges under all conditions.

Fig. 2.

A: frequency domain analysis of the ABR neural arc gain in the range from 0 to 0.5 Hz in 1 subject at rest (solid line) and during postexercise muscle ischemia (PEMI; dashed line) after 30% maximal voluntary contraction (MVC) isometric handgrip exercise. B: group-averaged transfer function gain between arterial blood pressure and muscle sympathetic nerve activity (MSNA; ABR neural arc) at rest and during PEMI after 15% and 30% MVC isometric handgrip. C: group-averaged coherence values for the ABR neural arc at rest and during PEMI after 15% and 30% MVC isometric handgrip. †P < 0.05 compared with rest; *P < 0.05 compared with 15% MVC. VLF, very low frequency; LF, low frequency; HF, high frequency.

In addition to the transfer function analysis, the ABR neural arc was also estimated from the linear regression of the relationship between spontaneously occurring variations in DBP and MSNA. From rest to PEMI, the linear relationship between MSNA burst incidence or total MSNA and DBP was relocated rightward (Fig. 3A), indicating a resetting of ABR control of MSNA to the higher BP during PEMI. More importantly, the slope of the linear regression line between MSNA burst incidence or total MSNA and DBP was not significantly different from rest to PEMI after IHG at 15% MVC, whereas both slopes were significantly increased from rest to PEMI following IHG at 30% MVC (Fig. 3). Furthermore, this increased sensitivity of ABR control of MSNA burst incidence and total MSNA was significantly related to the ABR neural arc gain derived from transfer function analysis between BP and MSNA. Indeed, there was a strong linear relationship between LF transfer function neural arc gain and the slope of the linear regression line of total MSNA-DBP (r = 0.709; P < 0.001) and MSNA burst incidence-DBP (r = 0.738; P < 0.001).

Fig. 3.

A: linear relationship between total MSNA and diastolic blood pressure (DBP) from 1 subject at rest (•) and during PEMI (○) after 30% MVC isometric handgrip. B and C: group summary data for the slopes of the linear regression lines between total MSNA or MSNA burst incidence and DBP at rest (black columns) and during PEMI (white columns) after 15% and 30% MVC isometric handgrip. †P < 0.05 compared with rest.

ABR peripheral arc.

Figures 4 and 5 summarize the sensitivity (i.e., operating gain) and coherence values for the ABR peripheral arc derived from the transfer function analysis between MSNA and FBV or FVCi, respectively, at rest and during PEMI following 15% and 30% MVC IHG. During PEMI after IHG at 15% MVC, the operating gain of the ABR peripheral arc in all frequency ranges was not different from rest. In contrast, PEMI following IHG at 30% MVC reduced the LF transfer function gain of the peripheral arc calculated using either FBV (−33 ± 7%; P = 0.039) or FVCi (−36 ± 8%; P = 0.017). Coherence values between MSNA and FBV or FVCi did not change from rest to PEMI following either 15% or 30% MVC IHG. In addition, the LF coherences remained >0.5 during PEMI in all subjects. Interestingly, the LF transfer function phase of the ABR neural arc was significantly less than the phase of ABR peripheral arc, suggesting that the latency of the ABR response of the neural arc was faster than that of the peripheral arc (LF neural arc phase 2.89 ± 0.38 radian vs. LF FBV peripheral arc phase 5.22 ± 0.13 radian, P < 0.001; or LF FVCi peripheral arc phase 4.81 ± 0.28 radian, P = 0.001). In addition to the MSNA and FBV or FVCi relationships, the ABR peripheral arc was also estimated from the transfer function of MSNA to BP. In contrast with the results for MSNA to FBV or FVCi, there was no significant difference in the ABR gain of peripheral arc calculated by the transfer function from MSNA to BP between rest and PEMI following 15% and 30% MVC (VLF, P = 0.243; LF, P = 0.500; HF, P = 0.235; data not presented).

Fig. 4.

A: frequency domain analysis of the ABR peripheral arc gain in the range from 0 to 0.5 Hz in 1 subject at rest (solid line) and during PEMI (dashed line) after 30% MVC isometric handgrip exercise. B: group-averaged transfer function gain between MSNA and FBV (ABR peripheral arc) at rest and during PEMI after 15% and 30% MVC isometric handgrip. C: group-averaged coherence values for the ABR FBV peripheral arc at rest and during PEMI after 15% and 30% MVC isometric handgrip. †P < 0.05 compared with rest.

Fig. 5.

A: frequency domain analysis of the ABR peripheral arc gain in the range from 0 to 0.5 Hz in 1 subject at rest (solid line) and during PEMI (dashed line) after 30% MVC isometric handgrip exercise. B: group-averaged transfer function gain between MSNA and femoral vascular conductance index (FVCi; ABR peripheral arc) at rest and during PEMI after 15% and 30% MVC isometric handgrip. C: group-averaged coherence values for the ABR FVCi peripheral arc at rest and during PEMI after 15% and 30% MVC isometric handgrip. †P < 0.05 compared with rest.

DISCUSSION

The present study is the first to identify a quantifiable difference between the operating gains of the neural and peripheral arcs of the ABR and their interactive relationship at rest and during graded muscle metaboreflex-mediated increases in BP and MSNA in humans. The major new finding is that the operational gains of both the neural and peripheral baroreflex arcs were altered by muscle metaboreflex activation. Indeed, an increase in the neural arc gain and a decrease in the peripheral arc gain were observed during PEMI following 30% MVC handgrip. Importantly, these alterations were not observed during PEMI following 15% MVC handgrip, indicating a dependence on the degree of muscle metaboreflex activation. Collectively, these findings suggest that the muscle metaboreflex interacts with and modulates ABR function in an intensity-dependent manner. Furthermore, during high intensity muscle metaboreflex activation an increased ABR operating gain of the neural arc appears to offset an attenuation of the operating gain of the peripheral arc to help maintain the overall ABR control of systemic BP.

The ABR neural arc.

In the present study, the operating point gain of the ABR neural arc was derived using the calculated transfer function gain from BP to MSNA. These data identified a clear increase in the operating gain of the neural arc during PEMI following 30% MVC handgrip. To further clarify our findings, we also employed the previously established methodology of applying a linear regression analysis to the relationship between spontaneously occurring variations in DBP and MSNA to identify the ABR neural arc operating gain (13, 15, 18, 21, 29). Importantly, we found no differences in the gain measurements obtained by the transfer function analysis or the linear regression analysis. Indeed, there was a strong relationship between the LF transfer function neural arc gain and the linear regression neural arc gain derived for total MSNA and MSNA burst incidence.

The finding of a significant increase in the ABR gain of the neural arc during PEMI after handgrip at 30% MVC is consistent with previous studies demonstrating that the ABR more effectively buffers reflex increases in MSNA during handgrip and PEMI compared with rest (13, 15, 34). Studies by Kamiya et al. (15) and Ichinose et al. (13) reported an increased gain of the ABR control of MSNA that was time dependent during the course of handgrip and was primarily due to muscle metaboreflex activation. The results from the current study are in agreement and extend our current understanding by indicating a dependence on the degree of muscle metaboreflex activation in that during PEMI following 15% MVC handgrip no increase in the ABR neural arc gain was observed. Thus it may be that increases in ABR-MSNA control are necessary to offset and restrain the profound increases in MSNA and BP evoked by high-intensity muscle metaboreflex activation.

Although the exact mechanism(s) underlying the increased ABR operating gain of the neural arc during PEMI are unclear, several points warrant discussion. Importantly, by isolating the muscle metaboreflex from the potential influence of both central command and the muscle mechanoreflex, we can suggest that the increased gain is related to muscle metaboreflex activation (24). This is in agreement with several previous studies (13, 15, 34) that have also employed isometric exercise in which increases in MSNA are primarily driven by the metaboreflex (23). In contrast, no changes in sensitivity of the baroreflex neural arc were found during mild to moderate intensity dynamic arm cycling (7, 29) and leg-kicking exercise (19, 20). However, recently, Ichinose and colleagues (12) reported a significant increase in ABR-MSNA gain during heavy and exhausting dynamic exercise, an effect that was absent during mild and moderate dynamic exercise. Collectively, these data and the current findings of no changes in the ABR neural arc after 15% MVC handgrip demonstrate an intensity-dependent modulation of the ABR regulation of MSNA during exercise where greater control (i.e., increased sensitivity) is exhibited at higher exercise intensities. Although the reason for the lack of changes in sensitivity of baroreflex neural arc during mild to moderate intensity dynamic exercise is unclear, these data suggest a minimal influence of central command and the muscle mechanoreflex in increasing the ABR neural arc operating gain. Indeed, Ichinose et al. (13) did not find an increase in the sensitivity of baroreflex neural arc during the first minute of handgrip in which central command and the muscle mechanoreflex are activated while stimulation of the muscle metaboreflex is minimal (23, 24). Collectively, these data support a role for the muscle metaboreflex in increasing the sensitivity of the ABR neural arc.

The ABR peripheral arc.

The present study is the first to examine the peripheral arc of the baroreflex during graded muscle metaboreflex activation in humans. The peripheral arc represents a relatively slow response at the efferent nerve-effector junction, representing the transduction of MSNA to an actual vasomotor response (14, 17). Indeed, in the present study, the transfer function phase of the ABR peripheral arc was significantly greater than the phase of the ABR neural arc indicating that the latency of ABR response of the peripheral arc was slower than that of neural arc. More importantly, the operating gain of ABR peripheral arc, obtained by using either FBV (Fig. 4) or FVCi (Fig. 5) as output signals, decreased from rest during PEMI following 30% MVC handgrip. In contrast, no differences in the ABR peripheral arc gain were observed for PEMI after handgrip at 15% MVC. Thus, similar to the ABR neural arc operating gain, these findings indicate that alterations in the peripheral arc are dependent on the degree of muscle metaboreflex activation. Interestingly, the reduction in the peripheral arc gain was only observed when an increase in the ABR neural arc gain was found. Although we cannot differentiate whether the changes in the peripheral arc operating gain preceded the alterations in the neural arc or vice versa, the findings clearly suggest that the neural and peripheral arcs are being differentially modulated (increased neural arc and decreased peripheral arc) likely to help maintain the overall sensitivity of systemic BP control via the ABR.

During PEMI the MSNA-FVCi relationship was reset to the right owing to the profound metaboreflex-mediated increase in MSNA, whereas no vertical shift was present due to the lack of significant changes in FVCi. The mechanism(s) contributing to the decrease in the gain of the peripheral arc during PEMI are unclear. However, the data of Lundvall and Edfeldt (22) may provide insight. These investigators demonstrated that sympathetically mediated changes in leg blood flow or leg vascular resistance reached a plateau before the sympathoexcitation stimulus had reached its maximum, primarily due to elevated sympathetic nerve activity. Indeed, the vascular response to a given change in sympathetic nerve activity was lower when the prestimulus sympathetic nerve activity was high compared with that under a condition of low sympathetic nerve activity (22). Thus it is plausible that during the large increases in MSNA induced by PEMI, the sympathetically mediated changes in leg vascular conductance were reduced compared with rest, thereby contributing to the decrease in the gain of the ABR peripheral arc. The lack of changes in the ABR peripheral arc during PEMI following 15% MVC in which MSNA was not significantly elevated from rest further supports the concept that high MSNA may alter sympathetically mediated changes in leg vascular conductance. However, the influence of local vascular regulatory mechanisms cannot be completely discounted (11). Additional studies are needed to further investigate the interaction between MSNA and vascular responsiveness during the activation of the muscle metaboreflex with PEMI.

Overall ABR function.

Kawada and colleagues (14, 16, 17) have demonstrated that the neural and peripheral arcs comprise the total ABR loop and contribute importantly to the overall control of arterial BP. In fact, these investigators have demonstrated that a decrease in the dynamic baroreflex gain of the neural arc resulted in a reduction in the total ABR gain. In the present study, we sought to better understand how increases in the ABR control of MSNA (i.e., neural arc) previously reported during PEMI influenced the transduction of ABR-mediated MSNA responses into a peripheral vascular response. Our rationale was based on previous studies demonstrating a maintained ABR gain for the control of BP during PEMI, suggesting that the increased ABR-MSNA control does not appear to manifest in an enhancement in ABR-mediated vascular responses and subsequent changes in arterial BP (6, 9, 10, 36). Indeed, we found that the ABR operating gain of the peripheral arc was decreased from rest during PEMI. These findings suggest that the increased ABR neural arc gain does not augment the ABR control of BP during PEMI but rather appears to offset an attenuation of the peripheral arc gain maintaining the overall ABR control of systemic BP.

Transfer function gain to assess ABR function.

Transfer function gains of the ABR can be identified by assessing different input-output relationships of cardiovascular variables (14, 17). However, this assessment of ABR gain is made within an integrated closed-loop system. The existence of a closed-loop feedback system makes it challenging to identify ABR function in human studies in which limited methodologies are available to open the loop. Importantly, Kawada et al. (14, 16, 17) have partitioned the ABR into a fast neural arc and a slow peripheral arc under closed-loop conditions, providing a means to more completely assess ABR function in humans. This is particularly true when considering the ABR peripheral arc since traditional pharmacological approaches do not allow for ABR vascular control to be assessed. Although limitations exist, these novel approaches can provide valuable insight into ABR function, particularly when specific efferent baroreflex measurements are made (i.e., MSNA and femoral vascular conductance). In this regard, a change in BP alters sympathetic efferent nerve activity via the ABR neural arc, whereas the resulting change in sympathetic nerve activity in turn affects arterial BP via the ABR peripheral arc (Fig. 1). However, for the estimation of the peripheral arc, a measure of the actual vascular response appears necessary. Previously, Ando et al. (1) identified ABR function of both neural and peripheral arcs in healthy subjects and patients with heart failure by using transfer function analysis. These investigators assessed the transfer function gain from MSNA to BP to estimate the peripheral arc. Because BP is dependent on both changes in cardiac output and total vascular conductance, BP is less specific than vascular conductance as an efferent-effector response of the ABR peripheral arc. Thus it is plausible that the use of BP as the output signal could underestimate or overestimate the peripheral arc gain of the ABR (1). To better understand this possibility, in the present study we analyzed the peripheral arc transfer function gain by using three different output signals: BP, FBV, and FVCi. The LF transfer function gain from MSNA to FBV and FVCi were both decreased from rest during PEMI after handgrip at 30% MVC; however, when BP was used to estimate the ABR peripheral arc gain, no changes were observed. This finding suggests that the transfer function gain from MSNA to BP may not be able to identify changes in the actual sensitivity of the ABR peripheral arc.

Several potential limitations in the design and interpretation of the present investigation should be considered. First, consideration should be given to the spontaneous methods used to assess ABR control. These noninvasive techniques only provide estimates of ABR sensitivity within the range of the normal spontaneously occurring beat-to-beat oscillations of BP. However, even though the range of pressures studied is limited by the naturally occurring fluctuations in BP, the DBP ranges obtained were fairly robust, approximating 15 mmHg. This range clearly provides significant inputs to the ABR and provides important information with regard to ABR control around normal operating pressures. Of note, although alternative methods such as the infusion of vasoactive drugs (i.e., modified Oxford technique) or the neck chamber technique may permit an assessment of baroreflex sensitivity over a greater range of DBP, these techniques also have shortcomings, particularly in assessing baroreflex control of vascular conductance. Indeed, vasoactive drugs would prevent our ability to measure the ABR peripheral arc gain. In addition, these other techniques would not have allowed us to obtain adequate ABR measures in the 3-min time period of PEMI. Second, the exact insonation angle of the Doppler probe was not able to be determined with the unit used. However, we were able to find a quality Doppler velocity signal in all subjects that was then kept constant with a custom-built holder throughout the experimental protocols. Thus, by fixing the probe to the skin and limiting movement, we are confident that the angle of insonation was held constant, which was of utmost importance for our analyses. Finally, the use of 3-min data segments for transfer function analysis may be considered limiting. However, we believe that the potential discomfort associated with longer periods of PEMI would have been problematic for the subjects.

In summary, the operating gains of both the neural and peripheral ABR arcs were altered by muscle metaboreflex activation in an intensity-dependent manner. These findings suggest that the muscle metaboreflex can interact with and modulate ABR control. Furthermore, during high-intensity muscle metaboreflex activation, an increased ABR gain of the neural arc appears to offset an attenuation of the peripheral arc gain to help maintain the overall sensitivity of the ABR control of systemic BP.

GRANTS

This research is the result of work supported with resources by National Heart, Lung, and Blood Institute Grant HL-093167 (to P. J. Fadel).