The cortical regions representing peripheral autonomic reactions in humans are poorly understood. This study examined whether changes in forebrain activity were associated with the altered physiological responses to lower body negative pressure (LBNP) following a single bout of dynamic exercise (POST-EX). We hypothesized that, compared with the nonexercised condition (NO-EX), POST-EX would elicit greater reductions in stroke volume (SV) and larger increases in heart rate (HR) and muscle sympathetic nerve activity (MSNA) during LBNP (5, 15, and 35 mmHg). Forebrain neural activity (n = 11) was measured using blood oxygen level-dependent (BOLD) functional magnetic resonance imaging. HR, SV, arterial blood pressure (ABP), and MSNA were collected separately. Compared with NO-EX, baseline ABP was reduced, whereas HR and total vascular conductance (TVC) were elevated in POST-EX (P < 0.05). In both conditions, 5 mmHg LBNP did not elicit a change (from baseline) in any physiological parameter. Compared with NO-EX, 35 mmHg LBNP-mediated decreases in SV and TVC produced greater increases in HR and MSNA during POST-EX (P < 0.05). The right posterior insula and dorsal anterior cingulate cortex demonstrated a larger decrease in BOLD at 5 mmHg LBNP but greater BOLD increase at 15 and 35 mmHg LBNP POST-EX vs. NO-EX (P < 0.005). Conversely, the thalamus and ventral medial prefrontal cortex displayed the opposite BOLD activity pattern (i.e., larger increase at 5 mmHg LBNP but greater decrease at 15 and 35 mmHg LBNP POST-EX vs. NO-EX). Our findings suggest that discrete forebrain regions may be involved with the generation of baroreflex-mediated sympathetic and cardiovascular responses elicited by moderate LBNP.
- postexercise recovery
- muscle sympathetic nerve activity
- cardiovascular regulation
- functional magnetic resonance neuroimaging
- lower body negative pressure
it has been well documented that a single bout of large-muscle dynamic exercise significantly affects the regulatory mechanisms responsible for the control of arterial blood pressure (ABP; see Ref. 12, 14, and 26). The physiological response following exercise includes increased systemic vascular conductance and arterial hypotension that can last ∼2 h in healthy normotensive individuals (17). Although local vasodilatory mechanisms contribute to these postexercise adjustments (29), there is convincing evidence that alterations in central autonomic outflow are also involved (10, 14, 26). The combination of a postexercise reduction in sympathetic outflow and augmented peripheral vasodilatation would suggest that the ability to defend against a subsequent orthostatic challenge may be compromised following exercise. This effect may be mediated by augmented venous pooling in the lower limbs followed by a greater reduction in cardiac venous return. In fact, previous research has observed a diminished tolerance to gravitational stress (3) and a greater incidence of postural hypotension (16) following prolonged dynamic exercise. Compared with the nonexercised state, the postexercise recovery period may provide an opportunity to manipulate the stimulus and reflex-mediated autonomic and cardiovascular responses to a real or simulated orthostatic challenge within the same individual.
Compared with preexercise, the baroreflex network is reset to a lower operating pressure during postexercise recovery, which is responsible for the reductions in peripheral sympathetic outflow (9, 14). The central neural pathways that contribute to this postexercise sympathoinhibition include barosensitive neurons within the nucleus tractus solitarius (6) and the rostral ventrolateral medulla (RVLM; see Ref. 19). These nuclei are also involved in mediating the physiological adjustments [i.e., elevated heart rate (HR) and sympathetic vasoconstriction] required to maintain ABP and cerebral perfusion during orthostatic challenges. Furthermore, neural activity within these brainstem regions can be modulated by higher cortical centers that regulate afferent baroreceptor input, efferent autonomic outflow, and cardiovascular function (5, 33, 41).
Human neuroimaging investigations into the functional anatomy of central cardiovascular control have identified discrete cortical structures concerned with the physiological reactions elicited both during (7, 23, 42) and following exercise (43). Of particular interest is determining the role or involvement of these forebrain regions under circumstances where variations in cardiovascular control are induced. Therefore, regions of the cortex involved with autonomic cardiovascular regulation should produce activity patterns similar to the changes observed in the peripheral physiological measures. For example, we identified forebrain regions that demonstrated activation or deactivation responses that matched sex-specific differences in the sympathetic and cardiovascular responses to lower body negative pressure (LBNP; see Ref. 22). The aim of the present investigation was to extend these between-subject observations and determine whether forebrain activity patterns during baroreceptor unloading would match within-subject postexercise differences in the sympathetic and cardiovascular responses to moderate levels of LBNP. We hypothesized that, compared with the nonexercised trial, the postexercise recovery period would elicit greater reductions in stroke volume (SV) and larger increases in HR and muscle sympathetic nerve activity (MSNA) during moderate levels of LBNP. Furthermore, based on our previous investigations (20, 22), it was hypothesized that the insular cortex, anterior cingulate cortex (ACC), ventral medial prefrontal cortex (vMPFC), and the medial dorsal nucleus of the thalamus would be involved with the altered autonomic cardiovascular responses to baroreceptor unloading.
Eleven young normotensive (resting blood pressure <140/90 mmHg) adults provided informed, written consent before participating in the present study, which was approved by The University of Western Ontario Health Sciences Ethics Review Board. The original study group consisted of 9 men and 2 women (aged 27 ± 4 yr, 171 ± 9 cm, and 72 ± 12 kg). All participants were screened by a licensed physician with a physical examination, including a detailed medical history and 12-lead electrocardiogram assessment. In addition, each subject completed a magnetic resonance imaging (MRI) readiness questionnaire to ensure safe compatibility within a high-magnetic-field environment. No subject smoked, exhibited a history of autonomic dysfunction, or was taking any cardiovascular-altering medications. None of the women were pregnant during the study, and all used oral contraceptives. To minimize potential variability in the autonomic control of cardiovascular variables because of reproductive hormone status, women were tested during the low-hormone phase of the oral contraceptive cycle.
Subjects participated in a total of four experimental test sessions. Two tests were dedicated to the collection of the functional neuroimaging data and the other two for collection of the peripheral physiological data. Each set of experiments (neuroimaging and physiological) was performed under the following two separate conditions: 1) after a minimum 20-min period of supine rest with no exercise (NO-EX) and 2) following a 60-min bout of moderate-intensity dynamic cycling exercise (POST-EX). The physiological tests were separated by a minimum of 3 wk and the neuroimaging sessions by a minimum of 2 wk. Each experiment was randomly assigned and performed at the same time of day, ≥3 h after a light meal and ≥24 h after the consumption of caffeinated products and alcoholic beverages.
During the exercise tests, subjects were instrumented with a telemetric HR monitor and remained seated on the cycle ergometer (model 874E; Monark, Kroonsväg, Sweden) until a stable resting HR was obtained. Following a 5-min period of loadless pedaling, the workload was increased every 2 min until the subject had reached the minimum HR range [i.e., 60% of maximal heart rate reserve (HRR)]. Each participant exercised for 60 min at a workload that maintained their HR between 60 and 70% of their maximal HRR using the American College of Sports Medicine heart rate reserve calculation (1). HRR was defined as the age-predicted maximal HR minus the resting seated HR obtained on the cycle ergometer before the start of exercise. Subjects were permitted to cycle at a self-chosen cadence level that was monitored and recorded. They were encouraged to cycle at this cadence throughout the 60-min exercise period and during the subsequent exercise test. During exercise, subjects were required to drink a minimum of 15 ml of water for every kilogram of body weight to help replace water loss because of sweating and to minimize the confounding effects that hypovolemia has on autonomic cardiovascular control (21). The exercise trials ended with a 5-min cooldown, and the subjects were immediately placed in the supine position and instrumented for the LBNP protocol (37).
The LBNP protocol used for all test sessions consisted of three randomly assigned levels of decompression at 5, 15, and 35 mmHg. The 5 mmHg level of LBNP was designated as the control task and was used to simulate the subjective sensory and emotional experience associated with the neuroimaging sessions (i.e., auditory sensations) without producing a significant cardiovascular challenge. Subsequently, both the 15 and 35 mmHg LBNP data were compared against 5 mmHg LBNP in an attempt to minimize any potential confounding forebrain neural activity patterns related to somatosensory input and therefore enhance our ability to detect cortical activity changes associated specifically with the autonomic and cardiovascular reactions to baroreceptor unloading. Subjects were positioned within the LBNP chamber for a minimum of 20 min before the collection of baseline measures. At all levels of LBNP, a sequence of four repeated 45-s LBNP exposures was administered, with each separated by a 30-s intervening rest period. For the neuroimaging sessions, a custom-designed LBNP chamber combined with a medical anti-shock trouser (David Clark, Worcester, MA) system was used to minimize subject head movement during scanning (20). The average time between the end of the exercise session and the start of the LBNP protocol was ∼30 min for both the physiological and neuroimaging sessions. Furthermore, all LBNP experiments were completed within 90 min following the end of exercise.
Peripheral physiological measurements and analysis.
During LBNP exposure, HR was determined by standard three-lead electrocardiogram methods. Cardiac stroke volume velocity (SVV) and brachial artery blood velocity (BBV) were obtained using pulsed Doppler ultrasound along with aortic root and brachial artery dimensions for the calculation of cardiac output (CO, n = 10) and forearm blood flow (FBF, n = 8), respectively. Aortic (2.5 MHz probe) and brachial artery (10 MHz probe) diameters were determined using two-dimensional B-mode echo Doppler images (GE/Vingmed System Five). Beat-by-beat measures of ABP were obtained using finger photoplethysmographic techniques (Finometer; Finapres Medical Systems) with the hand held at heart level. These blood pressure measures were corrected against sphygmomanometrically collected systolic and diastolic pressures intermittently throughout the experiment.
Multiunit recordings of postganglionic MSNA were successfully collected for both physiological test sessions in five subjects. Tungsten microelectrodes were inserted percutaneously into muscle fascicles of the right common fibular (peroneal) nerve with a reference electrode positioned subcutaneously 1–3 cm from the recording site (11). Neural activity was amplified 1,000 times by a preamplifier and an additional 75 times through a variable-gain isolated amplifier. The signal was bandpass filtered (0.7–2.0 kHz), full wave rectified, and integrated with a resistance-capacitance circuit (0.1-s time constant). Criteria for an acceptable MSNA recording included pulse-synchrony with the cardiac cycle and increased activity to a voluntary apnea but not to emotional arousal (i.e., a loud noise).
Baseline data were averaged over the 2- to 5-min period before the first applied level of LBNP. Peak HR changes within each repetition of lower body suction were determined, and the mean of these responses was used for statistical analysis. LBNP-mediated changes in SV, CO, FBF, and MSNA were determined over the 45-s period of suction, and the average over all repetitions was calculated. Only MSNA bursts with characteristic rising and falling slopes and amplitudes with a 2:1 or greater signal-to-noise ratio were measured for frequency per minute.
Neuroimaging measurements and analysis.
The neuroimaging experiments were performed on a Varian/UNITY INOVA (Palo Alto, CA) 4-Tesla whole body imaging system with a Siemens Sonata Gradient chain (Siemens, Erlangen, Germany). Subjects lay supine on the scanning bed in the LBNP chamber, and foam pads were placed on either side of the head to minimize movement during scanning. During the scanning session, HR was calculated from the pulse intervals recorded on a MRI-compatible oximeter (Nonin Medical Inc, 8600FO MRI, Plymouth, MN, USA) placed over the middle finger of the left hand.
Before functional imaging, a global shimming procedure (RASTAMAP) using first- and second-order shims was performed to optimize the magnetic field over the imaging volume of interest (24). Twenty-one interleaved contiguous axial slices (5 mm thick, 3.4 × 3.4 mm in-plane voxel resolution) were acquired in each volume. Volume acquisition time was 2.5 s with a time to repetition of 1.25 s (2 shots). A total of 133 volumes was collected per session. Five steady-state volumes were acquired before actual data collection to allow for magnetization equilibrium and discarded before data analysis. Functional data were collected using a segmented T2 (transverse relaxation time without refocusing)-weighted gradient spiral pulse sequence [TE (time to excitation) = 15 ms, flip angle = 60°, field of view = 220 × 220 mm] with navigator echo correction. A corresponding high-resolution gray/white matter contrast [i.e., T1 (longitudinal relaxation time)-weighted] anatomical image was acquired at the end of the same scanning session with a voxel resolution of 0.9 × 0.9 × 1.3 mm.
All neuroimaging data were processed using statistical parametric mapping software (SPM2; Wellcome Department of Imaging Neuroscience). Complete details of the functional imaging analysis have been reported previously (20). Briefly, all functional image volumes were realigned to the first volume collected. A mean functional image was created using the realigned volumes and coregistered to the anatomical image. For optimal coregistration, these images were bias corrected, skull and scalp stripped (36). The functional images were spatially normalized into a fixed stereotactic space using the Montreal Neurological Institute-152 template. Following this process, two of the subjects’ image volumes were subsequently removed from further analysis because of excessive head movement (>1.0 mm). To minimize intersubject differences in functional and cortical anatomy, the remaining nine image volumes were spatially normalized with a Gaussian kernel set at 8 mm full width at half-maximum. Furthermore, all functional images were high-pass filtered, corrected for serial correlations [autoregressive model, AR(1)], and normalized for global activations before statistical inference.
All data are expressed as means ± SD. Prior investigations have documented the physiological adjustments associated with the postexercise state (12, 14, 17). Therefore, the effect of exercise on baseline and 5 mmHg LBNP physiological data was analyzed using one-tailed paired t-tests. LBNP-mediated changes in all dependent variables were analyzed using a repeated-measures two-way (condition × LBNP) ANOVA. Tukey's post hoc analysis was performed to estimate differences among means. Probability levels during multiple pointwise comparisons were corrected using Bonferonni's approach. Statistical analyses for all cardiovascular and MSNA data were performed using a computer-based software program (SAS). A critical significance level of P < 0.05 was set for all cardiovascular and sympathetic comparisons.
A two-level statistical paradigm was used for all functional neuroimaging data. First, a within-subject analysis was performed to identify differences in signal intensity between baseline and LBNP periods. The change in blood oxygen level-dependent (BOLD) signal over repeated LBNP exposures was modeled using a canonical hemodynamic response function. This resulted in subject-specific contrast images containing whole brain information related to both sites of increased and decreased BOLD signal during LBNP. At the second level of analysis, these contrast images were included in a between-subjects mixed-effects ANOVA to identify significant (P < 0.005, uncorrected for multiple comparisons) changes in signal intensity that corresponded with condition-specific physiological responses measured during LBNP (i.e., condition × LBNP interaction effects). To reduce the risk of reporting false-positive results, a minimum cluster threshold of 10 voxels was included. Significant voxels were color coded for T-score and overlaid on a spatially normalized anatomical template. The effect size, representing the mean percent change in BOLD signal, was calculated for each significant cluster. All functional MRI data are represented in a neurological convention (i.e., subject's left appears on the left).
Based on our recent observations (20) and previous data highlighting the role of discrete cortical structures associated with central cardiovascular control (5, 7, 8, 15, 23, 27), we focused on signal intensity changes within specific cortical regions. These a priori anatomical regions of interest (ROI) included the insular cortex, ACC, vMPFC, and the thalamus. All ROI normalized masks were generated using the WFU_PickAtlas software program (version 1.04; see Ref. 28).
The exercise-related goal for the present study was to have all participants cycle for 60 min at 60–70% of their maximal HRR. The average workloads were 137 ± 22 and 134 ± 19 watts for the physiological and neuroimaging exercise sessions, respectively. HR increased from a seated resting value of 70 ± 6 to 150 ± 12 beats/min (66 ± 3% HRR) before the physiological recording session and from 69 ± 6 to 148 ± 12 beats/min (64 ± 6% HRR) for the neuroimaging session. The mean ratings of perceived exertion following exercise were 15 ± 6 and 14 ± 3 for the physiological and neuroimaging experiments, respectively. There were no differences in workload (P = 0.25), percent HRR (P = 0.49), or rating of perceived exertion (P = 0.11) between the two exercise sessions.
Baseline and 5 mmHg LBNP physiological responses following exercise.
Baseline (pre-LBNP supine rest) and 5 mmHg LBNP systemic and forearm hemodynamics and MSNA data are displayed in Table 1 for the NO-EX and POST-EX conditions. Compared with NO-EX, baseline HR, CO, and FBF were elevated, whereas systolic and MAP were reduced in POST-EX (Table 1; all P < 0.05). Similarly, TVC was elevated at baseline (22 ± 12%) and at 5 mmHg LBNP (21 ± 15%) in the POST-EX compared with the NO-EX condition. These systemic changes were associated with a postexercise rise in FVC at baseline (28 ± 14%) and at 5 mmHg LBNP (33 ± 20%). The application of 5 mmHg LBNP did not elicit a measurable change (from baseline) in any of the parameters during either condition (all P > 0.15). Sympathetic nerve burst frequency data at baseline (P = 0.11) and 5 mmHg LBNP (P = 0.13) were not different between the NO-EX and POST-EX conditions.
Physiological responses to LBNP.
Compared with 5 mmHg LBNP, reductions in SV (Fig. 1B), CO, TVC (Fig. 1C), FBF, and FVC (Fig. 2A) and increases in MSNA burst frequency (Fig. 2B) occurred in both conditions during 15 and 35 mmHg LBNP (P < 0.05). For the above comparisons, changes during 35 mmHg LBNP were greater than 15 mmHg LBNP (all P < 0.05). Steady-state MAP was not different (P > 0.19) from 5 mmHg LBNP in either condition during 15 mmHg LBNP (90 ± 9 vs. 89 ± 8 mmHg for NO-EX and 86 ± 9 vs. 84 ± 7 mmHg for POST-EX 5, and 15 mmHg LBNP) or during 35 mmHg LBNP in the NO-EX condition (90 ± 8 mmHg). However, there was a tendency for MAP to decrease during 35 mmHg LBNP following exercise (from 86 ± 9 at 5 mmHg LBNP to 82 ± 6 mmHg at 35 mmHg LBNP, P = 0.07). Importantly, condition × LBNP interactions were observed for LBNP-mediated changes in HR, SV, CO, TVC, FVC, and MSNA (all P < 0.05). Specifically, larger reductions in SV, CO, TVC, and FVC and greater elevations in HR and MSNA burst frequency occurred during 35 mmHg LBNP following exercise compared with NO-EX (Figs. 1 and 2).
Forebrain BOLD signal responses to LBNP: baseline vs. 5 mmHg LBNP.
Compared with NO-EX, greater reductions in BOLD signal (P < 0.005) were observed during 5 mmHg LBNP following exercise in the right posterior superior insula (−0.01 ± 0.09% NO-EX vs. −0.35 ± 0.13% POST-EX) and the dorsal anterior cingulate cortex (dACC; 0.07 ± 0.12% NO-EX vs. −0.84 ± 0.32% POST-EX). Conversely, greater increases (P < 0.005) in BOLD signal during 5 mmHg LBNP postexercise were identified in the medial dorsal nuclei of the thalamus (−0.04 ± 0.10% NO-EX vs. 0.41 ± 0.27% POST-EX) and vMPFC (0.23 ± 0.11% NO-EX vs. 0.66 ± 0.31% POST-EX).
Forebrain BOLD signal responses to LBNP: 15 and 35 vs. 5 mmHg LBNP.
To expose cortical structures involved with the sympathetic and cardiovascular responses to mild and moderate baroreceptor unloading, BOLD signal changes (from 5 mmHg LBNP) to 15 and 35 mmHg LBNP related to the condition × LBNP interactions for SV (Fig. 1B), TVC (Fig. 1C), FVC (Fig. 2A), and MSNA (Fig. 2B) were identified. The specific forebrain locations that demonstrated a similar pattern of BOLD signal change are presented in Table 2 and Fig. 3, A–D. There were no specific BOLD response patterns that matched the condition × LBNP interaction for HR (Fig. 1A).
In both conditions, the right posterior insular cortex and the dACC demonstrated graded elevations in BOLD signal during 15 and 35 mmHg LBNP (Fig. 3, A and B). The increases in BOLD signal were greater during 35 mmHg than 15 mmHg LBNP (P < 0.005). Compared with NO-EX, larger increases (P < 0.005) in BOLD signal were observed within these two forebrain regions during mild and moderate LBNP following exercise.
Conversely, progressive reductions in BOLD signal were observed during 15 and 35 mmHg LBNP in the medial dorsal nucleus of the thalamus and the vMPFC (Fig. 3, C and D). The magnitude of BOLD signal reduction was greater (P < 0.005) during 35 mmHg LBNP than 15 mmHg LBNP in both conditions. However, the decreases in BOLD signal during 15 and 35 mmHg LBNP within these cortical structures were greater (P < 0.005) POST-EX vs. the nonexercised condition.
In this study, the application of 35 mmHg LBNP following exercise elicited greater decreases in SV, TVC, and FVC and greater increases in HR and MSNA vs. the nonexercised condition. The main new findings of the present investigation were that these postexercise physiological differences during 35 mmHg LBNP corresponded with larger increases in BOLD signal within the right posterior insula and dACC, along with greater decreases in BOLD signal in the medial dorsal nucleus of the thalamus and vMPFC. The observation that significant BOLD signal responses were observed during 5 mmHg LBNP in the postexercise but not the nonexercised condition implies that alterations in baseline activity within these cortical regions may have contributed to the baseline hemodynamic adjustments observed following exercise (i.e., decreased MAP, increased TVC and FVC). However, the purpose of the present investigation was not to identify the cortical structures involved with the development of postexercise hypotension at baseline, and further study is required to address this issue more definitively. Nonetheless, a parallel postexercise difference in the LBNP-mediated peripheral physiological responses and BOLD signal changes within these discrete forebrain structures suggests that they are involved in the monitoring and/or regulation of baroreceptor afferent information.
Baseline physiological differences following exercise.
The exercise paradigm used in the present study was sufficient to produce a postexercise decrease in supine resting ABP consistent with prior findings in normotensive individuals (14, 43). The magnitude of hypotension and the increases in forearm and TVC were similar between baseline and 5 mmHg LBNP. This observation was consistent for all hemodynamic and sympathetic nerve data collected (Table 1). The absence of a measurable cardiovascular stress during 5 mmHg LBNP, in either condition, lends support for the use of this low level of suction as the control task from which 15 and 35 mmHg LBNP-mediated physiological and BOLD signal responses can be compared.
Both central neural (10, 14) and local vasodilatory (13, 14) mechanisms have been proposed to explain the postexercise hypotensive response and the rise in vascular conductance following exercise. Previous assessments of sympathetic outflow under similar conditions have revealed marked reductions in MSNA associated with a resetting of the arterial baroreflex toward lower blood pressures (14, 26). With a limited number of complete data sets, we were unable to replicate a parallel decrease in resting sympathetic outflow following exercise. Second, the accumulation of vasodilatory substances (i.e., histamine) has been shown to contribute significantly to the hypotension and increased TVC observed after physical exertion (29). Although this effect would be expected to be predominant within the vascular bed of the exercised muscles (i.e., the legs), it cannot be discounted that a translocation of these substances had occurred in nonexercised regions (i.e., the forearm; see Ref. 29). Nonetheless, the baseline elevation in FVC observed following exercise may provide indirect evidence for a general reduction in sympathetic tone (Table 1), as previously observed by others (10, 14, 26).
LBNP-mediated physiological differences following exercise.
It has been reported that mild levels of LBNP (<20 mmHg) influence both arterial and cardiopulmonary baroreceptor afferent information (38) although reflex changes in HR and MSNA are differentially affected. The application of 15 mmHg LBNP increased MSNA (Fig. 2B) and reduced SV, TVC (Fig. 1, B and C), and FVC (Fig. 2) similarly in both conditions. Neither HR nor ABP was affected by this level of LBNP in either condition. These observations suggest that baroreflex gain was unchanged at this level of baroreceptor unloading but continued to operate around a lower arterial pressure.
The decrease in SV during moderate (i.e., 35 mmHg) LBNP was associated with tachycardia, a greater reduction in TVC, and an augmented sympathetic nerve response. These adjustments are consistent with combined unloading of both cardiopulmonary and arterial baroreceptors (18). However, the magnitudes of these responses were augmented after exercise. The larger POST-EX drop in SV during this level of LBNP may have resulted from more venous pooling within the exercised limbs. The greater increases in HR and MSNA during 35 mmHg LBNP postexercise appeared to be the consequence of a larger reduction in SV (and presumably baroreceptor afferent traffic), since the relationships between increases in MSNA and HR (response variables) vs. the reductions in SV (stimulus variable) were not different between the two conditions (Table 3). Despite the greater POST-EX decrease in TVC during 35 mmHg LBNP, there was also a modest fall in ABP (−3.2 ± 1 mmHg, P = 0.07) concurrent with a larger reduction in CO. This drop in ABP was not observed during this level of suction in the nonexercised condition.
Forebrain BOLD signal responses following exercise.
Medullary mechanisms contributing to the depressed cardiovascular state following exercise include an increased excitatory drive to the nucleus of the solitary tract (NTS; see Ref. 6) and/or augmented inhibitory input to the RVLM (19). The precise source of these projections is unknown, but a recent neuroimaging study has implicated the involvement of higher brain structures, including the thalamus, anterior cingulate, and insular cortexes (42), which was confirmed in the present investigation.
The right insular cortex and dACC have previously been implicated with autonomic control of cardiovascular function (4, 30). The LBNP-mediated increase in BOLD signal within these regions suggests that activation of these structures is associated specifically with sympathoexcitation (Fig. 3, A and B). In support of these findings, electrical cortical stimulation and human neuroimaging studies have demonstrated that activation of the right insula and dACC mediate elevations in both sympathetic nerve activity and HR (4, 8, 22, 30). The specific central pathways responsible for mediating these sympathoexcitatory responses remain elusive, although direct neural connections have been discovered between these two forebrain centers and the NTS (33, 39), but not to the RVLM (34). The current data suggest that, at rest, activity within the right posterior insular cortex and dACC is suppressed by baroreceptor afferent information, since graded unloading of the baroreceptors (during 15 and 35 mmHg) produced progressively larger increases in BOLD signal. Furthermore, the larger POST-EX reduction in SV during 35 mmHg (vs. NO-EX) resulted in greater BOLD signal within these forebrain regions. The augmented postexercise rise in BOLD signal in the right posterior insula and dACC during 35 mmHg LBNP was matched by greater tachycardia and peripheral sympathetic nerve activity. Therefore, efferent connections from the right posterior insula and dACC may inhibit cardiovagal neurons and/or mediate sympathoexcitatory responses via the RVLM. Future research is required to determine the specific neural connections associated with these baroreflex-mediated responses.
The BOLD signal patterns within the medial dorsal thalamus and vMPFC were opposite to those of the right posterior insula and dACC and may therefore be involved with the facilitation of cardiovagal or sympathoinhibitory processes. The decreases in BOLD signal within these locations would be consistent with that experienced by the baroreceptor afferents (i.e., attenuated afferent traffic during LBNP) and suggest that these cortical regions may be tonically activated by primary baroreceptor input. The greater decrease in BOLD signal coupled with the augmented HR and MSNA responses during 35 mmHg LBNP postexercise suggest that an inhibitory drive may exist from these regions to sympathetic premotor neurons in the RVLM. With respect to the vMPFC, this idea is supported by observations that depressor responses evoked by stimulation of this region are accompanied by reductions in the discharge of sympathoexcitatory barosensitive neurons within the RVLM (40). In addition, activation of the vMPFC has also been linked with baroreflex-mediated parasympathetic cardiac control (32). Therefore, the greater POST-EX decrease in BOLD signal of the vMPFC during moderate LBNP may have contributed to the larger HR response observed in this condition via a further withdrawal of cardiovagal outflow. Although similar associations have not been observed between these autonomic nuclei and the medial dorsal subdivision of the thalamus, there is evidence of a direct projection between the thalamus and the NTS (2). Furthermore, reciprocal excitatatory connections do exist between the medial prefrontal cortex and the thalamus (25, 31), and baroreceptive neurons within the thalamus project to the posterior insula (43). Therefore, the medial dorsal thalamic nuclei may influence autonomic outflow from the RVLM or vagal preganglionic neurons destined for the myocardium and/or peripheral vasculature through synaptic relays within these cortical structures.
In conclusion, an acute bout of moderate-intensity dynamic exercise produced a postexercise baseline arterial hypotension associated with decreased systemic and forearm vascular tone. Compared with the nonexercised condition, the postexercise control task (5 mmHg LBNP) was associated with decreased BOLD signal in the right posterior insula and dACC and increased BOLD signal in the medial dorsal thalamic nuclei and vMPFC. The application of 35 mmHg LBNP following exercise elicited a larger drop in SV and a greater decrease in total and FVC with augmented HR and MSNA responses. These enhanced baroreflex-mediated efferent sympathetic and cardiovascular reactions corresponded with greater BOLD signal increase in the right posterior insula and dACC combined with a larger decrease in BOLD signal in the medial dorsal thalamus and vMPFC. Taken together, the data suggest that these cortical structures may influence the sympathetic and cardiovascular reactions associated with acute reductions in central blood volume.
This investigation was supported by the Ontario March of Dimes, the Canadian Space Agency, and the Heart and Stroke Foundations of Ontario and Canada (Grant nos. NA5020 and T5342). D. S. Kimmerly was a recipient of a Heart and Stroke Foundation of Canada Doctoral Research Award.
We thank the subjects for cheerful cooperation and acknowledge the expert technical assistance of Jason Bakker, Amanda Rothwell, Joe Gati, and Joy Williams.
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
- Copyright © 2007 by the American Physiological Society