despite early reports that “global” cerebral blood flow (CBF) remains unchanged during exercise, more recent studies have provided strong evidence to suggest that exercise causes regional increases in CBF (19, 24, 43, 45). Such changes in CBF have been associated with increases in cerebral activity and metabolism; however, hemodynamic, neurogenic, and neurohumoral factors have all been shown to modulate the CBF response to exercise (24, 43, 46). As such, the control of cerebral perfusion during exercise appears to be highly complex and remains incompletely understood.

It is well established that the cardiovascular responses to exercise are mediated by descending signals arising from higher brain centers (i.e., central command) (20, 28, 32) and by peripheral feedback arising from the exercising skeletal muscle, which is composed of mechanically (i.e., mechanoreflex) and metabolically (i.e., metaboreflex) sensitive afferents (5, 10, 32). Importantly, the activation of these neural pathways is also believed to contribute to the CBF response to exercise (13, 14, 25, 35, 49, 52, 53). Indeed, handgrip exercise-induced increases in regional cerebral perfusion are abolished by removal of muscle afferent feedback with regional anesthesia, thus implying that central command is less important for the CBF response to exercise (13, 14, 25). More recent work has demonstrated that exercise-induced elevations in middle cerebral artery velocity do not persist when metabolically sensitive muscle afferents are activated in isolation during postexercise muscle ischemia (PEMI) (26, 42). Taken together, these findings have led to the contention that CBF responses to exercise are dependent on the activation of mechanically sensitive muscle afferents but are independent of central command and the muscle metaboreflex (26, 42). However, whether muscle mechanoreflex activation accounts for the exercise-induced increases cerebral perfusion remains to be rigorously tested in humans.

In contrast, recent neuroimaging studies have reported exercise-induced increases in regional CBF that are specifically associated with the activation of central command (35, 49, 52, 53). Indeed, the insular cortex and anterior cingulate regions are activated during handgrip exercise but not during PEMI (53). Although such regional increases in CBF were convincingly shown to be independent of the muscle metaboreflex, the influence of mechanically sensitive afferents cannot be discounted, as the mechanoreflex would be activated during exercise along with central command. An alternative experimental approach to determine the relative importance of central command to the exercise-induced increases in cerebral perfusion is to compare the responses to voluntary exercise with those elicited by electrically stimulated exercise, where the parallel activation of central motor and cardiovascular centers is absent but peripheral feedback remains intact. Although several studies have considered the effects of experimental isolation or augmentation of central command on the cerebral perfusion responses to exercise in humans (36, 49, 52, 53), to the authors' knowledge how the absence of central command during exercise influences these responses has not been determined.

Given this background, the purpose of the present study was twofold. First, we tested the hypothesis that selective activation of muscle mechanoreceptors by passive calf stretch would increase anterior cerebral artery (ACA) mean blood velocity (V_mean). We chose to assess ACA V_mean because the ACA supplies the brain region associated with cortical representation of the calf muscles (3) and ACA V_mean is specifically increased during calf exercise (29, 33). Second, we examined whether the absence of central command would attenuate calf exercise-induced ACA V_mean responses. To achieve this, ACA V_mean responses to voluntary calf exercise (i.e., central command and mechanoreflex and metaboreflex activation) were compared with those elicited during electrically stimulated calf exercise performed at the same exercise intensity (i.e., mechanoreflex and metaboreflex activation but no central command). A period of PEMI was performed after all exercise bouts to assess the potential muscle metaboreflex contribution to ACA V_mean under each condition.

METHODS

Sixteen healthy volunteers (12 men and 4 women) with a mean age of 24 ± 1 yr, weight of 76 ± 4 kg, and height of 175 ± 2 cm (means ± SE) participated in the present study. All experimental procedures and protocols conformed to the Declaration of Helsinki and were approved by the institutional Ethics Committee. Before participation, each subject gave written informed consent and completed a medical health history questionnaire. No subject had a history or symptoms of cardiovascular, respiratory, metabolic, or neurological disease, and none were using prescribed or over-the-counter medications. Female participants were studied at the early follicular phase of the menstrual cycle, in which plasma estrogen and progesterone concentrations are generally low (22), although hormonal status was not directly assessed. Subjects were familiarized with all the equipment and procedures in two separate visits before any experimental sessions. Two study days were then conducted, on which passive stretch and exercise protocols were separately performed (described below). Subjects were requested to abstain from caffeinated beverages for 12 h and strenuous physical activity and alcohol for at least 24 h before experimental sessions. On the experimental days, subjects arrived at the laboratory a minimum of 2 h after a light meal. All experiments were performed at an ambient room temperature of 22–24°C with external stimuli minimized.

Experimental Measurements

Heart rate (HR) was continuously monitored using a three-lead electrocardiogram (Diascope DS 521, S&W Medicoteknik, Albertslund, Denmark) in the lead II position. Beat-to-beat blood pressure (BP) was measured using finger photoplethysmography (Portapres model 2, TNO Biomedical Instrumentation, Amsterdam, The Netherlands) obtained from the middle finger of the right hand, which was supported at the level of the right atrium on an adjustable padded support. In addition, an automated sphygmomanometer (Omron, Matsusaka, Japan) was used to measure brachial artery BP in the left arm every minute to verify the Portapres measurements of absolute BP (8). Stroke volume (SV) was calculated offline from the BP waveform using the Modelflow software program incorporating BeatScope version 1.0 software (TNOTPD, Biomedical Instrumentation, Amsterdam, The Netherlands) (51). This methodology has been shown to reliably estimate rapid changes in cardiac output (CO) during a variety of experimental protocols (34, 47), including isometric exercise (48). CO was calculated from the product of SV and HR. End-tidal Pco₂ (Pet_co2) was obtained from a capnogram acquired by means of a nasal cannula connected to a rapid response infrared CO₂ analyzer (Servomex 1440, Crowborough, East Sussex, UK). Changes in regional cerebral perfusion were assessed by the measurement of blood flow velocity in the left ACA. The proximal segment of the ACA was insonated through the temporal “window” using a 2-MHz pulsed-wave transcranial Doppler ultrasound (Multidop X, DWL, Sipplingen, Germany) with online spectrum analysis. The ACA was found at a depth of 60–75 mm, and velocity was shown as downwardly deflected pulse waves. After the optimum signal had been achieved, the probe was fixed with an adjustable headband and adhesive ultrasonic gel (Tensive, Parker Laboratories, Orange, NJ). ACA V_mean was expressed in centimeters per second. The cerebral vascular conductance index (CVCi) was calculated as ACA V_mean divided by mean arterial pressure (MAP). Signal outputs were transmitted to an analog-to-digital converter (1401plus, Cambridge Electronic Design, Cambridge, UK), sampled at frequencies of 1,000 Hz, and stored on a personal computer using Spike 2 software (Cambridge Electronic Design). Customized Spike 2 scripts files were used offline to determine beat-to-beat values for systolic BP (SBP), diastolic BP (DBP), MAP, and ACA V_mean. One-minute averages were calculated during the last minute of the rest period, passive calf stretch, calf exercise, and PEMI experimental phases, whereas the recovery phase was taken 30 s after the end of stretch or PEMI.

Experimental Procedures

Passive calf stretch.

Subjects were seated in a semirecumbent position in a Biodex System 3 Pro (Biodex Medical Systems, Shirley, NY) with the right knee flexed by 30° and the foot strapped to the footplate so the lower leg was horizontal to the floor (n = 10). Velcro straps were used to fix the foot and minimize heel lift during passive stretch (7). The maximum voluntary contraction (MVC) of the right calf plantar flexors was assessed by taking the highest torque produced in three attempts. Calf muscle stretch was assessed by manually dorsiflexing the foot to the end of the comfortable range of motion. After familiarization with all protocols, subjects performed separate bouts of sustained and rhythmic passive calf muscle stretch. After subjects were settled for 10 min, the protocol began with a 3-min rest period, and then the foot was passively dorsiflexed by the Biodex to the preset angle at a velocity of 30°/s, where it was held in position for the next 2 min. After this stretch period, the foot was returned to its starting position for 3 min of recovery. After a subsequent 15-min recovery period, the protocol was repeated but with rhythmic calf stretch, where the foot was passively dorsiflexed by the Biodex at a frequency of 0.5 Hz and an angular velocity of 60°/s. Using surface electromyogram recordings, it has previously been confirmed that the gastrocnemius and soleus muscles remain quiescent during these methods of calf stretch (7, 17).

Calf exercise.

Experimental sessions consisted of two exercise paradigms (i.e., isometric and rhythmic exercise), which were performed under both voluntary and electrically stimulated conditions (i.e., isometric voluntary exercise, isometric electrically stimulated exercise, rhythmic voluntary exercise, and rhythmic electrically stimulated exercise). The isometric and rhythmic trials were performed in a counterbalanced order, and within these conditions, the order of the voluntary or stimulated contractions was also counterbalanced. Exercise trials were separated by at least a 20-min rest period.

ISOMETRIC EXERCISE.

Subjects were seated in the dynamometer with the right thigh clamped horizontally with the ankle positioned at 85° (n = 8) (11). The upward force generated by the triceps surae was transduced, amplified, and transmitted to an analog-to-digital converter. Force output was displayed on a computer screen and recorded on Spike 2 computer software for offline analysis. While subjects were seated in the dynamometer, the subject's MVC was assessed and 35% of maximum was displayed on a computer screen positioned directly in front of them. For voluntary isometric contractions, subjects maintained the required force by matching the deflection produced on the computer screen to the predetermined 35% MVC line. Electrically stimulated isometric contractions were evoked using tetanic stimulation at 20 Hz with a pulse width of 50 μs was delivered percutaneous using established methodology (4, 11) (Digitimer DS7, Digitimer, Hertfordshire, UK). A force output of 35% MVC (i.e., the same relative and absolute force to that produced during voluntary exercise) was maintained by small adjustments of the stimulus current.

Once subjects were seated in the leg dynamometer and after the subject had attained a stable circulatory state, continuous recording of resting HR, BP, Pet_co2, and ACA V_mean were made for 5 min. Two minutes of voluntary or electrically stimulated isometric calf exercise was then performed followed by 2 min of PEMI. PEMI was achieved by inflation of a thigh cuff around the right thigh to suprasystolic pressure (220 mmHg) 5 s before the end of exercise using a rapid inflation unit (E20, Hokanson, Bellevue, WA). After the thigh cuff was deflated, a 5-min recovery period was conducted.

RHYTHMIC EXERCISE.

The rhythmic exercise protocol was conducted in the same manner as that described above for the isometric protocol; however, intermittent rhythmic contractions at 35% MVC were performed at a frequency of 1 Hz (n = 12, 0.5 s of contraction followed by 0.5 s of relaxation). For voluntary contractions, the contraction pace was dictated by a metronome, whereas for electrically stimulated exercise, a 20-Hz tetanic stimulus with a pulse width of 300 μs was delivered percutaneously once a second (12). Small adjustments of the stimulus current were made to maintain electrically evoked rhythmic contractions at 35% MVC (i.e., the same relative and absolute force to that produced during voluntary rhythmic exercise). All subjects were carefully familiarized with the stimulation procedures before actual data collection, and the level of electrical stimulation was well tolerated.

Statistical Analysis

Data are reported as means ± SE. In each experimental protocol, statistical comparisons of measured physiological variables were made using repeated-measures two-way ANOVA with the Greenhouse-Geisser correction, in which condition (i.e., voluntary vs. electrically stimulated exercise or sustained vs. rhythmic stretch) and phase (time effect) were the main factors (30). A Student-Newman-Keuls test was used post hoc to investigate significant main effects and interactions. Comparisons of changes from rest in the measured physiological variables, induced by isometric and rhythmic calf muscle exercise and passive calf muscle stretch, were made using paired Student's t-tests. Statistical significance was set at P < 0.05. All analyses were conducted using SigmaStat for Windows (Jandel Scientific Software, SPSS, Chicago, IL).

RESULTS

Passive Stretch

Table 1 shows the cardiovascular, cerebrovascular, and respiratory variables measured before, during, and after both sustained and rhythmic passive calf muscle stretch. At rest, there were no significant differences in BP, HR, CO, ACA V_mean, CVCi, or Pet_co2 between conditions (P > 0.05; Table 1). Sustained passive calf muscle stretch produced small but significant increases in MAP and DBP (P < 0.05 vs. rest; Table 1), which were more marked than those elicited during rhythmic passive calf muscle stretch (P < 0.05 between conditions; Fig. 1). In contrast, no significant changes were observed in ACA V_mean, CVCi, SBP, HR, or CO during passive calf stretch (P > 0.05 vs. rest; Table 1 and Fig. 1). Pet_co2 remained constant throughout all experimental phases, although it was slightly but significantly lower in the rhythmic calf stretch condition compared with the sustained calf stretch condition (P < 0.05 between conditions; Table 1). The torque elicited during sustained and rhythmic passive calf muscle stretch was equivalent to 31 ± 5 and 39 ± 5% MVC, respectively.

Isometric Exercise

At rest, there were no significant differences in any of the measured variables between the voluntary and electrically stimulated isometric exercise conditions (P > 0.05; Table 2). Voluntary and stimulated isometric exercise induced similar increases in BP (SBP, DBP, and MAP), HR, and CO from rest (P < 0.05; Table 2 and Fig. 1). ACA V_mean only increased during voluntary exercise (+15 ± 1%, P < 0.05 vs. rest) and remained unchanged during stimulated exercise. CVCi was not altered from rest during voluntary isometric exercise but was significantly reduced by stimulated isometric exercise (P < 0.05 vs. rest). During PEMI, BP (SBP, DBP, and MAP) remained elevated in both voluntary and stimulated isometric exercise conditions, although BP (DBP and MAP) during PEMI was higher after stimulated isometric exercise (P < 0.05; Table 2 and Fig. 1). In both conditions, HR and CO returned to baseline during PEMI, whereas in the voluntary isometric condition ACA V_mean recovered toward baseline so that no significant differences were observed from rest or between conditions (P > 0.05). CVCi was only significantly reduced from rest during PEMI after stimulated isometric exercise (P < 0.05). Pet_co2 remained unchanged throughout all experimental phases with no differences between voluntary or stimulated exercise conditions (P > 0.05; Table 2).

Rhythmic Exercise

Overall, the cardiovascular and cerebrovascular responses in the rhythmic exercise protocol were similar to those obtained during the isometric exercise protocol. No significant differences were observed in any of the measured variables at rest before voluntary and stimulated rhythmic exercise (P > 0.05; Table 3). During rhythmic exercise, the HR, CO, and pressor responses (SBP, DBP, and MAP) were similarly increased from rest in both the voluntary and electrically stimulated rhythmic exercises (P < 0.05 vs. rest; Table 3 and Fig. 1). In contrast, ACA V_mean was increased during voluntary rhythmic exercise (+15 ± 2%, P < 0.05 vs. rest) but was unchanged from rest during stimulated rhythmic exercise (Table 3 and Fig. 1). CVCi was unchanged from rest during rhythmic voluntary exercise but was significantly reduced during rhythmic stimulated exercise (P < 0.05 between conditions; Fig. 1). During PEMI, HR and CO returned to baseline levels, whereas BP (SBP, DBP, and MAP) remained elevated, although the pressor response (DBP and MAP) during PEMI was greater in the stimulated condition (P < 0.05; Table 3). ACA V_mean returned to baseline after voluntary rhythmic exercise during PEMI, whereas it remained unchanged in the rhythmic stimulated exercise condition, such that there were no difference between conditions (P > 0.05; Table 3). CVCi was significantly reduced from rest during PEMI in stimulated exercise condition. Pet_co2 remained constant throughout all experimental phases with no differences between conditions (P > 0.05).

DISCUSSION

The purpose of the present study was to determine the relative importance of peripheral feedback from mechanically and metabolically sensitive muscle afferents and central command to the exercise-induced increases in regional cerebral perfusion. The major findings of the present study were twofold. First, selective activation of either mechanically (i.e., passive calf stretch) or metabolically (i.e., PEMI) sensitive skeletal muscle afferents failed to increase ACA V_mean. Second, ACA V_mean was increased during voluntary calf exercise but not during electrically stimulated calf exercise (i.e., central command absent) performed at the same absolute exercise intensity. Therefore, we must reject our initial hypothesis that muscle mechanoreflex activation would increase regional cerebral perfusion during exercise and suggest that increases in ACA V_mean during voluntary contraction of an exercising calf muscle are independent of muscle mechanoreceptor and metaboreceptor activation but are dependent on the presence of central command.

Previous findings have indicated that the handgrip exercise-induced increases in regional cerebral perfusion, determined using either ¹³³Xe inhalation or transcranial Doppler, are abolished by blockade of afferent feedback with regional anesthesia of the arm (13, 14, 25). This observation has been interpreted to suggest that feedback from skeletal muscle afferents has a critical influence on the CBF response to exercise, whereas central command has a negligible influence in this regard. Although these studies were unable to differentiate between the specific contributions made by metabolically and mechanically sensitive muscle afferents to the CBF response to exercise (13, 14, 25), subsequent work demonstrated that muscle metaboreflex had no influence on regional cerebral perfusion (26, 42). Thus, by deduction, an important role of the muscle mechanoreflex has been suggested (26, 42). To directly test this possibility, we examined the ACA V_mean responses to passive calf muscle stretch, a well-established method of stimulating muscle mechanoreceptors in humans, independent of central command or muscle metaboreceptor activation (7, 18). However, contrary to our expectation, passive calf stretch had no effect on regional cerebral perfusion. Furthermore, in the present study, neither isolated muscle metaboreflex activation (i.e., PEMI) nor indeed combined muscle metaboreflex and mechanoreflex activation (i.e., electrically stimulated calf exercise) had an influence on ACA V_mean. Therefore, in direct contrast to previous work using pharmacological blockade of sensory feedback with regional anesthesia (13, 14, 25), we were unable to find a significant influence of muscle afferent feedback on regional cerebral perfusion. While these discrepant findings are difficult to reconcile, experimental differences in the exercise modality used (handgrip vs. calf exercise) and cerebral artery assessed (middle cerebral artery vs. ACA), or a direct central effect of the anesthesia (36), may have contributed to the contrasting observations.

An important finding of our study is that the presence of central command appears to be prerequisite for a calf exercise-induced increase in ACA V_mean. Indeed, ACA V_mean failed to increase in all the experimental paradigms we used where central command was not activated (i.e., passive calf stretch, PEMI, and electrically stimulated calf exercise). In contrast, clear increases in ACA V_mean occurred in the only experimental protocol used where central command was present (i.e., voluntary calf exercise). One factor that may have contributed to this increase in ACA V_mean is an increase in CO (23, 40). However, CO was similarly increased during volitional and electrically stimulated exercise (15), whereas ACA V_mean was only increased during volitional exercise. We speculate that central command mediated increases in ACA V_mean are linked to local increases in cerebral activity and metabolism. Although no indexes of regional cerebral metabolism were included in the present study, in general agreement with our observations, several previous neuroanatomic studies have revealed central command-mediated increases in regional cerebral activity (35, 36, 49, 52). Indeed, several brain structures have been functionally linked with the central command-mediated cardiovascular response to exercise, including the sensory motor area, supplementary motor area, insular cortex, anterior cingulate, and thalamic regions (35, 36, 49, 52). In a recent study (53), it was reported that CBF increases within regions of the insular cortex and anterior cingulate during exercise but not during isolated muscle metaboreflex activation. One potential limitation to this conclusion is that the contribution of muscle mechanoreflex activation to the exercise-induced changes in CBF cannot be ruled out. However, the anterior cingulate cortex is perfused by the ACA (3), and, given that in the present study increases in ACA V_mean were only observed during the activation of central command and not during isolated muscle mechanoreflex activation, our findings may support the proposed association between the anterior cingulate and central command (52).

It has previously been reported that middle cerebral artery blood velocity responses to passive and unloaded active upper limb movement are similar, which may suggest a predominant role of sensory feedback under both these conditions (31). However, while this form of limb movement may be expected to increase sensory feedback from joint receptors, muscle spindles, and possibly Golgi tendon organs, it is likely that only a weak of activation of the muscle mechanoreceptors (passive and unloaded active movement), and negligible or no activation of the metaboreceptors and central command (unloaded active movement) would be elicited (16, 17, 32). As such, it is unclear whether the cerebral perfusion changes observed (31) are representative of exercise per se, where these neural pathways (i.e., mechanoreflex, metaboreflex, and central command) are more robustly activated. Importantly, we found that mechanoreflex activation using calf stretch did not significantly increase ACA V_mean, unlike voluntary calf exercise, where, along with the mechanoreflex, central command and the metaboreflex were also activated. Of particular note, ACA V_mean was unchanged during rhythmic calf stretch, which, along with activation of the mechanoreflex, would also be expected to increase sensory feedback from joint receptors, muscle spindles, and possibly Golgi tendon organs (17). As such, our findings do not support a predominant role for sensory feedback on the regional cerebral perfusion responses to exercise. However, we cannot rule out the possibility that experimental differences in the muscle group studied and cerebral artery assessed may also have contributed to the contrasting findings with previous work.

The magnitude of the increase in ACA V_mean we observed during rhythmic calf exercise is comparable with previously reported increases in regional cerebral perfusion (i.e., middle cerebral artery blood velocity) during rhythmic handgrip, as determined using the transcranial Doppler technique (25, 42). Furthermore, ACA V_mean returned to resting levels during PEMI after voluntary calf exercise, in keeping with previous work examining middle cerebral artery blood velocity responses to PEMI after rhythmic handgrip (26, 42). However, in contrast to our finding that voluntary isometric exercise elicits increases in ACA V_mean, previous work has suggested that regional increases in cerebral perfusion do not occur during isometric exercise (26, 44), although this has not been a universal finding (13, 53). Indeed, Jorgensen et al. (26) noted that middle cerebral artery blood velocity does not increase during isometric leg extensor exercise but is increased during leg cycling exercise. However, in the present study, we observed similar regional increases in cerebral perfusion during volitional rhythmic and isometric exercise. One potential explanation for the unchanged middle cerebral artery blood velocity reported during isometric quadriceps exercise (26) but increased ACA V_mean during isometric calf exercise is the link between the neuroanatomic specificity of the CBF response and the active muscle group. Indeed, Linkis et al. (29) demonstrated that ankle flexion and extension specifically increased V_mean in the contralateral anterior cerebral artery, whereas contralateral middle cerebral artery blood velocity specifically increased during right hand contraction. These findings may be attributable to the close relationship between the cortical projection of the exercising limb and its arterial blood supply. Indeed, while the cortical representation of the leg is supplied by the hemispheric ACA (3), the area of the motor cortex that controls the arm is supplied by the central branch of the middle cerebral artery (2). Therefore, it may be reasonable to expect that isometric quadriceps exercise would not increase middle cerebral artery blood velocity (26) but would lead to increases in ACA V_mean, although this remains to be determined.

In the present study, ACA V_mean remained unchanged despite elevations in BP during electrically stimulated exercise. The relatively constant ACA V_mean at this time may be due to cerebral autoregulatory mechanisms (9, 41). Indeed, we observed that CVCi was significantly reduced during electrically stimulated exercise. Although the cause for this decrease in cerebral vascular conductance is unclear, we suggest that this response may be the normal and necessary autoregulatory response required to offset the BP response to electrically stimulated contraction in the absence of an increase in cerebral activation and consequent requirement for increased perfusion. However, other factors cannot be discounted, such as an enhanced activation of sympathetic outflow directed to the cerebral vasculature. Indeed, there is accumulating evidence for an important role of sympathetic neural control of the cerebral circulation, although this remains a controversial issue (1, 37–39). In contrast to stimulated exercise, we observed that CVCi was essentially unchanged during voluntary exercise, possibly indicating that local dilator influences resulting from increases in cerebral metabolism related to the activation of central command, predominate over any potential cerebral vasoconstrictor stimuli associated with exercise (21).

Methodological Considerations

ACA V_mean was used as an index of regional cerebral perfusion. We acknowledge that changes in ACA V_mean are only proportional to changes in CBF if the ACA diameter remains unchanged (43). As ACA diameter was not measured the present study, we cannot completely rule out that changes in vessel diameter influenced the ACA V_mean measurements. However, a recent review of the literature concluded that the diameter of large cerebral arteries does not change significantly during exercise and that the regulation of CBF takes place in the smaller downstream vessels (46). Nevertheless, potential factors that may be argued to influence ACA diameter in our study include arterial Pco₂, sympathetic nerve activity, and BP. However, Pet_co2 was not changed by any of the experimental procedures implemented (e.g., calf stretch, exercise, and PEMI). Furthermore, in agreement with previous work (42), BP and presumably sympathetic activation were increased during voluntary exercise and muscle metaboreflex activation (i.e., PEMI), whereas ACA V_mean was only elevated during exercise. Overall, this suggests that voluntary exercise-induced increases ACA V_mean are independent of changes in Pet_co2, sympathetic nerve activity, and BP per se. A further consideration is that the ACA supplies a brain area including both grey and white matter; however, due to the inherent limitations of the transcranial Doppler technique, we are unable to distinguish between regional differences in perfusion within these areas.

Recent work has suggested that electrically stimulated exercise may not induce an identical metabolic stimulus to that observed during voluntary exercise (50). Indeed, it is possible that this may have contributed to the more marked BP elevation observed during PEMI after stimulated exercise than after voluntary exercise. However, it is unlikely that this had any influence on the ACA V_mean responses observed in the present study, as ACA V_mean was unchanged from rest during PEMI after either voluntary or stimulated exercise. Due to technical considerations, it is unclear whether the exercise intensity used influenced the findings of the present study. Indeed, performing electrically stimulated exercise at an intensity greater than the 35% MVC used in the present study could induce subject discomfort, confounding the activation of higher centers and consequent alterations in cerebral perfusion. Similarly, more robust isolated activation of muscle mechanoreceptors than that used in the present study may also lead to subject discomfort (e.g., stretching a muscle beyond the comfortable range of motion). In agreement with previous studies, sustained passive calf stretch was sufficient to elicit an increase BP (6, 7), although the magnitude of this change was smaller than have reported previously. However, this may be attributable to calf stretch being performed under ischemic conditions in earlier work (6, 7). Of note, the more marked BP response to sustained passive calf stretch compared with rhythmic stretch supports previous animal (27) and human (17) studies suggesting that sustained and rhythmic perturbations may differentially modulate mechanoreflex activation. Nevertheless, neither sustained nor rhythmic passive calf stretch elicited responses in ACA V_mean.

Our study included four women; however, our results were not affected when female subjects were excluded. To minimize any potential sex differences, female participants were studied at the early follicular phase of the menstrual cycle, in which plasma estrogen and progesterone concentrations are low (22).

Summary

The present data support the concept that in conscious healthy humans, selective activation of mechanically or metabolically sensitive muscle afferents does not increase ACA V_mean and that increases in ACA V_mean during volitional contraction of an exercising calf muscle are dependent on the presence of central command.

GRANTS

L. C. Vianna was supported by Brazilian National Council for Scientific and Technological Development Grant 201587/2007-6.