on of poor time resolution of blood flow. Several other techniques have previously been developed that enable estimates to be made of arterial inflow, venous outflow, and local blood flow within a muscle [21- 27]. Whereas many of the techniques are impaired by different methodological limitations, the indicator thermodilution and the ultrasound Doppler method have both been found to give repeatable measurements of the same magnitude during both rest and dynamic knee extensor exercise [3,28] (Figure 2).
There is a positive linear relationship between leg blood flow in femoral artery and target work rate (10, 20,
Figure 2. Blood flow during incremental one-legged dynamic kicking (knee extensor) exercise at 60 contractions per minute, estimated from mean blood velocity and averaged on a beat-by-beat basis or continuously measured and analyzed in relation to muscle contraction cycle. The blood flow increases linearly with incremental target exercise intensities of work rate during steady-state exercise. This implies that an enhanced vasodilatation is elicited, in relation to the increased average muscle force exerted at high work rates, to meet the elevated metabolic activity. Figure adapted from Rådegran , reproduced with permission from The American Physiological Society.
30, 40, 50 and 60 watt) in relation to rhythmic thigh muscle contractions at 60 contractions per minute (Figure 2). With the rapid increase in energy requirements during exercise, equally rapid circulatory adjustments are essential in order to meet the increase need for oxygen and nutrients by the exercising muscle. In addition, thermodilution blood flow measurements obtained under similar experimental conditions by Andersen and Saltin  are closely related to those obtained by Doppler ultrasound. Thus, blood flow measured by Doppler ultrasound is valid not only at rest but also during incremental one-legged dynamic knee extensor exercise. The precision and accuracy of the Doppler technique has been improved by sampling the blood velocity (muscle contraction cycle) continuously instead of averaging the velocity in relation to each cardiac cycle (beat-by-beat cycle) (see ↔ in Figure 3).
The Doppler technique can be used to differentiate between physiological and methodological variations in flow, as well as detect rapid changes in flow induced by exercise (dynamic or static), different metabolic states (muscle contraction intensity or frequency), or any other type of vasodilatation such as the reperfusion period after arterial occlusion or infusion of a vasodilator substance.
3.2. Physiological Variations in Blood Velocity during Rhythmic Muscle Contractions
The continuous recordings can clearly determine the magnitude of the physiological variability in blood velocity by the contraction-relaxation-induced variations in muscle force, and consequently the intramuscular pressure variations, along with the superimposed influence of the blood pressure as well as the tonic influence of the state of vasodilatation [3,13,29]. The high intramuscular pressure during muscle contractions may consequently temporarily reduce or even reverse the blood velocity, depending on the relationship between the intramuscularand arterial blood pressure. The major extent of the blood velocity and flow consequently occurs during the muscle relaxation phase [3,13,30]. Blood velocity fluctuated in relation to the state of vasodilatation and the muscle contraction-relaxation duty cycles, indicated by the oscillations in muscle force.
In generally, the blood velocity increased to its highest value at the systolic blood pressure phase during muscle relaxation, and significantly decreased to its lowest value at the diastolic blood pressure phase during muscle contraction (Figure 3). The blood velocity showed an intermediate value at the systolic blood pressure phase during muscle contraction and at the diastolic blood pressure phase during muscle relaxation, respectively. The blood velocity curve was furthermore retrograde in the diastolic blood pressure phase during muscle contraction.
Figure 3. Continuous recording of blood velocity, blood pressure and muscle force during steady-state one-legged dynamic kicking (knee extensor) exercise at 20 watt and 60 contractions per minute. Figure adapted from Osada and Rådegran , reproduced with permission from Edizioni Minerva Medica.
In Figure 4 the limited view of blood velocity profile in relation to single muscle contraction-relaxation and single cardio systolic-diastolic beat, the blood velocities during the systolic and diastolic phases were found continuously in parallel with the blood pressure curve during the muscle contraction and muscle relaxation phases determined from the electromyography and the muscle force curve.
Four variations in the coupling between the blood pressure curve and the state of muscle contraction and relaxation were indicated; the systolic phase during muscle contraction, the diastolic phase during muscle contraction, the systolic phase during muscle relaxation, and the diastolic phase during muscle relaxation. The formation of the blood velocity profile and flow was influenced by the intramuscular pressure, as indicated by the muscle force curve, and the superimposed influence of the blood pressure in relation to the systolic and diastolic phases. The magnitude in blood flow value measured by the 4 variations indicates the large difference in work rates during the muscle relaxation phase and those of the muscle contraction phase at systolic and diastolic points, respectively (Figure 5).
The difference in the blood velocity values due to normal physiological variations raises the question as to how much sampling duration is necessary to determine the acceptable range of valid net-blood velocity value during steady-state rhythmic repeated voluntary muscle contraction exercise.
3.3. Muscle Contraction-Induced Blood Velocity Variability and Net-Blood Velocity Value Evaluated by Sampling Number of Muscle Contraction-Relaxation Cycle
A physiological variation in femoral artery is observed due to each muscle contraction during steady-state, onelegged, repeated kicking (dynamic, knee-extensor) exercise as mentioned above in Section 3.2. This was predo-
Figure 4. Blood velocity profile for the systolic and diastolic phases during the muscle contraction and muscle relaxation phases at 20 watt and 60 contractions per minute. The arrows down and up indicate the influence on the blood velocity, depending on the magnitude of, and temporal relation between, the muscle force (≈intramuscular pressure) and the blood pressure, respectively. Figure adapted from Osada and Rådegran , reproduced with permission from The Physiological Society of Japan.
minantly related to the muscle contraction-induced oscillations in intramuscular pressure and the influence of the superimposed waves on the cardiac cycle and arterial perfusion pressure.
During steady-state exercise with the rhythmic muscle contractions at 60 cpm (0.5 s muscle contraction-0.5 s muscle relaxation), it was found that the blood flow variability was smaller from 15-s~ to ~60-s (longer periods) than shorter periods, such as 3-s, as determined by averaging blood flow measurements over time periods . This difference in blood flow variability may depend on the sampling duration which may potentially include the various magnitudes in fragments of exercising blood velocity profile (see ↔ in Figures 3 and 4).
The magnitude of the muscle contraction-relaxationinduced physiological variability in blood velocity between “consecutive muscle contraction cycle” during steady-state rhythmic thigh muscle contractions at different contraction intensities (10, 20, 30 and 40 watt) and contraction frequencies (30 and 60 cpm) is examined (see simultaneous recording of hemodynamics parameters at 20 watt and 60 cpm in Figure 6). The steadystate suggests the condition of less variability of voluntary muscle contraction force, blood pressure and heart rate.
Measurement with the methodological concept for steady-state net-blood flow value requires determining whether the muscle contraction-relaxation cycle induced flow variability will be reduced by sampling more muscle contraction-relaxation cycles. In addition, it was important to determine the sampling duration of the muscle contraction-relaxation cycle that had the smallest variability required to obtain accurate steady-state netblood flow measurements at rest and during one-legged, repeated kicking (dynamic knee extensor) exercise. The blood flow variability was determined for seven con-
Figure 5. Blood flow at rest between systolic and diastolic phase at muscle contraction and relaxation during 5 - 40 watt exercise at 60 contractions per minute. There is clear increase in blood flow with an increase of work rate during muscle relaxation phase, however there is less change in blood flow between work rates during muscle contraction phase. Values are mean ± standard error. Figure adapted from Osada and Rådegran , reproduced with permission from The Physiological Society of Japan.
secutive cycles (1-, 2-, 5-, 10-, 15-, 20- and 30-cycles) for cardiac beat-by-beat cycles at rest, and for muscle contraction-relaxation cycles during steady-state exercise at 10, 20, 30 and 40 watt and for 30 and 60 cpm, as illustrated by Figure 6. The blood flow variability for seven consecutive cycles was estimated by the coefficients of variation.
In Figure 7, single cardiac beat-by-beat cyclesinduced blood flow variability was almost 15% at rest. Furthermore, single muscle contraction-relaxation cycleinduced physiological blood flow variability was coefficients of variation, range 11.4% - 13.2% at 30 cpm, and range 13.3% - 18.1% at 60 cpm between 10 and 40 watt. Both at rest and during steady-state exercise, the longer 30-cycles measurements represent a stable condition with minimal blood flow variations (Coefficients of variation, 4.1% - 4.3% at 30 cpm, 4.5% - 5.8% at 60 cpm). Furthermore, the 1-muscle contraction-relaxation cycle induced blood flow variability was similar between exercise intensity, but significant variations were seen between contraction frequencies at lower exercise intensities (below 15-cycles at 10 watt and below 5-cycles at
Figure 6. Analysis of blood velocity value for consecutive muscle contraction-relaxation cycles (20 watt and 60 contractions per minute). Blood velocity was measured at steady-state in relation to the muscle force curve for 1-, 2-, 5-, 10-, 15-, 20- and 30-consecutive contraction-relaxation cycles (CRcycles). Measurements were guided by the force curve from the initial phase of activity, which represents the initial contraction point (↑: Start of measurement). The letters depicted indicate; A: Muscle contraction at systolic blood pressure phase, B: Muscle contraction at diastolic blood pressure phase, C: Muscle relaxation at systolic blood pressure phase, D: Muscle relaxation at diastolic blood pressure phase corresponding to Figure 4. Figure adapted from Osada , reproduced with permission from Wolters Kluwer Health.
Figure 7. The coefficients of variations for blood flow determined by the consecutive muscle contraction-relaxation cycles (CRcycles). The blood flow variability was determined for seventh consecutive cycles (1-, 2-, 5-, 10-, 15-, 20- and 30-cycles) for cardiac beat-by-beat cycles at rest, and for CRcycles during steady-state exercise at 10, 20, 30 and 40 watt (W) and for 30 and 60 contractions per minute (cpm). The blood flow variability declined exponentially to a stable level; from 1- to 30 - cardiac beat-by-beat cycles at rest, and from 1- to 30-CRcycle during exercise. The blood flow variability is markedly reduced with a longer sampling measurement of at least 10-CRcycles with approximately 5%. Values are mean ± standard error. Figure adapted from Osada , reproduced with permission from Wolters Kluwer Health.
beat-by-beat cycle measurement may be closely related to the duration of the beat-by-beat cycle (precisely heart 20 watt. For steady-state rhythmic muscle contractions in the present exercise model, the findings demonstrated that blood flow variability was markedly reduced with a longer sampling measurement of at least 10-muscle contraction-relaxation cycle, which had coefficients of variation of approximately 5%. Above mentioned methodological consideration with samplings number (muscle contraction-relaxation cycle) may be one acceptable procedure for the determination of valid net-blood flow value such as rhythmic and dynamic muscle exercise. However, another exercise model such as intermittent muscle contraction (for instance, the ratio between muscle contraction and relaxation interval is not equal, or both intervals are too long) may cause differences in the magnitude in the muscle contraction-relaxation induced blood flow variability. Though the magnitude in blood flow variability may be different between exercise models, the information regarding its variability and impact on assessment of the steady-state exercising blood flow is available from the non-invasive Doppler method.
3.4. Comparison in Blood Flow Variability between Muscle Contraction-Relaxation Cycle and Cardiac Beat-by-Beat Cycle
The resting blood flow evaluated by the beat-by-beat cycle measurements may be a acceptable procedure because of the physiological close relation to magnitude of stroke volume. However, it is still unclear if the evaluation of exercising blood flow should be carried out by muscle contraction-relaxation cycle measurement or the beat-by-beat cycle measurement. The magnitude in netblood velocity determined between muscle contractionrelaxation cycle and beat-by-beat cycle is not similar as illustrated in relation to the blood pressure and force tracings (Figure 3). Therefore, the comparison in blood flow variability (the coefficients of variation) between each beat-by-beat cycle and each muscle contractionrelaxation cycle may provide the more acceptable information for the determination of physiological net-blood flow value during steady-state exercise at different exercise intensities.
Similar blood flow variability between different work rates is seen in muscle contraction-relaxation cycle (Figure 8(A)), however, larger variations in the blood flow with the range of ~18% - 29% were observed with an increase in work rate and a shorter duration of the 1- beat-by-beat cycle (Figure 8(B)). The duration of the 1- muscle contraction-relaxation cycle was, however, constant (≈1000 ms corresponding to 60 cpm) for the incremental exercise intensities (work rate) resulting in a smaller blood flow variability of approximately 15% .
Such blood flow variability determined for the 1- muscle contraction-relaxation cycle measurement has previously also been found to be closely related to the muscle force variability at 60 cpm but not at 30 cpm . On the other hand, the blood flow variability for the
Figure 8. Comparison in blood flow variability between muscle contraction-relaxation cycle (CRcycle) and cardiac beat-by-beat cycle (BBcycle) and its relationship to the duration of a cycle. (A) The coefficients of variation (CV) for blood flow are similar at each work rate when determined for the CRcycle. The CV for blood flow determined for the BBcycle were, however, significantly (*P < 0.05) higher at 40 watt (W) compared to at 10 W. A significant (†P < 0.01) difference in the CV for blood flow was furthermore seen between the BBcycle and CRcycle at 30 and 40 W, respectively. (B) As the CV in blood flow were closely (P < 0.01) related to the duration of the 1-BBcycle but not the 1-CRcycle, the duration of the 1-BBcycle may be assumed to influence the variations in blood flow. Values are mean ± standard error. Figure adapted from Osada and Rådegran , reproduced with permission from Edizioni Minerva Medica.
rate) with the relative influence of the variations in the blood pressure and intramuscular pressure variations.
The technological development of Doppler ultrasound may contribute to the examination of blood flow dynamics in the exercising muscles in human. The possibility of detection of blood velocity with high temporal resolution in real time at rest as well as during exercise is the advantage for using the Doppler method. However, the enhanced alterations in blood velocity profile may potentially confuse the evaluation of hemodynamics in the exercising muscles. In determining the exercising blood flow measures using non-invasive technique of Doppler ultrasound, the normal physiological variability should be considered when determining the precise physiological net-blood velocity/flow values for the research area in exercise and biomedical science. This review discusses how to obtain accurate measurements of spontaneous changes in exercising blood flow as measured by cardiac beat-by-beat cycle and the muscle contractionrelaxation cycle at various exercise intensities during rhythmmic leg exercise in humans.
The staff of The Copenhagen Muscle Research Centre is greatly acknowledged. This study was supported by the Danish National Research Foundation Grant 504-14, Uehara Memorial Foundation in 2002, a Grant-in-Aid for Young Scientists (B) in Scientific Research (No. 16700471) and the “Excellent Young Researchers Overseas Visit Program” in Scientific Research (No. 21-8285) 2010 from MEXT and JSPS.