Diffuse correlation spectroscopy (DCS) is an emerging technique which detects the motion of moving scatterers inside biological tissues. It has been experimentally verified that DCS signals originate primarily from the motion of red blood cells (blood flow) in the microvasculature of biological tissues [30
]. These studies were carried out on tissues that were essentially static. However, the motions of nuclei and mitochondria inside muscle cells/fibers during exercise
may significantly contribute to the DCS signal decay rates and consequently affect blood flow measurements in moving muscles. In this study, we evaluated and separated the true blood flow signals and those affected by the muscle fiber motion artifacts using three plantar flexion exercise protocols.
The toe up/down plantar flexion exercise (Protocol #1) can be easily performed without the need of dynamometer. Since the toe up/down plantar flexion supports the entire body weight, its intensity is likely higher than the 30% MVIC isotonic or isometric exercise. Moreover, the frequency of toe up/down plantar flexion exercise (1/4 Hz) was higher than those of isotonic (1/9 Hz) and isometric (1/6 Hz) exercises. The high intensity and high frequency of toe up/down plantar flexion exercise should cause more obvious increase in DCS signals seen in compared to those seen in (isotonic exercise) and (isometric exercise). With protocol #1, however, it is impossible to accurately determine and separate motion artifacts from true blood flow signals (see ), because the muscle status (i.e., static, movement, contracting, and non-contracting) cannot be precisely tracked. In fact, these motion artifacts during exercise led to blood flow quantification errors in our previous study using this exercise protocol [32
]. In addition, the exercise intensity of toe up/down flexion may vary between subjects.
The isotonic and isometric plantar flexion exercises (Protocols #2 and #3) require a dynamometer, which increases the complexity and cost of measurements. However, the dynamometer provides precise recordings of the muscle status during exercise. Furthermore, compared to the isometric exercise (Protocol #3), the isotonic exercise of pushing the footplate (Protocol #2) is more naturally similar to the toe up/down exercise, and its intensity can be precisely and objectively controlled.
The precise co-registration of the dynamometer recordings and DCS data during isotonic exercise enabled us to evaluate and separate the contributions of muscle fiber motion artifacts from DCS determined blood flow indices. During the isotonic exercise, the muscle fiber motions created larger DCS signal decay rates than those without motion [see ]. The difference between the static and moment DCS signals clearly demonstrates the effect of motion artifacts. Furthermore, larger angular velocity PF generated larger motion artifacts than that of smaller angular velocity DF [see ], verifying the motional sensitivity of the DCS measurements. By contrast, DCS data obtained when the muscle was under static status during isotonic and isometric exercise (see –) were not significantly influenced by any motion artifact. Therefore, we believe the latter ‘static’ DCS signals originate primarily from the muscle blood flow.
The muscle exercise activity has two different phases (contracting and non-contracting (relaxation)), which affect the muscle blood flow level [51
]. Each time the muscles contract, arterial inflow decreases due to extravascular compression, and then arterial inflow increases as the muscles relax. If blood flow were measured in the outflow vein, the venous outflow would increase during contraction and decrease during relaxation; the opposite of what occurs on the arterial side of the circulation. In this study a higher level of DCS blood flow was observed during muscle contraction compared to that during muscle relaxation (see ), which is most likely due to the large increase in venous outflow when the muscles contracted. Since DCS measures blood flow in muscle microvasculature including arterioles, capillaries and venules, the blood volume (65–70% at rest) residing in the venous vasculature [52
] may contribute more to DCS measurements than that in the arterial vasculature.
As expected, without motion artifacts the isotonic exercise (Protocol #2) produced similar level DCS signals compared to the isometric exercise (Protocol #3) (see ) when the exercise intensity was set up at the same level (30% of highest MVIC). This observation further verifies that DCS measures blood flow in muscles.
Because of the lack of adequate technologies for continuous monitoring of blood flow during exercise, researchers previously focused on the evaluation of post-exercise blood flow responses in muscles [16
]. DCS can continuously monitor muscle blood flow before, during and after exercise with relatively high temporal resolution and without interrupting blood flow. These advantages are critical not only for continuous monitoring of flow responses during exercise, but also for tracking of the rapid and dramatic flow immediately after exercise (see ). In addition, the substantial difference in post-exercise flow responses between the toe up/down exercise and isotonic/isometric exercise is most likely due to the different intensities and frequencies of the exercises. Although it is unclear why the flow recovery towards baseline at 10-second post isotonic exercise is slower than that of post isometric exercises (see ), the muscle fiber response due to the contraction type (i.e., isotonic vs. isometric) may have contributed to the difference observed. Further investigations are needed to understand these differences.
We note that while our findings are encouraging, measurements with a larger sample size is desirable to confirm observations. Finally, the relatively low temporal resolution (2.2 Hz) of our DCS system has restricted the highest frequency of exercise. However, it is expected that with faster DCS systems (e.g. ~100 Hz [25
]) we should be able to monitor exercise at high speed (e.g., plantar flexion exercise at 1 Hz) in the future, and the DCS approach will therefore offer even more possibilities for monitoring natural muscle activities. Of course, as we have demonstrated in the present paper, the DCS measurements can be gated via monitoring of muscle status by a dynamometer to derive more accurate blood flow information.
In conclusion, this case study demonstrates that muscle fiber motions during exercise introduce artifacts into traditional (low repetition rate, ~1 Hz) DCS measurements of muscle blood flow responses to exercise. However, we show that the DCS signals due solely to blood flow can be precisely separated from those affected by motion artifacts using concurrent and co-registered dynamometer recordings with DCS data during exercise.