1.

Introduction

Pressure ulcer is a localized wound to the skin and/or underlying tissue, which affects up to 3 million adults in the United States.¹ Pressure ulcers occur when the skin and underlying tissues above the bone are compressed for a long period of time. Constant pressure against the tissue reduces blood supply to that area, which results in tissue ischemia/hypoxia and eventual necrosis.^2,3 The development of pressure ulcer can lead to several complications including sepsis, localized infection, cellulitis, osteomyelitis, pain, and depression.^4,5 Ultrasound imaging results have recently shown that early pressure ulcers may originate from deep tissues attached to the bone and spread upward, eventually to the skin.⁶ Therefore, measurements of deep tissue hemodynamics may provide information for early diagnosis and therapeutic monitoring of pressure ulcers before they are apparent on the skin surface.

Conventional ulcer care and treatment include debridement, dressing, nutrition, and surgical repair,^4,7 which are labor intensive and associated with high cost as they require the cooperation and skills of an interdisciplinary health care team. As one of the adjunctive therapies for ulcer management, transcutaneous electrical stimulation (ES) is simpler and less labor intensive with lower cost compared to conventional therapies. Transcutaneous ES has been used to treat both superficial tissues and deeper structures. ES at relatively high frequency (100 to 1000 Hz) and low current intensity ($<1\mathrm{mA}$) is often used to treat ulcers in superficial tissues^8–10 via triggering electrochemical repair mechanism of cells.^11,12 The delivered electrical potentials can influence activities and functional capacities of fibroblasts, macrophages, and neutrophils,^11,13,14 thus enhancing wound healing.

Deep muscles are found to be more susceptible to the external pressure, which deprives tissue of oxygen supply.¹⁵ More severely, injuries to muscle evolve undetectably until massive damages occur.¹⁵ Therefore, the treatment and prevention of pressure ulcers in deep muscles are of critical importance. As mentioned earlier, tissue ischemia/hypoxia resulting from the prolonged local pressure on the tissue is the major cause that eventually leads to pressure ulcer.^2,3,16 Correspondingly, improving tissue blood flow and oxygenation in deep muscles by ES is a straightforward way for the treatment and prevention of pressure ulcers. ES at relatively low frequency (40 Hz) and high current intensity (35 mA) can be used to treat ulcers in deep tissue.¹⁷ Although ES electrodes are placed on the surface of skin, the delivered electrical pulses can be transmitted to muscle end-plate, activating the muscle pump and eliciting muscle fiber contractions.^18,19 The contractile muscles pump arterial blood directly into capillaries,¹⁵ thus improving both microcirculation and tissue oxygenation. ES-induced muscle activities are especially critical for patients with spinal cord or brain injury since many of them lose an awareness of the warning signals that would prompt action against the prolonged pressure.⁷

The ulcer treatments by ES, however, have been performed empirically and ES parameters were selected randomly across a wide range.^{7,11,15,20–23} Treatment outcomes were simply evaluated by visible changes of tissue mass and/or wound volume^7,11 rather than intrinsic hemodynamic changes in ulcerous tissues. The random selection of ES parameters is mainly due to the lack of quantitative technologies for instantly and continuously evaluating tissue hemodynamic improvements that are closely associated with the increased muscle activity during ES. Moreover, minimizing muscle fiber motion artifacts in hemodynamic measurements during ES is known to be a technical challenge.

Technologies currently used to evaluate blood flow improvements by ES include Doppler ultrasound^21,22 and laser Doppler flowmetry (LDF).²³ Although Doppler ultrasound can assess blood flow in large vessels, the pressure ulcer is a condition caused mainly by poor microcirculation. LDF measures microvascular blood flow only in superficial tissues, not in deep muscles. Tissue oxygen level ($\mathrm{P}{\mathrm{O}}_2$) has been measured using ${\mathrm{O}}_2$ microelectrode or transcutaneous oxygen measurement ($\mathrm{TcP}{\mathrm{O}}_2$) to evaluate ES impact on muscle oxygenation.²⁴ However, ${\mathrm{O}}_2$ electrode measurement is invasive, and $\mathrm{TcP}{\mathrm{O}}_2$ is a complicated interaction between skin, fat, and muscle. Furthermore, although previous studies have evaluated ES impacts on either tissue blood flow^21–23 or oxygenation,^24–26 none of them have measured both quantities simultaneously. Since tissue blood flow and oxygenation are usually interactive, there is a critical need to develop noninvasive techniques that can continuously monitor both blood flow and oxygenation in deep muscles during ES.

Near-infrared spectroscopy (NIRS) offers a noninvasive, fast, portable, and low-cost alternative for repetitive bedside monitoring of deep tissue oxygenation.^27–33 NIRS can probe tissue hemodynamics millimeters to centimeters below the tissue surface as tissue absorption in the NIR range (650 to 950 nm) is relatively low. The NIR absorption spectra of major tissue chromophores, particularly oxy- and deoxy-hemoglobin, differ significantly. As a result, NIRS is sensitive to the changes in hemoglobin concentration and blood oxygen saturation (${\mathrm{StO}}_2$) in tissue microvasculature. NIRS has been used to evaluate ES impact on ${\mathrm{StO}}_2$ in skeletal muscles,^25,26 although it does not directly measure tissue blood flow. An emerging dynamic NIR technique, namely diffuse correlation spectroscopy (DCS), has also been developed for direct and noninvasive measurement of blood flow in various tissues,^31,34–44 including human skeletal muscles.^45–56 Similar to NIRS, DCS also utilizes NIR light to penetrate deep tissues. However, NIRS monitors tissue blood oxygenation variations through quantifying the relatively slow changes in light intensity level due to the slow changes in tissue absorption and scattering, whereas DCS monitors directly the motion of moving scatterers (mainly red blood cells) in tissue through quantifying the fast temporal fluctuations in light intensity.^47,57,58 Therefore, DCS is considered as a dynamic NIR technique as opposed to the static NIRS.³⁴ DCS measurements of tissue blood flow changes have been extensively validated to other standards, including power spectral Doppler ultrasound,⁴³ Doppler ultrasound,^35,40 LDF,^31,41 Xenon computed tomography,³⁹ fluorescent microsphere flow measurement,⁴⁴ and arterial-spin-labeled magnetic resonance imaging.⁵⁴ DCS has also been combined with NIRS in hybrid instruments to simultaneously measure tissue blood flow and oxygenation.^{36,38,47,49,52,53}

The hybrid DCS/NIRS instruments and combined fiber-optic probes are relatively large, complex, and expensive.^31,39,47 Recently, our group has developed a portable dual-wavelength DCS flow-oximeter⁴⁶ for simultaneous measurement of variations in tissue blood flow and oxygenation. This portable device has been used to investigate hemodynamic changes in muscle,^46,48,50,51 brain,^41,42 and tumor.³⁷ The goal of the present study was to integrate the novel DCS flow-oximeter and a muscle electrical stimulator to noninvasively evaluate ES treatment effects and optimize ES parameters for maximizing hemodynamic improvements in deep tissues. We applied ES pulses at low, medium, or high frequency (2, 10, or 50 Hz) to generate a single twitch, multiple twitches, or fused tetanus of the quadriceps muscle. A novel gating algorithm was developed to control data acquisition at time points when muscle fiber motion was minimal. Muscle hemodynamic responses to ES were continuously monitored, and the results at different stimulation frequencies were compared for the optimization of treatment effect. In addition, local hemodynamic changes in contralateral leg muscle (without ES) as well as systemic changes (blood pressure and heart rate) were also continuously monitored to evaluate the interactions among the local muscles and the global system of whole body. To the best of our knowledge, we report in this article the first simultaneous and continuous measurement of muscle blood flow and oxygenation responses during ES.

2.

Methods

2.1.

Integrated System for Hemodynamic Monitoring of Muscle Responses to ES

Figure 1 illustrates an integrated system including a custom-made DCS flow-oximeter [Fig. 1(a)],^{36,41,42,45,46,48,49,52,53} a laptop controller [Fig. 1(b)], a rectify circuit [Fig. 1(c)], and a four-channel commercial muscle stimulator [Theratouch 4.7, Rich-Mar Corp., Tennessee, Fig. 1(d)]. The DCS flow-oximeter was used to monitor muscle hemodynamic responses to ES delivered by the muscle stimulator. The biphasic pulses generated by channel #1 (CH1) of the stimulator were rectified by a custom-made circuit. The output from the circuit was read by a digital I/O line of an analog-to-digital (A/D) board (USB-6009, National Instrument Corp., Texas) to judge the status of ES “on” or “off.” A gating algorithm based on the ES status was designed to control the data acquisition at time points when muscle fiber motions were minimum (i.e., ES was off). The integrated system was controlled automatically by the A/D board connected to the laptop controller via an USB port. Details for each component of this integrated system are described below.

Fig. 1

A schematic illustration of the integrated system, including a custom-made diffuse correlation spectroscopy (DCS) flow-oximeter (a), a laptop controller (b), a rectifying circuit (c), and a four-channel muscle stimulator (d). A pair of electrodes was placed on the right leg and two fiber-optic probes were taped on the two legs, respectively (e).

2.4.

Gating Control for Minimizing Muscle Motion Artifacts in Optical Measurements

Optical measurements were gated based on ES outputs indicating no stimulation such that muscle fiber would be at resting status (relaxation) during data collection to minimize motion artifacts. This gating algorithm is similar to that reported in our previous study.⁴⁹ Both algorithms aimed to record data at the time when the muscle is at resting status. In our previous study,⁴⁹ forearm muscle status during a handgrip exercise was judged directly from the hand position output by a dynamometer. In the present study, we built a circuit to detect the occurrence of ES pulses on the thigh muscle. As an example, the gating algorithm for ES at 2 Hz is illustrated in Fig. 2. The gating algorithms for ES at 10 and 50 Hz followed the same principle. A rectifying circuit [a silicon diode plus a 200 Ω resistor, Fig. 1(c)] was designed to cut off the negative pulses from CH1, and its output was connected to a digital I/O line of the A/D board for reading the “on” or “off” status of ES pulses. Since the outputs from all channels of the stimulator were synchronized, muscle status (contraction or relaxation) controlled by CH2 could be read from the CH1 output (on or off). A digital counter with a temporal resolution of 100 ns was set in the A/D board to detect the ES pulses generated from CH1 [Fig. 2(a)]. When ES was on, pulse counts increased continuously from the initial value of zero. Once pulse counts remained unchanged within a short period of time slightly longer than the interval of ES pulses (e.g., 500 ms at 2 Hz), ES was considered to be “off” and the counter was then reset to zero [Fig. 2(b)]. After a short period of delay (300 ms) to account for the residual motion of muscle fibers after contraction [Fig. 2(c)], the DCS flow-oximeter started to collect one frame of hemodynamic data within 1.5 s [Fig. 2(d)]. This gating algorithm enabled data recording only at time points when muscle fiber motion was minimal.

Fig. 2

A gating algorithm for optical data acquisition during electrical stimulation (ES) at 2 Hz. The outputs from channel #1 (CH1) (a) and channel #2 (CH2) (c) of muscle stimulator were synchronized; thus the muscle status (contraction or relaxation) controlled by CH2 could be read from the CH1 output (“2 s on” or “2 s off”). A digital counter in the A/D board was used to detect the ES pulses generated from CH1. The pulse counts increased continuously when ES was on and remained unchanged when ES was off (b). Optical data were recorded within 1.5 s when the ES was off (d).

3.

Results

All subjects completed the protocols without adverse complications. The ATTs over 10 subjects were $0.76\pm 0.07$ and $0.74\pm 0.06\mathrm{cm}$ in muscles with and without ES, respectively. The ES currents applied to the subjects were $48.9\pm 4.2\mathrm{mA}$ at 2 Hz, $33.4\pm 3.4\mathrm{mA}$ at 10 Hz, and $25.7\pm 3.3\mathrm{mA}$ at 50 Hz, respectively. The subjects tended to tolerate less ES current intensity at higher frequency.

3.1.

Effect of Gating Control in Optical Measurements

Figure 3 shows results from a preliminary test in a single subject demonstrating the effect of gating control in DCS flow measurements during 3-min ES at 2 Hz. The two trials were performed using the same ES current intensity and separated by more than 30 min. Without gating, ES-induced muscle fiber motions introduced large artifacts in DCS data, leading to substantial overestimates of rBF [Fig. 3(a)]. The gating algorithm greatly reduced noises due to motion artifacts [Fig. 3(b)].

Fig. 3

Relative blood flow (rBF) responses during 3-min ES at 2 Hz, measured by the DCS flow-oximeter without (a) or with (b) the gating algorithm. The time period between the first and second marks (vertical lines) indicates the transition to adjust the ES current from zero to the maximum tolerant level (i.e., ES setup). The optical data collected during the period of ES setup were excluded due to the large noises induced by the adjustment of ES current. The ES started from the second mark and ended at the third mark. The gating algorithm (b) significantly reduced noise due to motion artifact.

3.2.

Time Course Physiological Responses to ES

Figures 4 and 5 show average time course physiological responses to ES over 10 subjects measured from stimulated (right leg) and control (left leg) muscles, respectively. Physiological variables measured included rBF, $\mathrm{\Delta }[\mathrm{Hb}{\mathrm{O}}_2]$, $\mathrm{\Delta }[\mathrm{Hb}]$, ABP, and heart rate. Physiological responses to ES at low (2 Hz), medium (10 Hz), and high (50 Hz) frequencies are demonstrated in (a), (b), and (c), respectively. Data obtained from 10 subjects were aligned based on the marks indicating the beginning and end of the ES (see the vertical lines in figures). For clarity of data presentation, a 1-min averaging window was applied on and shifted over the raw data for each subject. The processed datasets from 10 subjects were then averaged and displayed as mean and standard errors (error bars).

Fig. 4

The average time course physiological responses [rBF, $\mathrm{\Delta }[\mathrm{Hb}{\mathrm{O}}_2]$, $\mathrm{\Delta }[\mathrm{Hb}]$, arterial blood pressure (ABP) and heart rate] to ES over 10 subjects measured from the stimulated muscles in the right legs with ES at 2 Hz (a), 10 Hz (b), and 50 Hz (c). Data obtained from 10 subjects were aligned based on the marks (vertical lines) made at the beginning and end of ES.

Fig. 5

The average time course hemodynamics responses to ES over 10 subjects measured from the control muscles in the left legs without ES. The hemodynamic responses (rBF, $\mathrm{\Delta }[\mathrm{Hb}{\mathrm{O}}_2]$, $\mathrm{\Delta }[\mathrm{Hb}]$) to ES at low (2 Hz), medium (10 Hz), and high (50 Hz) frequencies are displayed in (a), (b), and (c), respectively. The two vertical lines (marks) indicate the beginning and end of ES.

In the muscles with ES treatment (Fig. 4), rBF increased persistently during ES at all frequencies and maintained at a level higher than its baseline for 20 min post ES. There were remarkable increases in $\mathrm{\Delta }[\mathrm{Hb}{\mathrm{O}}_2]$ during and post ES at 10 Hz [Fig. 4(b)], but not at other two frequencies [2 and 50 Hz, Figs. 4(a) and 4(c)]. Changes in $\mathrm{\Delta }[\mathrm{Hb}]$ were not apparent during and post ES at all frequencies. ABP increased during ES and maintained higher than its baseline post ES, while heart rate varied around baseline throughout ES at all frequencies.

In the contralateral control muscles without ES (Fig. 5), rBF increased slightly and $\mathrm{\Delta }[\mathrm{Hb}{\mathrm{O}}_2]$ and $\mathrm{\Delta }[\mathrm{Hb}]$ varied around baseline throughout ES. However, these changes were much smaller than those observed in the stimulated muscles with ES (Fig. 4).

3.3.

Average Physiological Changes During and Post ES

As described in Sec. 2.6, the mean value of data points during the last 3 min of ES or recovery period was used to quantify the physiological changes (relative to pre-ES baseline) during or post ES, and results are presented in Figs. 6 and 7, respectively. Two-way RMA analyses showed that muscle site significantly affected the changes in rBF and $\mathrm{\Delta }[\mathrm{Hb}]$ among different time periods ($p=0.01$ and 0.03 for rBF and $\mathrm{\Delta }[\mathrm{Hb}]$ respectively), leading to subsequent RMA analyses of time period adjusting for ES frequency in stimulated and control muscles, respectively. Two-way RMA analyses also showed that both muscle site and ES frequency significantly affected the changes in $\mathrm{\Delta }[\mathrm{Hb}{\mathrm{O}}_2]$ ($p=0.002$ and 0.03 for the interactions between time period and muscle site and between time period and ES frequency, respectively), leading to subsequent RMA analyses of time period in each combination of muscle site and ES frequency.

Fig. 6

Average physiological changes ($\text{mean}\pm \mathrm{SE}$) over 10 subjects in rBF (a), $\mathrm{\Delta }[\mathrm{Hb}{\mathrm{O}}_2]$ (b), $\mathrm{\Delta }[\mathrm{Hb}]$ (c), ABP (d), and heart rate (e) during ES at 2, 10, and 50 Hz. The changes for each subject were calculated by averaging the data over the last 3 min during ES.

Fig. 7

Average physiological changes ($\text{mean}\pm \mathrm{SE}$) over 10 subjects in rBF (a), $\mathrm{\Delta }[\mathrm{Hb}{\mathrm{O}}_2]$ (b), $\mathrm{\Delta }[\mathrm{Hb}]$ (c), ABP (d), and heart rate (e) after ES at 2, 10, and 50 Hz. Individual changes were calculated by averaging the physiological data over the last 3 min of the recovery period.

In stimulated muscles, RMA showed that rBF differed significantly among different time periods. Compared to pre-ES baseline values, rBF increased significantly and persistently during ES [$p<0.001$, post hoc LSD, Fig. 6(a)] and post ES [$p<0.001$, post hoc LSD, Fig. 7(a)], with the adjustment of ES frequency effect.

Significant increases in $\mathrm{\Delta }[\mathrm{Hb}{\mathrm{O}}_2]$ (compared to pre-ES values) were also observed during ES [$p=0.04$, post hoc LSD, Fig. 6(b)] and post ES [$p=0.02$, post hoc LSD, Fig. 7(b)] at 10 Hz in stimulated muscles. There was also significant increase in $\mathrm{\Delta }[\mathrm{Hb}{\mathrm{O}}_2]$ during ES at 2 Hz [Fig. 6(b)], but not post ES at this stimulation frequency. Since two-way RMA analyses showed that ES frequency significantly affected $\mathrm{\Delta }[\mathrm{Hb}{\mathrm{O}}_2]$ changes among different time periods, additional analyses were performed to evaluate the effects of ES frequency on $\mathrm{\Delta }[\mathrm{Hb}{\mathrm{O}}_2]$ changes. The increases in $\mathrm{\Delta }[\mathrm{Hb}{\mathrm{O}}_2]$ at 10 Hz were significantly larger than those during ES at 50 Hz ($p=0.04$, post hoc LSD) and those post ES at 2 and 50 Hz ($p=0.03$ for both 2 and 50 Hz; post hoc LSD).

Taken together, ES at medium frequency (10 Hz) significantly and persistently increased both blood flow (rBF) and oxygenation ($\mathrm{\Delta }[\mathrm{Hb}{\mathrm{O}}_2]$) in stimulated muscles throughout ES (i.e., both during and post ES).

In addition, no significant differences (post hoc LSD) of hemodynamic changes were found in stimulated muscles between during and post ES. Linear regressions revealed that rBF changes were not significantly correlated with tissue oxygenation changes in most conditions. Exceptions included that rBF changes were correlated with $\mathrm{\Delta }[\mathrm{Hb}{\mathrm{O}}_2]$ during ES at 50 Hz ($R^2=0.51$, $p=0.02$) and post ES at 2 Hz ($R^2=0.67$, $p=0.004$).

In the contralateral (control) muscles without ES, some variables demonstrated significant changes during ES [$\mathrm{\Delta }[\mathrm{Hb}]$ decreased, Fig. 6(c)] or post ES [rBF increased, Fig. 7(a)]. However, additional analyses revealed that rBF increases post ES in control muscles were significantly ($\mathrm{p}<0.001$, post hoc LSD) smaller than those observed in stimulated muscles [Fig. 7(a)]. Furthermore, there were no significant increases in $\mathrm{\Delta }[\mathrm{Hb}{\mathrm{O}}_2]$ during [Fig. 6(b)] and post [Fig. 7(b)] ES in control muscles at all three frequencies.

Significant increases ($p<0.001$) in ABP were also observed post ES [Fig. 7(d)]. The changes in heart rate were not significant [Figs. 6(e) and 7(e)]. No correlations were found among muscle hemodynamic responses, ABP, and heart rate throughout ES.

4.

Discussion and Conclusions

The goal of our ES treatment is to maximize tissue hemodynamics in local muscles. Continuous measurement of muscle hemodynamics during ES provides valuable information for objectively optimizing ES effect. However, technologies available for this purpose are limited. In this study, we developed a unique integrated system (see Fig. 1) that incorporated a novel DCS flow-oximeter and a muscle stimulator for noninvasive and continuous monitoring of muscle blood flow and oxygenation responses to ES. A gating algorithm was designed to control the data collection only at time points when muscle was under relaxation (see Fig. 2), which greatly reduced noises due to muscle fiber motion during ES (see Fig. 3). We also compared tissue hemodynamic improvements generated by different types of muscle contractions. ES at relatively low frequency (2 to 40 Hz) and high current intensity (10 to 45 mA) have been previously found to be efficient in enhancing tissue blood flow for the treatment of pressure ulcers.^2,61,17 Based on those previous results, electrical stimulations at 2, 10, and 50 Hz were thus selected in the present study as representative frequencies to generate a single twitch, multiple twitches, and fused tetanus on the muscle, respectively. The current intensities (14 to 75 mA) were selected based on subjects’ maximal pain tolerance.

Our ultimate goal is to use ES to treat pressure ulcers in patients with spinal cord injury (SCI). However, this pilot study aimed to adapt our novel optical techniques for evaluating ES treatment effects and to optimize ES parameters for maximizing hemodynamic improvements in deep tissues. Healthy subjects were recruited for this study because they have the ability to sense pain, which allowed us to determine their maximal tolerance to ES currents. Moreover, the use of healthy subjects for repeated testing permitted us to optimize our techniques and ES protocols. It is not practical to optimize our techniques and ES parameters using patients with SCI due to their poor mobility and loss of pain sensation. Although patients with SCI may have partial or complete impairments in their central nervous system function that controls muscle contractions, externally applied ES pulses can still stimulate their muscle pump activity by triggering the cross-bridge cycling between actin and myosin.¹⁹ Consequently, ES-induced blood flow increases in patients with SCI have been observed to be similar to those in healthy subjects.^19,62 Therefore, the knowledge gained from the present study with healthy subjects is expected to be beneficial for future studies using SCI patients with pressure ulcers.

For future translation of the developed ES protocols to SCI patients with pressure ulcers, we may gradually increase the ES current intensity to the point of visible muscle contractions (for those without pain sensation) or maximal pain tolerance (for those with pain sensation). In addition, muscle contraction can be evaluated by detecting the electrically evoked force,⁶³ which may be an alternative standard for selecting ES current intensity.

Experimental results from this study demonstrate that the short-term ($\sim 20\min$) ES can significantly increase muscle rBF during ES (see Figs. 4 and 6) and maintain a higher level than its pre-ES baseline for at least 20 min post ES (see Figs. 4 and 7), regardless of the applied stimulation frequencies. Previous studies also found significant blood flow increases in quadriceps muscle^21,22 or skin²³ when applying ES at medium (10 Hz)²² or high (30 Hz)^21,23 frequency. In those studies, muscle or skin blood flow was measured by Doppler ultrasound^21,22 or LDF.²³ Although ES-induced increases in rBF had led to some improvements in tissue oxygenation (i.e., increases in $\mathrm{\Delta }[\mathrm{Hb}{\mathrm{O}}_2]$ and decreases in $\mathrm{\Delta }[\mathrm{Hb}]$, see Figs. 4 to 7) during or post ES at some ES frequencies, only ES at medium frequency (10 Hz) improved $\mathrm{\Delta }[\mathrm{Hb}{\mathrm{O}}_2]$ significantly and persistently during and post ES (see Figs. 6(b) and 7(b)). Maintenance of elevated rBF or tissue oxygenation post ES has also been reported in other studies,^15,23,64 and it was attributed to the reactive hyperemia induced by ES.¹⁵

The level of tissue blood oxygenation reflects the balance between oxygen supply (blood flow) and oxygen consumption. These three parameters (i.e., blood flow, blood oxygenation, tissue oxygen consumption) may or may not be coupled, depending on physiological manipulations/procedures. Tissue oxygen consumption may substantially affect the coupling between the blood flow and blood oxygenation levels. For example, a high intensity foot pedal plantar flexion exercise can lead to the increases in blood flow and oxygen consumption as well as a decrease in tissue blood oxygen saturation in calf muscles due to the high tissue oxygen consumption induced by the exercise.^47,65 In the present study, ES at 10 Hz generated effectively multiple twitches of muscle fibers, which may lead to large variations in tissue oxygen consumption affecting the coupling of blood flow and tissue oxygenation. This could be the reason of why no significant correlation existed between the rBF and $\mathrm{\Delta }[\mathrm{Hb}{\mathrm{O}}_2]$ when using ES at 10 Hz. The poorly coupled hemodynamic responses further emphasize the need for simultaneous measurements of both blood flow and oxygenation changes. Previous studies have also reported either poorly coupled^47,48,65,66 or well-correlated³⁹ changes between the blood flow and tissue oxygenation in response to different physiological manipulations.

A single ES pulse can generate a cycle of muscle fiber twitch including a fast contraction (30 to 50 ms) and a relatively slow relaxation (100 to 300 ms).^67–69 When the interval of ES pulses (e.g., 500 ms at 2 Hz) is larger than a twitch cycle ($<350\mathrm{ms}$), different twitches are isolated and muscle fibers present the feature of a single twitch.^67,70 When the pulse interval (e.g., 100 ms at 10 Hz) is smaller than a twitch cycle ($<350\mathrm{ms}$) but larger than a contraction time ($<50\mathrm{ms}$), twitches are accumulated during the period of relaxation and muscle fibers present the feature of multiple twitches.^67,70 When the pulse interval (e.g., 20 ms at 50 Hz) is smaller than a contraction time ($<50\mathrm{ms}$), twitches are accumulated throughout the ES time and muscle fibers present the feature of fused tetanus.^67,69,70

The multiple twitches of muscle fibers generated by ES at 10 Hz repeatedly pump arterial blood into tissue, thus constantly elevating oxygen level. Although the single twitch generated by ES at 2 Hz also improves muscle perfusion in intermittent pattern, the transient oxygen increment from single muscle contraction may be rapidly balanced by tissue oxygen consumption. Thus, limited muscle fiber twitches generated by ES at low frequency (2 Hz) are likely not sufficient to generate sustained elevation of tissue oxygen levels. ES at 50 Hz recruits bulk muscle fibers to contraction and generates large contractile force, but the fused tetanus pattern restricts the release of muscle contraction. Therefore, ES at high frequency (50 Hz) actually generates a single and longer-term contraction of bulk muscle fibers in pumping blood to tissue, which is functionally similar to ES at low frequency (2 Hz). Taken together, it is not surprising that ES at 10 Hz promotes muscle oxygen level efficiently.

ES also causes significant increases in local rBF of contralateral control leg muscles for some frequencies [see Figs. 6(a) and 7(a)] and in global ABP [see Fig. 7(d)]. These changes are likely due to systemic influences of ES on the whole body. Such systemic influences were previously found in other studies^25,64 and were attributed primarily to the increase in sympathetic nerve activity due to ES.^20,25 The sympathetic nerve activity regulates ABP as well as the release of neurotransmitters that control the muscle vessel vasodilation and vasoconstriction.^20,25 Nevertheless, hemodynamic changes in control muscles without ES were found to be significantly lower than those in stimulated muscles with ES (see Figs. 4 to 7), and no correlations were observed among muscle hemodynamic responses, ABP, and heart rate throughout ES. These findings suggest that hemodynamic improvements in stimulated muscles with ES are mainly due to the local muscle contraction rather than the systemic impact.

One concern for this study is that the DCS flow-oximeter for deep muscle measurements may be influenced by the overlaying skin and adipose tissues. It is known that the penetration depth of NIR light in biological tissues is approximately half of the S-D separation.⁷¹ The probing depth of $\sim 1.3\mathrm{cm}$ with the S-D separation of 2.5 cm used in the present study is much larger than the ATTs ($0.76\pm 0.07$ and $0.74\pm 0.06\mathrm{cm}$ on top of the stimulated and control muscles, respectively), thus allowing for the detection of deep muscle hemodynamics. Nevertheless, future study will use optical probes with multiple S-D separations to detect hemodynamic responses in different tissue layers (i.e., skin, adipose tissue, and muscle) and correct the influence of overlying tissues on top of the muscle.

Another concern is the laser wavelengths used in DCS flow-oximeter. The laser at 785 nm is not ideal for tissue oxygenation measurement as the extinction coefficients of oxy- and deoxy-hemoglobin at this wavelength are not greatly different. Selection of the two wavelengths was restricted by the commercial availability of long-coherence lasers ($>5\mathrm{m}$) for concurrent DCS blood flow measurements. In a previous study, we have validated a dual-wavelength (785 and 854 nm) DCS flow-oximeter for tissue oxygenation measurements against a commercial tissue oximeter (Imagent, ISS Inc., Illinois); the oxy- and deoxy-hemoglobin concentration changes measured by the two devices were found to be highly correlated.⁴⁶ Furthermore, this dual-wavelength DCS flow-oximeter has been broadly used in the study of skeletal muscle,^46,48,50,51 brain,^41,42 and tumor.³⁷ Nevertheless, optimization of wavelengths will be the subject of our future work.

In conclusion, we have integrated the novel DCS flow-oximeter with a muscle stimulator to noninvasively and continuously monitor muscle hemodynamic improvements during and post ES at low (2 Hz), medium (10 Hz), and high (50 Hz) frequencies. Muscle fiber motion artifacts in optical measurements during ES were mitigated using a novel gating algorithm. Multiple physiological parameters, including blood flow and oxygenation in local leg muscles with or without ES, blood pressure, and heart rate, were concurrently measured during ES for the first time. Multiparameter measurements provide deeper insight into tissue physiology and allow us to investigate the complex relations of local (muscle) and systemic (blood pressure and heart rate) responses to ES. We found that multiple twitches of muscle fibers generated by ES at medium frequency (10 Hz) significantly and persistently increased both muscle blood flow and oxygenation throughout ES. Importantly, none of the 10 participated subjects experienced any discomfort during this ES intervention. This study supports the application of the integrated system to quantify muscle hemodynamic improvements for the optimization of ES treatment and holds potential for improving treatment evaluation for patients suffering from diseases caused by poor blood circulation and low tissue oxygenation (e.g., pressure ulcer).