Wavelet Phase Coherence Analysis of Oxyhemoglobin and DeoxyHemoglobin Oscillations to Investigate the Relationship Between Cups of Cupping Therapy

ABSTRACT

Research has not demonstrated whether multiple cups of negative pressure cupping therapy would induce interactions of hemodynamic responses between different areas. A multichannel near‐infrared spectroscopy (NIRS) was used to assess oxyhemoglobin and deoxyhemoglobin oscillations in response to cupping therapy. Wavelet transform and wavelet phase (WPC) coherence were used to quantify NIRS signals. Three levels of negative pressure (−75, −225, and −300 mmHg) were applied to the gastrocnemius in 12 healthy adults. Oxyhemoglobin coherence between the two inside‐cup areas was higher at −75 mmHg compared to −300 mmHg in both metabolic (WPC = 0.80 ± 0.11 vs. 0.73 ± 0.13) and neurogenic (WPC = 0.70 ± 0.11 vs. 0.60 ± 0.17) controls. Deoxyhemoglobin coherence was also higher at −75 mmHg compared to −300 mmHg in both metabolic (WPC = 0.78 ± 0.11 vs. 0.66 ± 0.14) and neurogenic (WPC = 0.67 ± 0.11 vs. 0.58 ± 0.13) controls. Our study provides first evidence on the interaction of hemodynamic responses between the two cups of cupping therapy using WPC analysis of NIRS signals.

The locations of the near‐infrared spectroscopy channels are labeled (Channels 1–16), and the skin areas are defined as inside or outside the treatment area of cupping therapy are I, II, III, and IV. Areas I and II are on the distal side while the other two areas are on the proximal side. The Areas I and III are outside the cup and the other two areas are inside the cup. The measured hemodynamic responses from the treated areas are analyzed using wavelet phase coherence to assess the interactions between different treated areas under two cupping cups.

1. Introduction

Cupping therapy is defined as an intervention by inducing negative pressure on an area of the skin or soft tissue [1, 2]. Recently, cupping therapy has been increasingly incorporated into management of muscular discomfort, myofascial trigger points, and musculoskeletal pain. The mechanical effect of cupping therapy is to cause deformation of underlying tissues, potentially leading to an increase in blood circulation in the treated area [3]. The increased blood flow within the skin and muscles induced by cupping therapy may facilitate the transport of nutrients and oxygen to cells and the removal of metabolic waste products [4].

On the one hand, cupping therapy has been shown to affect local blood flow regulation [5]. Studies using laser Doppler flowmetry (LDF) with wavelet analysis have demonstrated that cupping therapy can increase skin blood flow and such an increase is regulated through certain skin blood flow control mechanisms [6, 7]. Studies using near‐infrared spectroscopy (NIRS) found that cupping therapy increases the blood volume and oxyhemoglobin concentration in the cupping area [8, 9]. Additionally, increased blood flow could significantly activate the heme oxygenase‐1 (HO‐1) system, which has antioxidant and anti‐inflammatory health benefits [4]. On the other hand, cupping parameters such as the magnitude of pressure also significantly influence blood flow regulation. Studies demonstrated that cupping at −300 mmHg results in higher oxyhemoglobin and deoxyhemoglobin levels in muscle tissue compared to −225 and −75 mmHg [9]. These findings collectively suggest that cupping therapy could enhance blood flow of treated area, potentially leading to improved health and rehabilitation outcomes.

Nevertheless, it is important to consider potential risks of cupping therapy due to the negative pressure stretching the skin and underlying tissue. Case studies demonstrated that prolonged cupping could break superficial papillary capillaries causing capillary damage and bullae [10]. Negative pressure may result in subdural hematoma after cupping therapy [11]. A case of burn injury was reported due to the improper application during cupping therapy [12]. These adverse events associated with cupping therapy raise attention to further investigate the mechanical mechanism of action of cupping therapy. Given the mixed outcomes and potential complications, further research is needed to elucidate both the therapeutic benefits and the potential adverse effects of cupping therapy.

Microcirculation has been reported to be regulated under both the central and peripheral mechanisms in healthy individuals [13, 14]. It is unknown whether cupping therapy could disrupt or enhance physiological synchronization of microvascular networks treated by cupping therapy [15]. Researchers have demonstrated that wavelet phase coherence (WPC) could quantify coherence between certain frequency bands of blood flow oscillations [16]. Tankanag and colleagues investigated how controlled respiration modifies cardiovascular oscillations by using WPC to assess coherence between HR variability and peripheral blood flow oscillations, and attributed microvascular regulation to sympathetic nerve activity [17, 18]. These studies highlight the need to examine cupping therapy's impact on the physiological synchronization of microvascular networks by WPC. Therefore, to better understand the effect of cupping therapy on managing musculoskeletal impairment, local hemodynamic responses have been proposed as a potential mediating factor for promoting musculoskeletal healing after cupping therapy [19, 20]. Exploration of the interaction between different physiological controls may be an essential process needed for effective cupping therapy [21, 22]. WPC is a signal processing method to examine potential interactions between two hemodynamic responses [21, 22], and could be used to analyze synchronized blood flow oscillations after cupping therapy [23].

Negative pressure improves local microcirculation in the treated muscle and skin under both single‐cup [9] and multi‐cup interventions [24], and this increase in local blood flow can be attributed to physiological controls of the microvascular system in the muscle [13]. Although there is evidence on the physiological mechanism of the single‐cup condition, there is usually lack of discussion about the multi‐cup condition in the literature, which is commonly seen in clinical practice [25]. Specifically, the cupping effect has been observed beyond the area inside the cup to encompass the surrounding area in a single‐cup intervention [18, 26]. Therefore, it is important to consider how cupping may cause an interaction between different areas of the skin and muscles (e.g., areas inside and outside the cup). NIRS quantifies oxyhemoglobin and deoxyhemoglobin changes in tissues and is widely used in research for assessing muscle oxygen metabolism [27]. Research has demonstrated that cupping therapy improves local oxygen intake and blood microcirculation, lowers deoxyhemoglobin, and increases oxyhemoglobin [28]. NIRS could be an instrumental tool for understanding the effect of cupping therapy on muscle oxygen metabolism and hemodynamics [29]. The use of multichannel NIRS could investigate whether there is an interaction between two cups on microvascular systems of different areas. These interactions may partly contribute to the therapeutic outcomes of cupping therapy.

Therefore, the aim of this study was to assess the interactions between the two‐cup cupping on the muscle hemodynamic response using the WPC analysis of multichannel NIRS signals. The hypotheses of this study were: (a) The two cupping areas shows higher WPC values under the higher magnitude of pressures than the lower magnitude of pressures because the areas under two cups have interactions on each other's hemodynamic response. (b) The cupping area and its adjacent non‐cupping area show lower WPC values under the higher magnitude of pressures than the lower negative pressure because of the pressure difference between the area inside and outside the cup.

2. Method

The study adopted a repeated‐measures design with three visits to investigate the main effect of the pressure factor and the location factor, and the interaction effect between the pressure and location factors of cupping therapy. The use of repeated‐measures design allowed each subject to participate in all three protocols for a higher statistical power in this study. The pressure factor included three levels at −75, −225, and −300 mmHg. The selection of these values of negative pressure was based on common combinations in clinical practice [9]. The applied negative pressure usually ranges from −225 to −300 mmHg with a duration from 5 to 10 min [30]. Each participant completed three protocols for three visits including (A) –75 mmHg for 10 min, (B) –225 mmHg for 10 min, and (C) –300 mmHg for 10 min. The counterbalanced design was implemented to minimize the order effect of three cupping protocols. The testing order of three protocols for all participants is shown in Table 1. This study was approved by the Institutional Review Board of the University of Illinois at Urbana‐Champaign (IRB #23334).

TABLE 1.

The specific test orders of three cupping therapy protocols, including (A) –75 mmHg, (B) –225 mmHg, and (C) –300 mmHg.

Participants	The test order of three protocols
HA001	B, A, C
HA002	B, C, A
HA003	C, A, B
HA004	A, C, B
HA005	B, A, C
HA006	A, C, B
HA007	B, C, A
HA008	C, B, A
HA009	A, B, C
HA010	B, A, C
HA011	A, C, B
HA012	C, A, B

Note: The test orders for the 12 participants followed the same principle of counter‐balanced design to minimize the order effect.

2.1. Participants

Eligibility for the study was determined by several criteria. Participants had to be aged 18–60 years, have no visible wounds or scars, exhibit a non‐blanchable response on the skin over the gastrocnemius muscle of the dominant side, and have no history of diagnosed cardiovascular diseases, diabetes, or smoking. All participants gave written consent before participating in this study.

2.2. Instrumentation

The different negative pressures within the cups were generated using an electronic cupping therapy device (P1000‐PCS, Medical Device Manufacturing Facility, CA). This approach is called dry cupping. By configuring the device to a precise negative pressure value, the intensity of negative pressure could be accurately achieved. In this study, cups with an inner diameter of 45 mm and an outer of 53–55 mm were applied [31]. Hemodynamic changes in the gastrocnemius muscle, such as oxyhemoglobin and deoxyhemoglobin levels (all measured in micromolars), were monitored using the fNIRS device (fNIR Imager 1000, fNIR Devices, LLC, Potomac, MD) with the sampling rate at 2 Hz. The fNIRS sensor band featured 10 photodetectors and 4 LED light sources, spaced 2.5 cm apart, facilitating signal acquisition through 16 channels, and each channel has a 1.25‐cm detection depth [32]. Oxyhemoglobin, the oxygenated variant of hemoglobin, exhibits peak absorption between 850 and 900 nm, whereas deoxyhemoglobin, its deoxygenated counterpart, shows peak absorption in the 730–750 nm range [33].

The study recorded the time‐series signals in muscle hemodynamic responses following cupping therapy (post‐cupping–pre‐cupping muscle hemodynamic responses). For the measurement of 5‐min pre‐cupping and 10‐min post‐cupping, the fNIRS probe will be taped to the skin of the lateral head of gastrocnemius of the dominant leg, which was adjusted to align with the curvature of the muscle belly. The oxyhemoglobin (Δ[HbO₂]) deoxyhemoglobin (Δ[Hb]) values were reported during both pre‐cupping and post‐cupping periods. Settings on the spectroscopy device, including brightness and sensitivity, were adjusted for each participant to ensure that the intensities of the infrared light‐emitting diodes remained within an operational range, thus preventing both dark noise and signal saturation. The fNIRS signals of 5‐min pre‐cupping period were used to calculate the relative change of fNIRS signals in the concentration of oxyhemoglobin and deoxy‐hemoglobin. In this study, the channels' number and the area names are shown in Figure 1, which consists of the two cupping areas (II and IV indicates the areas inside the cup, while Area II is on the distal side and Area IV is on the proximal side) and the other non‐cupping areas (I indicates an area outside the cup and adjacent to only one cup and III indicates an area outside the cup and adjacent to two cups).

FIGURE 1.

Photographs of the position of fNIRS channels. The photos show the areas defined in this study including Areas I, II, III, and IV. Areas I and II are on the distal side, while the other two areas are on the proximal side. Areas I and III are outside the cup, while the other two areas are inside the cup. The photo shows the experimental setup of a participant.

2.3. Experimental Procedures

All experiments were performed in the Rehabilitation Engineering Lab in the Disability Resources and Educational Services, University of Illinois at Urbana‐Champaign. Figure 2 shows the diagram of the whole protocol for each participant. First, the participants were asked to complete a demographic form and undergo heart rate (HR) and blood pressure (BP) assessments to determine the eligibility. Second, the participant was in a prone position on a mat table for the intervention and data collection. Then, during the 5‐min pre‐cupping condition, the fNIRS probe was securely taped along the axis of the lateral head of the gastrocnemius, positioned to cover the muscle belly. As shown in Figure 1, the edges of the fNIRS sensor and the intended cup‐rim center were marked on the skin once the baseline data collection was completed. Third, during the 5‐min intervention, one of the three pressures (−75, −225, and −300 mmHg) was tested using two cups placed in the corresponding positions of Channels 13–16 and Channels 5–8 in the fNIRS sensor. Fourth, fNIRS was taped to the marked position for 10 min after the removal of the two cups. The participant repeated the above procedures for the second and third visits for the rest of two negative pressures. The second and third protocols of cupping therapy were applied after a week of the last protocol to prevent the carryover effect.

FIGURE 2.

Study design and experimental protocols for each participant.

2.4. Data Analysis

Wavelet analysis of fNIRS signals was implemented to quantify the responses of associated physiological mechanisms. Wavelet‐based analysis (WA) offers both high time resolution for high‐frequency components and superior frequency resolution for low‐frequency components. WA has been used to assess physiological and pathophysiological properties of microvascular control mechanisms [34, 35]. Its procedures are described briefly as follows. The continuous wavelet transform (CWT) technique enables the intricate conversion of time‐series data from its original time domain into a combined time–frequency domain. This process entails the convolution of the time series, denoted as g(u), with an array of basic functions. These functions, typically non‐orthogonal, are derived from a primary wavelet known as the mother wavelet [13]

$$W_{\left(s,t\right)}=\frac{1}{\sqrt{s}}{\int }_{-\mathrm{\infty }}^{+\mathrm{\infty }}\mathit{\Psi }\left(\frac{u-t}{s}\right)g\left(u\right)du$$ (1)

where $W_{\left(s,t\right)}$ is a wavelet coefficient and $\left(s,t\right)$ are the scaling factor $s$ and the temporal position $t$ on the signal translation. The Morlet mother wavelet Ψ is a complex sinusoid modulated by the Gaussian function with basic frequency ${\omega }_0$:

$$\Psi \left(u\right)={\pi }^{-\frac{1}{4}}·e^{-i{\omega }_{0}u}·e^{-\frac{u^2}{2}}$$ (2)

where ${\omega }_0=2\pi .$

The WPC analysis is an effective way for revealing the phase relations and the degree of phase coherence between two time series. It detects potential linkages by calculating the extent to which two signals maintain phase coherence during an ongoing time‐series process. The WPC value was determined by averaging the amplitude of the instantaneous phase difference in the frequency domain over time [36, 37]. The two instantaneous phases of two signals, which were ${\mathrm{\varphi }}_{1k,n}$ and ${\mathrm{\varphi }}_{2k,n}$, were calculated at each time $t_n$ and $f_k$. The relative phase difference is obtained as

$$\mathrm{\Delta }{\mathrm{\varphi }}_{k,n}={\mathrm{\varphi }}_{2k,n}-{\mathrm{\varphi }}_{1k,n}$$ (3)

The phase coherence function is established by computing the sine and cosine components of the phase differences, which are then time‐averaged across the full duration of the signal [38]

$$\begin{array}{c}{C}_{\mathrm{\varphi }}\left(f_k\right)=\sqrt{{\left(\cos \mathrm{\Delta }{\mathrm{\varphi }}_{k,n}\right)}^2+{\left(\sin \mathrm{\Delta }{\mathrm{\varphi }}_{k,n}\right)}^2}\hfill \end{array}$$ (4)

For the signal analysis, Python 3.10 platform was considered and executed in Pycharm IDE with some basic packages taken like Numpy, Pandas, and Openpyxl. Specifically, the PyWavelets library was utilized, which is dedicated to facilitating wavelet transformations and WPC analysis.

In this study, only three characteristic frequency bands were considered, which were metabolic endothelial (0.0095–0.021 Hz), neurogenic (0.021–0.052 Hz), and myogenic (0.052–0.145 Hz), and the systematic factors including respiration and cardiac controls were not studied because of local effect of cupping therapy. The metabolic band is associated with endothelial activity related to nitric oxide [13]. The neurogenic band is regulated by a combination of neurovascular coupling and partial autonomic control [39]. The myogenic band is believed to originate from the intrinsic myogenic activity of smooth muscle cells in resistance vessels, and this mechanism may be partly influenced by autonomic control [40]. For the WA values, signals from all channels within each cupping area (Channels 5–8 and 13–16) under each pressure condition were subjected to wavelet transformation. For each pressure condition, the average WA across all the eight channels was recorded as the representative WA value for this condition. For the WPC values, the coherence analysis was performed between each signal in one area and each signal in the other area. The mean WPC was then calculated to represent the coherence between the two areas. An example of the WPC between two cupping areas was shown in Figure 3.

FIGURE 3.

An example of the WPC values between two areas.

2.5. Statistical Analysis

All values were expressed as means and standard deviations (SDs). To evaluate the interaction level of adjacent bi‐cups under different cupping intervention (−75, −225, and −300 mmHg), the one‐way repeated‐measure analysis of variance (ANOVA) was used to investigate the difference of the WPC value in each frequency band for both changes of oxyhemoglobin (Δ[HbO₂]) and deoxyhemoglobin (Δ[Hb]). The significance level was set at p < 0.05. All the statistical analyses were implemented in the SPSS (Version 24, Chicago, IL).

3. Results

3.1. Participants

There are 12 participants (six males and six females) recruited into this study. Their characteristics were (mean ± SD): age 22.5 ± 0.9 years, height 168.5 ± 7.2 cm, and weight 60.4 ± 11.0 kg, and the participants were free from any manifestations of hypertension.

3.2. Wavelet Average Amplitude

Figure 4 shows the mean values of Δ[HbO₂] and Δ[Hb] in the wavelet average amplitude after different cupping pressures. For the myogenic control, the −300‐mmHg cupping protocol (0.049 ± 0.026) was significantly higher than the −75‐mmHg cupping protocol (0.030 ± 0.016, p = 0.004) in Δ[HbO₂]. There was no significant difference in the other comparisons of Δ[Hb].

FIGURE 4.

(a) Comparison of wavelet amplitudes of metabolic endothelia, neurogenic, and myogenic for Δ[HbO₂], and (b) comparison of wavelet amplitudes of metabolic endothelia, neurogenic, and myogenic controls for Δ[Hb]. *indicates p < 0.05.

3.3. WPC

Figures 5 and 6 show the WPC values of each pair of Δ[HbO₂] signals between two areas. In Figure 5a,b, there is no coherence between the distal cupping area and the other two non‐cupping area. As shown in Figure 5c, in the metabolic control of Areas IV and I, the WPC values have a significant difference between the −75‐mmHg protocol and the −225‐mmHg protocol. Similarly, in the metabolic control of Areas IV and III, the WPC values have a significant difference between the −75‐mmHg protocol and the −225‐mmHg protocol in Figure 5d. As shown in Figure 6, the coherence comparison of two cupping areas in the metabolic control shows a significant difference in WPC value between the −75‐mmHg protocol and the −300‐mmHg protocol. For the neurogenic control, the WPC values have a significant difference between the −75‐mmHg group and the other two groups.

FIGURE 5.

The WPC value between each inside‐cup area and each outside‐cup area comparison in metabolic endothelia, neurogenic, and myogenic controls for Δ[HbO₂] among three pressures. The coherence values of distal inside‐cup Area II and its two adjacent outside‐cup areas are shown in (a) and (b). The coherence values of proximal inside‐cup Area IV and the other two adjacent outside‐cup areas are shown in (c) and (d). *indicates p < 0.05.

FIGURE 6.

The WPC value between two cupping Areas II and IV. Comparison of WPC values in metabolic endothelia, neurogenic, and myogenic controls for Δ[HbO₂] among three pressures. *p < 0.05.

Figures 7 and 8 show the WPC values of each pair of Δ[Hb] signals between two areas. As shown in Figure 7a,b, in the neurogenic control, the WPC values of Areas II and I have a significant difference between the −75 mmHg group and the other two groups (−225 and −300 mmHg). Similarly, in the neurogenic control, the WPC values of Areas II and III have a significant difference between the low‐pressure group (−75 mmHg) and the high‐pressure group (−225 and −300 mmHg). In the metabolic control, the WPC values have a significant difference between the −75‐mmHg protocol and the −225‐mmHg protocol. In Figure 7c,d, in the neurogenic control, the WPC values of Areas IV and III have a significant difference between the low‐pressure group (−75 mmHg) and the high‐pressure group (−225 and −300 mmHg), and in both metabolic and myogenic controls, the WPC values of Areas IV and III have a significant difference between the −75 and −225 mmHg protocols. As shown in Figure 8, in both metabolic and neurogenic controls, the WPC values of two cupping areas have a significant difference between the −75 and −225 mmHg protocols.

FIGURE 7.

Comparison of WPC values between each inside‐cup area and each outside‐cup area in metabolic endothelia, neurogenic, and myogenic controls for Δ[Hb] among three pressures. The coherence values of distal inside‐cup Area II and its two adjacent outside‐cup areas are shown in (a) and (b). In addition, the coherence values of proximal inside‐cup Area IV and the other two adjacent outside‐cup areas are shown in (c) and (d). * indicates p < 0.05.

FIGURE 8.

The WPC between two cupping Areas II and IV. Comparison of WPC values in metabolic endothelia, neurogenic, and myogenic controls for Δ[Hb] among three pressures. *indicates p < 0.05.

4. Discussion

The findings of this study indicate the following: (1) For the comparison of coherence levels within two cupping areas, the WPC values of both Δ[HbO₂] and Δ[Hb] were significantly higher at −75 mmHg compared to the other two pressure conditions (−225 or −300 mmHg) in both metabolic control (0.0095–0.021 Hz) and neurogenic control (0.021–0.052 Hz). (2) When comparing the interaction level between the cupping area and the adjacent non‐cupping area, the WPC values of Δ[HbO₂] were higher at the proximal cupping position in the metabolic control (0.0095–0.021 Hz). Additionally, the WPC values of Δ[Hb] were significantly higher at lower magnitude of pressures (−75 mmHg) than at higher magnitude of pressures in the neurogenic control (0.021–0.052 Hz). (3) For the WA values, comparison under different cupping pressures showed that the myogenic control (0.052–0.145 Hz) of Δ[HbO₂] was significantly lower at −75 mmHg compared to −300 mmHg. This study revealed that based on the WA result in myogenic control, higher magnitude of pressure increases muscle engagement or disrupts normal regulatory mechanisms. For the coherence at lower magnitude of pressure (−75 mmHg), both metabolic and neurogenic controls show higher coherence, indicating that this pressure level may facilitate better synchronization of hemodynamic regulations between cupping areas. In addition, the higher magnitude of pressures (−225 or −300 mmHg) may cause more chaotic or less organized physiological responses and behavior [41], which would reduce the coherence between cupping areas. To our knowledge, this is the first study to use WPC analysis to explore the relationship between two cupping areas and their adjacent non‐cupping areas on hemodynamic responses.

The WA value of Δ[HbO₂] in myogenic control (0.052–0.145 Hz) was significantly lower at −75 mmHg compared to −300 mmHg. Myogenic control plays a crucial role in local vascular regulation, particularly in skeletal muscle circulation, which involves pressure‐dependent mechanisms that contribute to basal vascular tone, autoregulation of blood flow, and protecting against harmful circulatory effects [42, 43]. This result shows that in the cupping therapy, the −300‐mmHg condition increased the blood flow, especially the Δ[HbO₂] activity, by adjusting the basal vascular tone much more than the other two conditions. This result aligned with the previous study that with higher magnitude of pressure, in which the cupping therapy has been shown to notably enhance oxyhemoglobin levels, increase blood volume, and improve oxygenation [9]. This result demonstrated that the myogenic control may be the main regulation system in muscle blood related activity in cupping intervention.

In terms of signal coherence, chaotic effects are often marked by irregular and unpredictable behavior, which can lead to reduced synchronization between signals [44, 45]. The findings of this study suggest that higher magnitude of pressures (−225 or −300 mmHg) may induce more chaotic, disorganized hemodynamic responses, as evidenced by reduced coherence levels. These disruptions in coupling between metabolic and neurogenic controls may result in irregular oscillations. In contrast, the results also indicate that lower magnitude of pressures (−75 mmHg) foster greater synchronization, reflected by higher coherence values. This supports the concept that physiological systems are better regulated and less influenced by chaotic effects at lower magnitude of pressure conditions.

By comparing the Δ[HbO₂] coherence levels, the WPC values were significantly higher under the −75‐mmHg condition than under the −225‐mmHg condition between the inside‐cup Area IV and the other two outside‐cup areas (I and III) in metabolic control. Within the metabolic control (0.0095–0.021 Hz), endothelial metabolic activity primarily facilitates the delivery of oxygen from the blood to the muscle tissue [46]. This result suggests that when cupping with a relatively low magnitude of pressure, the cupping area and its adjacent non‐cupping area may adjust oxygen consumption in smooth muscle cells more synchronously compared to the other two pressure conditions (−225 and −300 mmHg). This may be due to the direction of oxygen delivery from proximal to distal areas and the characteristics of oxyhemoglobin, which include its active and efficiently binding, transporting, and releasing of oxygen. The regulation of oxygen plays a crucial role in maintaining proper respiratory function in organisms. The blood regulation was more activated under the inside‐cup area than the outside‐cup area, as there is a significant pressure difference, which results in the coherence difference.

Regarding the Δ[HbO₂] coherence levels in two cupping areas under different pressures, the WPC values between two cupping areas showed significant differences under different pressures in metabolic control. Cupping therapy, which applies negative pressure to the skin, increases local blood volume, oxygenation, and oxyhemoglobin concentration compared to non‐cupping conditions [18]. These changes are more pronounced at lower magnitude of pressure (−75 mmHg). This mechanism may explain the results of the present study, as different pressures cause various hemodynamic responses in the cupping area, leading to different coherence responses between two intervention areas. This study may provide new insights into the physiological mechanisms behind various negative pressure selections in cupping therapy.

Similarly, as for the coherence of neurogenic control within inside‐cup areas, the coherence value was significantly higher at −75 mmHg compared to the other two pressure conditions (−225 and −300 mmHg). The neurogenic control mainly represents the neurogenic activity. Under the negative pressure from cupping, nervous activity was activated in microcirculation, leading to the regulation oxygen transportation by oxyhemoglobin. Combined with previous studies [9], this result indicates that cupping therapy has a significant effect only under a threshold negative pressure, as lower coherence values show that the cupping area functions differently from non‐cupping areas. This result shows that the −75‐mmHg neurogenic control mechanisms may be more effective in maintaining or enhancing the synchronization of hemodynamic response patterns between the same regions.

For the comparison of Δ[Hb] coherence levels between each inside‐cup area and each outside‐cup area, the WPC value was significantly higher under the −75‐mmHg condition than under the −225‐mmHg condition when comparing the outside‐cup area (Area III) to each of the two inside‐cup areas (Areas II and IV) in metabolic control. The levels of Δ[Hb] can indicate the amount of oxygen that has been delivered and released into the tissues, which is crucial for cellular respiration and energy production. The endothelial metabolic activity predominantly aids in the transfer of oxygen from the blood to the muscle tissue [46]. Area III, located between two cupping areas, is easily affected by adjacent cupping areas, leading to a similar hemodynamic response under the −75‐mmHg condition. This result suggests that when cupping with a relatively low magnitude of pressure, the cupping area and its adjacent non‐cupping area may adjust oxygen consumption in smooth muscle cells more synchronously compared to the other two pressure conditions (−225 and −300 mmHg).

As for the Δ[Hb] coherence values in the neurogenic control, a significant difference in signal synchronization was shown between each pair of areas under the −75‐mmHg pressure condition and the other two pressure conditions (−225 and −300 mmHg) within each pair in neurogenic control except the inside‐cup area IV and the outside‐cup area I. This result shows that neurogenic control is a major factor in the activation of Δ[Hb] under different cupping pressures. Since Δ[Hb] also reflects the amount of oxygen delivered and transported, this result shows that the cupping pressure significantly affects oxygen release within the cupping area and its adjacent areas. Also, the higher magnitude of pressure (−225 and −300 mmHg) creates a significant physiological environment for the inside‐cup area and outside‐cup area; thus, it responds in a lower coherence level. This result indicates that the neurogenic control may be more effective in maintaining or enhancing the synchronization of hemodynamic response patterns between the inside‐cup region and outside‐cup region. Additionally, the WPC value was significantly higher at −75 mmHg than at −225 mmHg between outside‐cup Area III and inside‐cup Area IV in myogenic controls.

For the comparison of Δ[Hb] coherence levels between the two inside‐cup areas, the WPC value was significantly higher at −75 mmHg than at −225 mmHg in both metabolic and neurogenic controls. This result is similar to the result of Δ[Hb] coherence levels with different pressures, which shows the different coherence response of the oxygen transport under different negative pressure. It suggests that the inside‐cup tissue is inclining to regulate its microcirculation by neurogenic control and metabolic control independently when under the common −225 and −300 mmHg [6, 18].

In summary, a lower magnitude of pressure (−75 mmHg) suggested a more subdued vascular response and reduced vasomotion of the muscle microvasculature. This may affect microvascular regulation and blood flow to meet oxygen demands. Higher magnitudes of pressure (−225 and −300 mmHg) induced chaotic effects, which disturb the normal synchronization between cupping area and the region around the rim of the cup. Additionally, higher magnitudes of pressures led to greater fluctuations in oxyhemoglobin and deoxy‐hemoglobin, reflecting stronger hemodynamic responses.

There are some limitations in this study. First, although we intended to investigate the effect of multiple cups of cupping therapy on muscle hemodynamic responses, we assessed the relationship from two cups. For clinical practice, there may be more cups involved. Further research is needed to investigate the influence of different numbers of cups in cupping therapy. Second, the distance between the two cups used in this study was chosen based on the distance between the sensors of the NIRS sensor pad. It is needed to explore the effect of different distances between two cups on the coherence between hemoglobin and deoxyhemoglobin as well as blood flow controls. Third, the therapeutic effects of cupping are measured in the long term, while this study only evaluated the immediate effect of cupping therapy. In the future study, it is necessary to focus on the long‐term effects of cupping. Lastly, there were multiple comparisons between the four areas of the leg under three levels of negative pressure in this study. Therefore, the statistical power may not be sufficient for these repeated comparisons. Future studies should recruit a larger sample size to validate our results.

5. Conclusion

Cupping therapy has been becoming popular in sports and rehabilitation interventions. Evidence to support the effectiveness of cupping therapy remains insufficient. In this study, we conducted the first study to assess phase coherence of hemoglobin and deoxyhemoglobin oscillations from the areas under two cups of cupping therapy. Our results demonstrated that under various negative pressure levels, oxyhemoglobin coherence levels between the two inside‐cup areas were significantly higher under the −75‐mmHg condition compared to the −300 mmHg condition in the metabolic control, and the WPC value was significantly higher under −75 mmHg than −225 and −300 mmHg in the neurogenic control. For the wavelet amplitude values, the myogenic control in oxyhemoglobin was significantly lower at −75 mmHg compared to −300 mmHg. Our study indicates that the use of WPC analysis of oxyhemoglobin and deoxyhemoglobin measured from multichannel near‐infrared spectroscopy could be used to investigate the mechanism of action of cupping therapy in sport and rehabilitation interventions.

Conflicts of Interest

The authors declare no conflicts of interest.

Data Availability Statement

The data that support the findings of this study are available on request from the corresponding author. The data are not publicly available due to privacy or ethical restrictions.