Physical Therapy Korea

The gluteus medius (GM) is commonly referenced as a primary hip abductor and biomechanically stabilizes the pelvis in unilateral weight-bearing against the effects of gravity and stabilizes the pelvis during gait as a pelvic rotator [1-3]. Additionally, it has been previously reported that the GM might be an important dynamic stabilizer of the hip and pelvis during activities that can cause a sudden loss of balance and potential injury [4]. Therefore, GM activity has been the focus of many therapeutic exercise protocols for its prevention and rehabilitation.

There has been controversy over whether GM is primarily activated during hip external rotation or internal rotation [5,6] because most studies examining GM activity used only one or two electrodes [5-8]. It may be inappropriate to extrapolate the activity of one or two subdivisions of the GM to the muscle as a whole because of the functional subdivisions within each muscle [2,9-11]. Previous studies have suggested that the GM consists of distinct functional subdivisions (anterior, GMa; middle, GMm; posterior, GMp) according to the alignment of the muscle fibers [2,6,10]. These studies demonstrated that the GMa and GMm are hip abductors and internal rotators [2,6], but the GMm primarily functions as a hip abductor and is not highly influenced by hip rotation forces [6]. They also reported that the GMp is an external rotator and is relatively inactive during hip abduction [2,10]. Few studies have examined muscle activities in all three GM subdivisions during non-functional and functional tasks, and these studies showed significant differences in activities between each subdivision [2,9-11].

Closed kinetic exercises are similar to many functional movements and have been suggested for rehabilitation exercises because they result in a reduced shear force on the knee joint compared with open kinetic exercises [12]. In addition, closed kinetic exercises demonstrated significantly greater GM activity than open kinetic exercises [13]. As a closed kinetic exercise, the pelvic drop exercise (PD) is often used in rehabilitation sessions because of its ease of application [14] and requires the ability of the GM as a hip abductor to control hip adduction and abduction eccentrically and concentrically [15].

When the GM is weak during PD, it is believed that there is synergistic dominance of tensor fasciae latae (TFL) and contralateral quadratus lumborum (QL) [14]. The dominance of the TFL, which is a synergist with the GM as a hip abductor and also hip internal rotator, can induce excessive hip internal rotation and lateral patellar displacement, and these have been associated with iliotibial band friction syndrome and patellofemoral pain syndrome [7]. Moreover, the dominance of the contralateral QL can occur as a synergist with the GM to prevent pelvic drop, and an imbalance between the GM and QL or a decrease in the GM/QL activity ratio can induce movement impairment and low back pain [7].

Minimizing synergistic dominance and selective contraction of weak muscles are the major components in the development of therapeutic exercises [16]. Piran et al. [7] concluded that posteriorly directed force on the pelvis of the non-weight bearing side using a tensiometer cable to induce external rotation forces on the weight-bearing leg during PD can be suggested as an effective intervention to increase GM activity while minimizing the TFL and contralateral QL activity in the healthy group and genu valgum deformity group. They demonstrated that increased GM activity and GM/TFL and GM/QL activity ratios were the results of GM function as a hip stabilizer that maintains the transverse-plane position of the pelvis during the application of external rotation forces on the weight-bearing leg [7].

Meanwhile, the tensiometer cable used in the previous study can be alternated with thera-band [6]. Thera-band can be used easily in worksite training, rehabilitation in hospitals, or in-home use because of its affordability, availability, and portability [17]. However, it is unclear how isometric hip extension or flexion of the non-weight bearing leg induced by anteriorly or posteriorly directed force, respectively, using a thera-band at the ankle influences the activities of the three subdivisions of the GM and synergists with the GM during PD in patients with GM weakness.

Therefore, the purpose of this study was to determine whether isometric hip extension or flexion of the non-weight bearing leg using thera-band at the ankle can influence 1) the activities of GMa, GMm, GMp, TFL, and contralateral QL; and 2) the activity ratios of GM/TFL and GM/QL during PD in patients with GM weakness. We hypothesized that there would be differences among PD, PD with an isometric hip extension of the non-weight bearing leg (PDE), PD with an isometric hip flexion of the non-weight bearing leg (PDF) on the activities of GMa, GMm, GMp, TFL, and contralateral QL, as well as the activity ratios of GM/TFL and GM/QL during PD in patients with GM weakness.

1. Participants

The sample size was calculated using G-power software (ver. 3.1; Heinrich-Heine-Universit^ät D^üsseldorf, Germany). The required sample size of six participants was calculated from data obtained from a pilot study of five participants that measured the same variables in the same conditions to achieve a power of 0.80, an effect size of 0.61 (calculated by partial η2 of 0.27 from the pilot study), and an α level of 0.05. This study included 23 participants (Table 1).

Table 1 . Subject characteristics (N = 23).

Variable
Sex (male/female)	23/0
Age (y)	23.5 ± 2.4
Height (cm)	170.0 ± 4.7
Weight (kg)	74.0 ± 11.2
Body mass index (kg/m2)	24.2 ± 3.3

Values are presented as number or mean ± standard deviation..

The inclusion criteria were GM weakness, no history of surgery on the lower limb, no current knee pain [13], no balance deficits [18], and the ability to perform three consecutive repetitions of PD. The GM weakness was confirmed by performing a manual muscle testing. To confirm GM weakness, the participants were positioned in a side-lying position with the test leg uppermost on the treatment table. The test leg was aligned with the rest of the trunk and the hip of the test leg was abducted to 50% of the full range of hip abduction. The contralateral leg was flexed to ensure the participant’s stability. The principal investigator (PI) applied resistance downward 10 cm proximal to the lateral femoral epicondyle, and an isometric hold was performed twice for 5 seconds against resistance [13,19]. The PI provided verbal cues to encourage maximal performance and instructions to avoid any compensations, such as hip hiking through recruitment of the QL and internal rotation or flexion of the hip through recruitment of the TFL [19]. Participants took a 3-minute rest between the two trials [13]. Strength was graded as 0, 1, 2, 3, 4, or 5/5, then grouped as either ‘weak’ (3/5 or less) or ‘strong’ (4 or 5/5) based on the method described by Kendall et al. [20]. The four grade indicated the ability to hold against gravity plus moderate resistance [20]. The participants who failed to hold their leg abducted 50% of the full range of hip abduction against gravity plus moderate resistance were classified as ‘weak GM’ group, and only the ‘weak GM’ group participated in this study. To assess balance, participants were able to perform single-leg balance with their symptomatic leg on even ground with eyes open for 30 seconds [18].

The exclusion criteria included any past or present musculoskeletal disorder of the lower limbs, neurological or cardiopulmonary diseases, complaints of pain during any physical activity that could restrict the performance of PD, and obesity or status as overweight (body mass index > 25 kg/m2); as fatty tissue acts as a low-pass filter for electrical signals [21]. In addition, participants positive for the Trendelenburg sign were excluded. To check for the Trendelenburg sign, the PI stood behind the participant while visually observing and palpating the iliac crests, and the participant was instructed to lift one foot off the ground by flexing the hip. The sign was considered negative when the participant was able to maintain the pelvis in neutral or with elevating the pelvis of the non-weight bearing side, and positive when the participant was unable to maintain the pelvis level or shift the trunk to maintain the pelvis level [22]. Before the study, participants read and signed a written consent form.

2. Instrumentation

Surface electromyography (SEMG) was used to measure the muscle activities of GMa, GMm, GMp, TFL, and QL. SEMG data were collected using DTS EMG 542 sensors and a Tele-Myo DTS Belt receiver system (Noraxon, Inc., Scottsdale, AZ, USA). The sensor features were a sampling rate of 1,500 Hz, an overall gain of 500, common mode rejection (CMRR) > 100 dB, and input impedance > 100 Mohm. The Myo-Research Master’s Edition (ver. 3.16; Noraxon, Inc., Scottsdale, AZ, USA) was used to analyze SEMG data. A digital bandpass filter between 10 and 450 Hz was used to filter the raw signals, and a notch filter was used to reject 60 Hz. The root-mean-square values were calculated using a moving window of 200 ms before recording.

3. Procedures

Before the experiment, the participants underwent a familiarization period of approximately 20 minutes to achieve proper exercise performance capability. During the familiarization and exercise periods, the participants were instructed to avoid any compensatory movements, such as trunk lateral flexion over the weight-bearing side, the elevation of the contralateral pelvis, and rotation of the pelvis through verbal feedback [23]. Nonetheless, when the participants failed to perform or maintain the standardized position after the familiarization period or during the exercise period, data collection was stopped.

Each exercise was initiated at the command, “ready, go.” Subsequently, the participants consecutively performed a set of three repetitions of PD, PDE, and PDF. The participants performed PD in advance, and the other two exercises (PDE and PDF) were performed in a randomized order by drawing lots to reduce carryover or learning effects. The participants had a 3-minute rest period between each exercise to minimize muscle fatigue [24]. The Metronome Beats application (Metronome Beats, Stonekick, London, England) set at 60 beats per minute (bpm) was used to control exercise speed such that the descending and ascending phases of PD took 3 seconds each. The participant was instructed to keep pace with the metronome during the familiarization period.

PD^. The participants stood barefoot on a step with a leg that was confirmed as having GM weakness while maintaining the head and neck straight, and the opposite leg had no support. The participants were allowed to lightly touch the wall with one hand to maintain balance when necessary, but not to hold onto the wall [18]. The center of the heel and the second toe of the symptomatic leg were placed on a guide tape parallel to the wall. Participants were instructed to maintain touching of the target bar, which was placed horizontally and perpendicular to the wall with both anterior superior iliac spines before the exercise period, and not to move their bodies away from the target bar during the exercise period [7]. The participants started PD from the position where the feet were even [14]. While maintaining extension of both knees, participants were asked to lower the non-weight bearing leg toward the floor until there were no compensatory motions and then return the foot until the feet were even [10,14].

PDE. The participants performed the PDE in the same way as PD, except for the isometric hip extension of the non-weight bearing leg using a thera-band at the ankle. The participants were instructed to maintain contact with the target bar, which was placed at a hip extension of 0° with the calf of the non-weight bearing leg during the PDE (Figure 1). The tension in the thera-band was controlled by the color and length of the thera-band and determined when the participant was able to perform no more than ten repetitions of hip extension 0° of the non-weight bearing leg in a one-leg standing position without any compensation using the thera-band while maintaining consistent metronome speed [25].

Figure 1. Pelvic drop exercise (PD) with an isometric hip extension of the non-weight bearing leg (PDE) inducing isometric hip external rotation of weight bearing leg.

PDF. The participants performed the PDF in the same way as PD, except for isometric hip flexion of the non-weight bearing leg using a thera-band at the ankle. The participants were instructed to maintain contact with the target bar, which was placed at a hip flexion of 0° with the shank of the non-weight bearing leg during the PDF (Figure 2). The tension in the thera-band was controlled by the color and length of the thera-band and determined when the participant was able to perform no more than 10 repetitions of hip flexion 0° of the non-weight bearing leg in a one-leg standing position without any compensation using the thera-band while maintaining consistent metronome speed [25].

Figure 2. Pelvic drop exercise (PD) with an isometric hip flexion of the non-weight bearing leg (PDF) inducing isometric hip internal rotation of weight bearing leg.

EMG data collection. The electrode sites were prepared by shaving any hair from the immediate vicinity of the muscle belly, cleaning the skin with isopropyl alcohol, applying a sterile gauze pad to reduce EMG signal impedance, and allowing for proper electrode fixation. For the GMa, electrodes were placed at 50% of the distance between the anterior superior iliac spine (ASIS) and greater trochanter [10]. For the GMm, electrodes were placed 50% of the distance between the greater trochanter and the iliac crest, similar to previous research [10]. For GMp, electrodes were placed at 33% of the distance between the posterior ilium and the greater trochanter [10]. For the TFL, the electrodes were placed 2 cm inferior and slightly lateral to the anterosuperior iliac spine [26]. For the QL, electrodes were placed in the QL muscle contralateral to the weight-bearing leg, 4 cm lateral from the vertebral groove of the muscle belly of the erector spinae, at a slightly oblique angle at half the distance between the 12th rib and the iliac crest [26].

The entire PD period was analyzed with no differentiation among the descending, holding, and ascending phases because patients normally complete these three phases together as part of their rehabilitation program [10]. A series of three repetitions were recorded for each exercise, and an average of three repetitions was calculated and used for data analysis. To calculate the activity ratio of GM/TFL, the muscle activity of GMp was used because its function opposes that of the TFL in the transverse plane [13]. To calculate the activity ratio of GM/QL, the muscle activity of GMm was used, because most previous studies that examined the activity ratio of GM/QL during PD used the same electrode placement for the GM as GMm in the present study [7,14,27].

Maximal voluntary isometric contraction (MVIC) in the standard manual muscle test position was used to normalize the GMa, GMm, GMp, TFL, and QL. Many previous studies have simply used abduction as a suitable action to determine the MVIC of the GM [13,18]. However, since the GM acts to rotate as well as abduct, it was decided to also assess EMG activity during maximal isometric internal and external rotation [10]. Hip abduction was tested on the side lying on the treatment table with the test leg uppermost, the contralateral hip flexed at 45°, and the knee flexed at 90° to secure the participant’s stability. The test leg was abducted to 50% of the full hip abduction range of motion, with slight hip extension and external rotation. The investigator provided downward force to the ankle while maintaining the participant’s hip position with another investigator’s hand [20]. Internal and external hip rotations were tested in the prone position with the hip in neutral rotation and the knee flexed to 90°. Resistance was applied 2 cm superior to the lateral malleolus during internal/external rotation [10]. The participants performed MVIC twice in each direction for GM. The standard method for MVIC described by Kendall et al. [20] was used to normalize the EMG signal amplitudes for the TFL and QL. To obtain the MVIC value for the TFL, the participants were assumed to be in a supine position on the treatment table with the hip of the test leg flexed and slightly medially rotated with the knee extended. The investigator provided a downward force to the ankle. To obtain the MVIC value for the QL, the participants were assumed to be in a side-lying position, with the knees extended, in a neutral position of the hips with the upper limbs crossed at the chest, and the hands on the contralateral shoulder. An adjustable strap was positioned around the ankle to prevent abduction, and the investigator provided the force to resist the lateral flexion of the trunk at the shoulder. The participants performed MVIC twice for the TFL and QL. The participants had a three-minute rest between each repetition and between the tested muscles [28]. The participants were instructed to gradually increase their muscular contraction over 2 seconds against the resistance until they were at their maximal effort and then maintain that effort for 5 seconds [29]. The middle 3-second contraction, excluding each 1-second at the beginning and end, was used for data analysis. The mean value of two repetitions for the maximal contraction in each GM subdivision was calculated, and the highest mean value from any hip contraction direction was used for MVIC for each GM subdivision. The mean value of two repetitions for the maximal contraction in the TFL and QL was taken as the MVIC. Normalized muscle activity was expressed as a percentage of MVIC (%MVIC).

4. Statistical Analysis

IBM SPSS Statistics for Windows, version 24.0 (IBM Co., Armonk, NY, USA) was used to perform all statistical analyses. Kolmogorov-Smirnov Z-tests were performed to assess the normality of the distribution. One-way repeated-measures analysis of variance was used to assess the statistical significance of the activities of GMa, GMm, GMp, TFL, and QL, and the activity ratios of GMp/TFL and GMm/QL during PD, PDE, and PDF. Statistical significance was set at 0.05. If a significant difference was found, Bonferroni correction was performed to avoid type I errors (α = 0.05/3 = 0.017).

There were significant differences in GMa activity (F = 15.221, p < 0.001), GMm activity (F = 9.451, p < 0.001), GMp activity (F = 8.181, p = 0.001), TFL activity (F = 22.963, p < 0.001), QL activity (F = 6.270, p = 0.004) (Table 2, Figure 3) and GMp/TFL activity ratio (F = 8.689, p = 0.001) among PD, PDE, and PDF (Table 3, Figure 4). However, there were no significant differences in the GMm/QL activity ratio (F = 1.743, p = 0.187) among the PD, PDE, and PDF groups (Table 3, Figure 4). GMa, GMm, and GMp were significantly greater during PDF than during PD and PDE (p < 0.001, p = 0.001; p = 0.001, p = 0.005; p = 0.004, p = 0.004; respectively) (Table 2, Figure 3). The TFL activity was significantly greater during PDE than during PD and PDF (p < 0.001, p < 0.001, respectively) (Table 2, Figure 3). The QL activity was significantly greater during PDF than during PD (p = 0.003) (Table 2, Figure 3). The GMp/TFL activity ratio was significantly lower during PDE than during PD and PDF (p = 0.001, p = 0.001, respectively) (Table 3, Figure 4).

Table 2 . Comparison of muscle activity (%MVIC) in the GMa, GMm, GMp, TFL, and QL among PD, PDE, and PDF.

Variable	Muscle activity (%MVIC)				95% CI
	PD	PDE	PDF		PD	PDE	PDF
GMa	19.2 ± 10.6	18.5 ± 8.0	28.9 ± 16.8		14.6–23.9	15.0–22.0	21.6–36.1
GMm	21.1 ± 10.1	21.7 ± 11.4	26.7 ± 10.7		16.8–25.5	16.8–26.6	22.1–31.4
GMp	16.8 ± 7.2	17.6 ± 7.4	21.7 ± 7.9		13.7–20.0	14.5–20.8	18.3–25.1
TFL	11.3 ± 5.3	22.1 ± 9.1	13.6 ± 6.3		9.0–13.6	18.1–26.0	10.9–16.3
QL	28.9 ± 18.6	35.4 ± 23.1	40.6 ± 24.2		20.8–36.9	25.4–45.8	30.2–51.1

Values are presented as mean ± standard deviation or number. %MVIC, %maximum voluntary isometric contraction; GMa, anterior subdivision of gluteus medius; GMm, middle subdivision of gluteus medius; GMp, posterior subdivision of gluteus medius; TFL, tensor fasciae latae; QL, quadratus lumborum; PD, pelvic drop exercise; PDE, pelvic drop exercise with an isometric hip extension of non-weight bearing leg; PDF, pelvic drop exercise with an isometric hip flexion of non-weight bearing leg; CI, confidence intervals..

Table 3 . Comparison of muscle activity ratio in the GMp/TFL and GMm/QL among PD, PDE, and PDF.

Variable	Muscle activity ratio				95% CI
	PD	PDE	PDF		PD	PDE	PDF
GMp/TFL	1.9 ± 1.3	0.9 ± 0.5	2.1 ± 1.5		1.3–2.5	0.7–1.1	1.4–2.7
GMm/QL	1.1 ± 1.0	0.9 ± 0.8	0.9 ± 0.7		0.7–1.6	0.6–1.2	0.6–1.2

Values are presented as mean ± standard deviation or number. GMp, posterior subdivision of gluteus medius; TFL, tensor fasciae latae; GMm, middle subdivision of gluteus medius; QL, quadratus lumborum; PD, pelvic drop exercise; PDE, pelvic drop exercise with an isometric hip extension of non-weight bearing leg; PDF, pelvic drop exercise with an isometric hip flexion of non-weight bearing leg; CI, confidence intervals..

Figure 3. Comparison of muscle activity (%MVIC) in the (A) GMa, (B) GMm, (C) GMp, (D) TFL, and (E) QL among PD, PDE, and PDF. %MVIC, %maximum voluntary isometric contraction; GMa, anterior subdivision of gluteus medius; GMm, middle subdivision of gluteus medius; GMp, posterior subdivision of gluteus medius; TFL, tensor fasciae latae; QL, quadratus lumborum; PD, pelvic drop exercise; PDE, pelvic drop exercise with an isometric hip extension of non-weight bearing leg; PDF, pelvic drop exercise with an isometric hip flexion of non-weight bearing leg. *By Bonferroni correction, p < 0.017.

Figure 4. Comparison of muscle activity ratio in the (A) GMp/TFL and (B) GMm/QL among PD, PDE, and PDF. GMp, posterior subdivision of gluteus medius; TFL, tensor fasciae latae; GMm, middle subdivision of gluteus medius; QL, quadratus lumborum; PD, pelvic drop exercise; PDE, pelvic drop exercise with an isometric hip extension of non-weight bearing leg; PDF, pelvic drop exercise with an isometric hip flexion of non-weight bearing leg. *By Bonferroni correction, p < 0.017.

This study aimed to determine whether isometric hip extension or flexion of the non-weight bearing leg using a thera-band at the ankle influences the activities of GMa, GMm, GMp, TFL, and contralateral QL, and the activity ratios of GMp/TFL and GMm/QL during PD in participants with GM weakness. We hypothesized that there would be differences among PD, PDE, and PDF in the activities of GMa, GMm, GMp, TFL, and contralateral QL, as well as the activity ratios of GM/TFL and GM/QL during PD in participants with GM weakness. The results of this study support this hypothesis.

The GMa and GMm activities were significantly greater during PDF than during PD or PDE. This finding is consistent with those reported in previous studies. Earl et al. [6] reported significantly greater GMa and GMm activities in isometric hip abduction with internal rotation than in isometric hip abduction and hip abduction with external rotation during single-leg stance. Furthermore, Piran et al. [7] and Schmitz et al. [8] reported that isometric hip internal rotation increased GM activity in response to hip external rotation forces on the weight-bearing leg during single-leg stance and PD, respectively, in healthy participants. Piran et al. [7] and Schmitz et al. [8] placed their GM electrode the same as that of GMm in this study, and their findings are in agreement with our study. The results of the current study can be explained by the functions of GMa and GMm. The main roles of GMa and GMm are hip abduction and internal rotation because of the alignment of the muscle fibers [2,6]. In the current study, isometric hip flexion of the non-weight bearing leg during PDF induced hip external rotation forces on the weight-bearing leg because posteriorly directed thera-band tends to move the iliac crest of the non-weight bearing leg backward in the transverse plane compared with the weight-bearing leg [7,8,11]. In other words, the hip internal rotation force in the muscle of the weight-bearing leg is required to maintain the transverse position of the pelvis during PDF [11]. In contrast, the isometric hip extension of the non-weight bearing leg during PDE induced hip internal rotation forces on the weight-bearing leg because the anteriorly directed thera-band tends to move the iliac crest of the non-weight bearing leg forward in the transverse plane compared with the weight-bearing leg [11]. In other words, the hip external rotation force in the muscle of the weight-bearing leg is required to maintain the transverse position of the pelvis during PDE [11]. Thus, GMa and GMm activities were significantly greater during PDF than during PD and PDE to perform hip abduction with isometric hip internal rotation of the weight-bearing leg.

GMp activity was significantly greater during PDF than during PD or PDE. However, GMp is known as a hip external rotator because of the greater horizontal alignment of the muscle fibers [10]. Lewis et al. [5] also reported that GMp activity increased as a hip external rotator to counteract hip internal rotation force during resisted side stepping. Unlike previous studies [5,10], our study investigated GMp activity during closed kinetic conditions in participants with GM weakness. Thus, the results of GMp activity during PDF in this study could be explained by the concept of synergistic dominance as a stabilizer rather than the function of GMp, whether it is an internal or external rotator [10]. GM weakness was confirmed by manual muscle testing for hip abduction. This method might be more influenced by the strength of GMa and GMm than by that of GMp because GMa and GMm are hip abductors [2,6] but GMp is relatively inactive during hip abduction [10]. In other words, this method could confirm GMa and GMm weaknesses but not GMp weaknesses. The subdivisions of the GM work together synergistically [10]. Thus, GMp might be synergistically dominant during PDF as a pelvic stabilizer to maintain the transverse position of the pelvis in the participants who were included in this study [10]. In contrast, during PDE, although we did not measure SEMG data of the gluteus maximus and the deep hip external rotator, they might be synergistically dominant and have more influence on maintaining the transverse position of the pelvis than GMp because they are primary hip external rotator [30]. Meanwhile, there has been controversy over whether GMp is primarily activated during internal or external rotation in previous studies [5,9]. Contrary to anatomically based observations of GMp, O'Dwyer et al. [9] reported that hip internal rotation in open kinetic conditions resulted in significantly higher GMp activity than hip external rotation in healthy participants. They demonstrated that GMp does not necessarily activate anatomically based observations during hip external rotation under open kinetic condition [9]. Thus, they concluded that excessive hip adduction and internal rotation resulting from weakness of the hip abductor and external rotator muscles related to lower limb injuries should be considered in other hip muscles, such as the gluteus maximus and the deep hip external rotator, rather than in GMp [9]. Our study sheds some light on these conflicting findings, but future studies are needed to evaluate the activities of all subdivisions of the GM during various non-functional or functional activities not only in healthy participants but also in patients with GM weakness.

The TFL activity was significantly greater during PDE than during PD and PDF. Although the TFL is an internal hip rotator, it has zero horizontal plane leverage when standing upright [30]. Thus, isometric hip internal rotation of the weight-bearing leg during PDF might not influence the TFL activity. TFL also acts to balance and provide stability together with all subdivisions of the GM as a synergist during functional activities [2]. The PDE and PDF require greater muscle activities to control for proper balance or stability of the weight-bearing leg because of the application of the thera-band at the ankle of the non-weight bearing leg [31,32]. If the activity of one of the synergists decreases, the activities of the other muscles of the synergists increase to complete the movement [11,27]. In this study, the TFL needed to be more activated during PDE than PDF to control balance or provide stability because of the relatively lower activities of all GM subdivisions.

The QL activity was significantly greater during PDF than during PD, and although not significant, the QL activity was greater during PDF than during PDE. This finding is consistent with those reported in previous studies [33,34]. Previous studies have found the greater activity of the anterior layer of the QL during isometric trunk rotation in the direction to the side of the recorded muscle than in the other direction while standing [33,34]. In our study, the electrode placement of the QL was the same as that of the anterior layer of the QL in the previous study. The QL was divided into three layers, with fascicles arising from the iliac crest to the 12th rib (anterior layer), from the transverse processes of L1 to L4 to the 12th (middle layer), and from the iliac crest to the transverse processes of L1 to L4 (posterior layer) [35]. The interpretation of QL function is considerably variable because various studies recorded muscle activity of the different layers of QL, and these layers may differ in function [36]. In this study, the force of trunk rotation to the weight-bearing leg side is induced during PDF because the posteriorly directed thera-band tends to move the iliac crest of the non-weight bearing leg backward in the transverse plane compared with the weight-bearing leg while the participant maintained the head and neck straight. In other words, isometric trunk rotation in the direction of the non-weight bearing leg was performed during PDF to avoid moving their bodies away from the target bar with ASIS. Thus, the QL of the non-weight bearing side was more activated during PDF than during PDE or PD.

Previous studies have investigated the design of various therapeutic exercises to increase GM activity and decrease TFL and QL activity to prevent lower extremity injuries, such as iliotibial-band friction syndrome, patellofemoral pain syndrome, movement impairment, and low back pain, during PD induced by synergistic dominance [7,14]. In our study, we also investigated the GMp/TFL and GMm/QL activity ratios to provide clinical information. GMp/TFL activity was significantly lower during PDE than during PD or PDF. This result suggests that PD and PDF may have fewer risks of inducing iliotibial band friction syndrome and patellofemoral pain syndrome than PDE. Meanwhile, there were no significant differences in GMm/QL activity ratios among PD, PDE, and PDF, although QL was significantly greater during PDF than during PD and PDE. This result suggests that PDF may not have more possibilities of inducing lower back pain and other movement impairments than PD and PDE due to QL activity. Thus, PDF can be more effective than PDE and PD in increasing GMa, GMm, GMp activities, and GMp/TFL activity ratio in participants with GM weakness.

Our study had several limitations. First, the cross-sectional design prevented the assessment of long-term effects. Second, generalization is limited because all the participants were males aged 20–27 years. Third, although all safety measures were taken, SEMG always involves a risk of crosstalk between nearby muscles and adjacent muscle subdivisions [26]. The optimal electrode placement locations for GMa, GMm, and GMp are unknown; thus, we selected the electrode placement location based on previous studies [6,9,10]. In a previous study, it was possible that the electrode placement location used may have affected the results and was not optimal [6,9,10]. Fourth, the deep and inferior parts of the GMp lie deep in the gluteus maximus; therefore, it is inaccessible to SEMG [10]. The activity of GMp in this study reflects only the superior part [9,10]. Despite these limitations, this study is the first to evaluate the muscle activity of all three GM subdivisions and synergists of the GM during PD, PDE, and PDF. The results may help clarify existing confusion in the literature and guide both clinical practice and future studies on clinical populations.

This study investigated whether isometric hip extension or flexion of the non-weight bearing leg using a thera-band at the ankle influences the activities of GMa, GMm, GMp, TFL, and contralateral QL, and the activity ratios of GMp/TFL and GMm/QL during PD in patients with GM weakness. During PDF, the activities of GMa, GMm, and GMp and the activity ratio of GMp/TFL were the highest. Thus, PDF may be an effective exercise to increase the activities of all subdivisions of the GM and the GM/TFL activity ratio in patients with GM weakness.

None.

None to declare.

No potential conflict of interest relevant to this article was reported.

Conceptualization: SC, SK, SB, HC. Data curation: SC, SK, SB, HC. Formal analysis: SC, SK, SB, HC. Investigation: SC, SK, SB, HC. Methodology: SC, SK, SB, HC. Project administration: SC, SK, SB, HC. Resources: SC, SK, SB, HC. Software: SC, SK, SB, HC. Supervision: SC, SK, SB, HC. Validation: SC, SK, SB, HC. Visualization: SC, SK, SB, HC. Writing - original drafting: SC, SK, SB, HC. Writing - review and editing: SC, SK, SB, HC.