Rectus abdominis muscle thickness change and activation increase during planks performed on different surfaces

Abstract

Aims

The plank is a common exercise used to evaluate core function. Surface electromyography (sEMG) and ultrasound can be used simultaneously to measure muscle activity. We aimed to compare the %-thickness and %-activation during the plank performed on three surfaces and to determine agreement and relationship between rectus abdominis (RA) %-thickness of a rested tabletop position and %-activation normalized to quiet tabletop position during the plank on three surfaces.

Methods

In this cross-sectional study, ultrasound and sEMG measured RA muscle function during the first 5-s and last 5-s of a plank performed on a table, yoga mat, and fitness ball. A repeated measures ANOVA compared differences in %-thickness change and Friedman’s tests compared differences in %-activation, alpha set a priori p ≤ 0.05. Bland–Altman plots measured agreement between instruments. Spearman’s rho determined relationships between instruments.

Results

There was no difference between %-thickness change across surfaces during the first 5-s or last 5-s, or between %-activation during the last 5-s. The %-activation of the RA during the first 5-s performed on the fitness ball was higher than the table and yoga mat (p < 0.001). Ultrasound and sEMG had weak relationships across all surfaces (ρ = − 0.078 to 0.116).

Conclusion

The first 5-s of the plank performed on the fitness ball requires a greater RA activation. Ultrasound could not detect changes in %-thickness of the RA during the plank which may be influenced by the type of contraction. Comparison between these measurement tools during isometric exercise should be used with caution.

Introduction

Core stability can be defined as the ability to maintain the position of the trunk over the pelvis and withstand the transfer of forces throughout the lumbopelvic-hip complex [1]. Muscles of the core include but are not limited to the rectus abdominis (RA), transverse abdominis, lumbar multifidus, external oblique, gluteus medius, and erector spinae [2, 3]. The RA attaches to multiple joints creating the anterior portion of the core and stabilizes the core through isometric contraction and vertebral flexion [3]. The influence of the RA on core stabilization is assessed in rehabilitation, and laboratory settings [4]. Increasing core strength and endurance can decrease the risk of injury and improve athletic performance [5]. In individuals with chronic symptoms such as low back pain and patellofemoral pain, core strengthening exercises can decrease pain and improve physical function [6, 7].

A common exercise used to evaluate and improve core function is the prone bridge plank (plank). When completed on the floor, as it is frequently in a clinical and fitness setting, the plank targets the anterior abdomen by activating the RA up to 41%-maximum voluntary isometric contraction (MVIC) [8]. Progression of the plank can be beneficial to enhance strength and endurance [4, 8]. Fitness balls (ball) enhance exercise difficulty and also increase RA activation [8]. A 13%-MVIC increase was observed during the plank performed on a ball compared to a table [8]. The endurance-based exercise can be held until individuals are unable to maintain the appropriate position, although in the existing literature, the plank is often designed to be held for a pre-determined time only (e.g., 5-s, 10-s) [8, 9]. The plank can be performed until maximal fatigue [10], or until pain or discomfort is felt, causing discontinuation. Discontinuation occurs when the participant stops an exercise voluntarily due to fatigue, discomfort, pain or motivation [2]. Multiple repetitions of shorter durations may not be fully representative of an individual’s core endurance throughout a plank completed until termination. Different individuals may be able to hold the plank for a wide range of time, which necessitates an understanding for clinicians of the beginning and end of a maximal hold regardless of the total hold time. Implementing different surfaces aids clinicians, researchers, and patients in understanding if termination is occurring due to abdominal fatigue or other causes. Measurements of onset and conclusion of muscle activation, morphology, and periods of fatigue can be measured at given time points during the exercise completed until discontinuation. For example, muscular fatigue causes a decrease of the EMG signal frequency content [11, 12].

Non-invasive surface electromyography (sEMG) and normalization to a quiet resting value are reliable in measuring muscle activation during the plank under various conditions [13]. sEMG provides insight into electrical activity of the abdominal muscles throughout the endurance-focused task by recording activity of the broadcast area directly beneath the sensor. Motion mode or M-mode ultrasound imaging can produce a one-dimensional image of anatomy over time [14]. In comparison to sEMG, ultrasound produces an image of the thickness of the muscle versus the broadcast area of the sensor. Visualization of fascial tissue movement during an exercise through M-mode ultrasound (e.g., %-thickness change) in addition to sEMG (e.g., amplitude) allows for simultaneous measurement of muscle activity in the time domain. The sEMG signal may be influenced by crosstalk of surrounding muscles, whereas ultrasound does not have this limitation. Minimal evidence has used both sEMG and ultrasound during the plank exercise as most plank literature implements sEMG alone [2, 9]. Ultrasound is usually captured at a single time point or as a measure before and after the plank to calculate change in abdominal muscle thickness [15, 16]. Although sEMG and ultrasound measurements are independently reliable during abdominal exercises [17, 18], agreement between these measurements during the plank has yet to be determined. One view suggests that the onset of a contraction measured by M-mode ultrasound is not synonymous with sEMG measures [19]. However, a strong linear relationship between both measurements has also been concluded during the onset of contraction [20]. The relationship between these measurements is influenced by the type of muscle contraction and joint activity. For example, when the external oblique acts as an agonist muscle, the correlation between the ultrasound and sEMG measures increases [21]. The use of ultrasound and sEMG during the plank performed without pre-determined times can provide novel quantifiable measures of muscle. Implementing normalization techniques for both ultrasound and sEMG allows for observation of a visual and electrical representation of a muscle of interest. These measures can help clinicians improve prevention and rehabilitation protocols that utilize ultrasound and sEMG (e.g., biofeedback).

Therefore, the purpose of this study was to compare the %-thickness and %-activation during the plank performed on 3 different surfaces. We hypothesized that the ball would have the highest %-thickness and highest %-activation during the first and last 5-s. The secondary purpose of this study was to determine the agreement and relationship between RA %-thickness of a rested tabletop position and sEMG normalized to quiet tabletop position during the plank performed on a treatment table (table), yoga mat (mat), and ball. We hypothesized that sEMG and ultrasound measurements of the RA during the plank would have poor agreement because previous comparison between these methods did not detect a clear relationship during abdominal bracing, hollowing contractions or isometric contractions of the core [22].

Materials and methods

Study design

This was a cross-sectional study that measured RA %-thickness change and sEMG amplitude simultaneously during the first 5 and last 5-s of the plank to failure performed on a table, mat, and ball by a single assessor (LD). The independent variable was surface with 3 levels (table, mat, and ball). The dependent variables were ultrasound %-thickness change and sEMG amplitude normalized to a quiet resting position of the right RA.

Participants

Forty-four (22 male, 22 female), physically active adults with no previous year history of low back, hip, or core injury were recruited from the local community (Table 1). G*Power (G*Power, Version 3.1.9.4, Dusseldorf, Germany) [23] determined 40 individuals were necessary to detect a mean difference of 10%-MVIC in RA sEMG recruitment (effect size = 0.46) at alpha level = 0.05 [8]. We accounted for 10% attrition due to potential screen fails which increased the total to 44 participants. Participants were included if they self-reported being physically active at least 30 min, 5 days/week [24], and were between 18 and 45 years of age. All participants provided informed consent and the study was approved by the University of Central Florida Institutional Review Board.

Table 1.

Participant demographics and plank hold times (mean ± standard deviation)

Characteristic	(N = 44)
Age (years)	23.27 ± 4.79
Height (cm)	170.28 ± 7.71
Mass (kg)	68.93 ± 10.93
BMI (kg/m2)	23.67 ± 3.07
IPAQ (mets)	5697.17 ± 3647.55

Surface	Hold times (s)
Table	106.53 ± 47.41
Mat	83.18 ± 31.31
Ball	58.30 ± 26.26

N number of participants, cm centimeters, kg kilograms, BMI body mass index, IPAQ International Physical Activity Questionnaire

Procedures

After providing informed consent, participants completed questionnaires regarding general health history and physical activity using the IPAQ [25]. Age, sex, height, and weight were recorded. Delsys Trigno wireless system (Delsys, Inc., Boston, MA, USA) with Trigno Avanti sensors and EMG works® 4.54 (Delsys, Inc., Boston, MA, USA) was used for sEMG acquisition with a sampling frequency of 2000 Hz and a gain of 500. Shaving, debridement, and cleansing of the skin occurred prior to electrode placement. The right RA electrode was placed 2 cm lateral to the umbilicus parallel to the muscle fibers and was confirmed according to previous recommendations [26]. The participant was instructed to lie supine in a hook-lying position on the table for the sEMG quiet rested values. At the conclusion of the sEMG quiet collection, an 8-MHz linear transducer connected to a portable GE LOGIQ Book XP (GE Healthcare, Waukesha, WI, USA) ultrasound unit was placed over the RA while the participant was supine in a hook-lying position (Fig. 1). The RA image was obtained by using anatomical landmarks as previously described [27–29]. The transducer and gel were positioned between the umbilicus and sEMG electrode and was adjusted until the right RA fascial borders were visible on the screen using brightness mode; this image was verified by the investigator before measurements occurred. Resting RA images were recorded with the participant placed in a hook-lying position [28]. The transducer was then secured in the same location with a foam block and elastic belt prior to plank performance [30]. The same investigator (LD) performed all ultrasound imaging and had 2 years of imaging experience. The plank was performed first on a table, then a mat, and finally a ball. The size of the ball used was determined by manufacturer guidelines regarding height. The body positioning for the plank is pictured in Fig. 2. The participant was asked to hold the plank until they could not maintain the appropriate position. Two minutes of rest were provided between each exercise [8]. The same investigator (LD) instructed participants on the correct form and observed each participant’s body positioning during exercise completion.

Fig. 1.

Ultrasound transducer and sEMG electrode placement on the right rectus abdominis

Fig. 2.

Plank positioning and the belt used to secure the ultrasound transducer and muscle thickness measured in centimeters during the first 5-s of the plank

Data processing and statistical analysis

RA sEMG amplitude was collected during each of the 3 planks and was bandpass filtered between 50 and 500 Hz and processed with a root mean square envelope between 0.125- and 0.050-time constants. The software was defaulted to process data over moving windows with 50% overlap between successive windows. The peak amplitude in millivolts of sEMG during the first 5 and last 5-s was normalized to the quiet resting trials and expressed as a % increase (mean, [standard deviation]). %-activation was calculated by dividing the peak amplitude by the resting amplitude. The thickest portion of the RA throughout the first 5 and last 5-s from the M-mode ultrasound was measured in centimeters and calculated as a percentage of the participants’ rested thickness (%-thickness change), (mean, [standard deviation]) (Fig. 2). %-thickness change was calculated by subtracting the rested thickness from the contracted thickness and then dividing by the rested thickness multiplied by 100 [31]. Normality of the data was assessed by skewness and kurtosis. A repeated-measures ANOVA compared differences in %-thickness change, alpha set a priori p ≤ 0.05. Bonferroni post-hoc tests were conducted for significant findings. Friedman’s tests compared differences in %-activation, alpha set a priori p ≤ 0.05. A Wilcoxon signed-rank test was conducted for significant findings.

A Bland–Altman plot was created with the mean difference between the %-activation increase and %-thickness change as well as 95% limits of agreement (LOA) for the first 5 and last 5-s of the 3 planks. Spearman’s rho (ρ) correlation coefficients were used to measure the relationship between %-activation increase and %-thickness change during both the first 5 and last 5-s. SPSS version 22.0 (IBM Corporation, Armonk, NY) was used for all statistical analyses.

Results

The RA %-thickness mean differences during the plank performed on the table, mat, and ball are presented in Table 2. %-thickness change values confirmed the assumption of normality. %-activation change failed the assumption of normality. There was a statistically significant difference in %-activation across surfaces during the first 5-s, X²(2) = 35.318, p < 0.001. Median (IQR) %-activation for the first 5-s of the plank performed on the table, mat, and ball were 7.6% (4.74–16.60), 10.4% (5.67–20.66), and 20.75% (12.8–43.85), respectively. A Wilcoxon signed-rank test showed that the RA %-activation during the first five seconds of the plank on the table (15.01 ± 18.26%), had significantly less activation compared to the plank performed on the ball (29.57 ± 23.96%), (Z = − 4.376, p < 0.001). Also, the first 5-s of the plank on the mat (18.61 ± 26.44%) had significantly lower %-activation compared to the ball (29.57 ± 23.96%), (Z = − 3.898, p < 0.001). The %-activation of the RA during the first 5-s was not significant between the table (15.01 ± 18.26%) and mat (18.61 ± 26.44%) (Z = − 0.969, p = 0.333). There was not a significant difference in %-activation between surfaces during the last 5-s, (table: 42.29 ± 55.94%, mat: 36.09 ± 45.51%, ball: 31.44 ± 27.13%), X²(2) = 0.864, p = 0.649.

Table 2.

%-Thickness mean difference of plank during table, mat, and ball

	Mean difference (%-thickness)	Sig	95% CI
First 5-s
Table vs mat	− 5.15	0.586	− 14.90, 4.60
Table vs ball	− 9.33	0.10	− 20.05, 1.40
Mat vs ball	− 4.18	0.684	− 12.68, 4.33
Last 5-s
Table vs mat	− 9.50	0.20	− 22.10, 3.09
Table vs ball	− 4.80	0.83	− 15.72, 6.10
Mat vs ball	4.70	0.72	− 5.17, 14.57

Bland–Altman plots for the first 5-s on the table, mat, and ball are presented in Figs. 3, 4, and 5, respectively. All measures demonstrated a weak relationship between ultrasound and sEMG [table, first 5-s (ρ = 0.02); last 5-s (ρ = 0.12)]. The mat had a weak relationship between the instruments (first 5-s ρ = 0.09; last 5-s ρ = 0.09), as well as the plank on the ball (ρ = 0.02 and − 0.08).

Fig. 3.

Bland–Altman plots describing the relationship between ultrasound and sEMG during the plank performed on the table

Fig. 4.

Bland–Altman plots describing the relationship between ultrasound and sEMG during the plank performed on the yoga mat

Fig. 5.

Bland–Altman plots describing the relationship between ultrasound and sEMG during the plank performed on the ball

Discussion

The purpose of this study was to compare the %-thickness and %-activation during the plank performed on 3 surfaces. There was not a significant difference in %-thickness change between the different surfaces during the first 5-s or last 5-s. Although the ball did not have the highest %-thickness change during the first and last 5-s on all surfaces, the plank on the ball did have the highest %-activation during the first 5-s. The exercise progression was chosen based off previous evidence evaluating sEMG during a plank performed on the floor and a ball [8]. A mat was implemented for the plank performed on the floor to minimize potential discomfort felt during the exercise [2]. Contrary to predetermined time periods or repetitions, the plank performed in the current study was held until participants could not maintain the proper form [8, 15].

A difference in %-thickness of superficial and deep musculature of the core may not occur during the plank performed until discontinuation due to the type of muscular contraction. The RA produces an isometric contraction throughout the exercise with the origin and insertion of the muscle in a static position, creating stability. An isometric contraction will not provide the same change in muscular thickness as a concentric contraction due to the static muscle fibers. The variance in the %-thickness change during the first and last 5-s of the 3 surfaces was high, suggesting that consistent thickness changes of the RA may not occur during the plank performed on different surfaces. The absence of statistical significance in %-thickness during an isometric contraction contrast with significant findings when measuring muscle thickness during the abdominal draw in maneuver, bridging, roll-up or one leg raise [16].

Our findings support changes in muscle thickness during the plank performed on stable and unstable surfaces are not significantly different [15]. Contrary to our findings, a plank performed with a cushion under participants’ feet resulted in significant differences in transverse abdominis and internal oblique muscle thickness [32]. The participants in the previously mentioned study completed the plank until 3 exhalations occurred, differing with our participants holding the plank until discontinuation. The duration of the plank and the location of the unstable surface may affect the thickness change of the abdominal musculature. Thickness during the first and last 5-s of exercise was measured, but the remaining time between the measurements, although variable among participants (Table 1), may provide insight of a maximal thickness or increased thickness change.

Isometric contraction may not increase thickness during the plank but that type of contraction can influence %-activation during the first 5-s of the plank. We found a significant increase in %-activation on the unstable surface compared to the stable surfaces during the first 5-s of the plank. Our findings are in agreement with previous findings evaluating the raw sEMG activity and % MVIC of the rectus abdominis during a plank performed on stable and unstable surfaces [33]. The planks held in the previous study were completed for 5-s repetitions, which is similar to our first 5-s analysis [33]. The significant difference between the unstable surface (ball) and the stable surfaces (table, mat) may be explained by perturbations created by the unstable surface. As the participant performs the plank on the ball, there is an increased demand to withstand the forces placed on the body, especially as they work to achieve the initial stabilization on the ball. Additionally, the angle of the body during the ball plank was not parallel to the ground as with the table or mat, which may increase the biomechanical challenges placed on the individual.

The secondary purpose of this study was to determine the agreement and relationship between RA %-thickness of a rested tabletop position and sEMG normalized to quiet tabletop position during the plank performed on a table, mat, and ball. The Bland–Altman plot allowed for quantification of %-activation and %-thickness change (Figs. 3, 4, and 5). The rested values for both instruments were measured in supine, whereas the plank was performed prone. The abdominal tissue is suspended and forced inferiorly by gravity during the completion of the exercise. The findings of this study indicate that the 2 measurements are assessing separate aspects of muscle function and should be cautioned when comparing values regarding the plank. Although the tools are measuring separate components of muscle function, they are valuable as stand-alone measurement tools.

This study was not without limitations. The normalization methods, type of muscle contraction, and the role of the muscle (agonist, antagonist) used during an endurance-based exercise do not allow for ultrasound and sEMG values to be used interchangeably [21]. Although ultrasound imaging mitigates some of the limitations of sEMG (e.g., cross-talk), there are shared drawbacks of both methods as they only collect from the area immediately beneath the electrode or transducer. A single peak measure was reported to represent activation and thickness change during the 5-s intervals, which assumes consistent variability of both measures. Future research is needed to evaluate the relationship between ultrasound and sEMG when the RA is performing as an agonist or during concentric muscle contractions, in addition to assessment of other muscles that contribute to core stability.

Conclusions

Differences in RA muscle thickness could not be detected during the plank performed on stable and unstable surfaces with dynamic ultrasound. The type of muscle contraction may limit the ability to measure %-thickness change even if measured on various surfaces of increased difficulty. The type of contraction may not influence %-activation because the electrical activity is being measured, not physical thickness change. The relationship and agreement between %-thickness change and %-activation during the plank are weak. The tools are valuable when used separately and can describe their respective aspects of muscle function.

Data availability

The datasets generated during and/or analysed during the current study are available from the corresponding author on reasonable request.

Declarations

Funding

The authors declare that no funds, grants, or other support were received during the preparation of this manuscript.

Conflict of interest

The authors have no relevant financial or non-financial interests to disclose.

Ethical approval

The study was performed in line with the principles of the Declaration of Helsinki and approved by the Institutional Review Board at the University of Central Florida.

Consent to participate

Informed consent was obtained from all individuals participants included in the study.

Author contributions

LD and LCM contributed to the study conception and design. Material preparations were performed by all authors. Data collection was performed by LD and LCM. Analysis was performed by all authors. The first draft of the manuscript was written by LD and all authors commented on previous versions of the manuscript. All authors read and approved the final manuscript.