Comparison of Trunk Muscle Activities during Pressure Feedback Monitoring among Individuals With and Without Chronic Low Back Pain

Background: Pressure biofeedback unit (PBU) is a widely used non-invasive device for the monitoring of transverse abdominals (TA) and multifidus (MF) muscles in patients with low back pain (LBP). There is little research that compared trunk muscle activities with pressure feedback during trunk muscle contraction in a seated position. The aim this study was to compare muscle activities between deep local trunk muscles (TA and MF) and different target pressures (50, 60 and 70mmHg) of PBU in individuals with and without cLBP. Methods: Twenty-two patients with chronic LBP (cLBP) and 24 age matched healthy individuals were recruited. Electromyography (EMG) signals were recorded from the TA and MF muscles while the TA and MF were contracted to achieve PBU pressure value of 50, 60 and 70mmHg in random order. The average EMG amplitude (AEMG) of 3 replicate trials was used in the analysis after normalization to %MVIC. %MVIC is defined as the mean of the three AEMG divided by the AEMG of MVIC. Two-way ANOVA was performed to assess the effects of groups (Healthy and cLBP) and the three different target pressures of PBU. Spearman’s correlation analysis was performed in the cLBP group to determine potential correlations between EMG activity, NPRS and ODI. Results: The %MVIC of the TA and MF in the cLBP group were statistically higher than the control group at each pressure value (P0.05). The slope of the %MVIC-pressure of the cLBP subjects was significantly steeper than the healthy subjects (TA: P=0.01, MF: P0.001). During maximal voluntary isometric contraction (MVIC) of TA and MF, compared with pain-free group, cLBP subjects showed a decrease in EMG amplitudes (P≤0.001). The MVIC of MF was negatively correlated with Numerical Pain Rating Scale (r = -0.48, P=0.024) and Oswestry Disability Index (r = -0.59, P=0.004). Conclusions: The study demonstrated the feasibility of using PBU to assess muscle contraction that corresponds with changes of muscle activity as employed in this study is made up of a three-chamber air-filled pressure cell, a catheter and a sphygmomanometer gauge. The pressure cell of the PBU was made from latex-free rubber material, and the unfolded dimension of the cell was 16.7 × 24.0cm. The sphygmomanometer was calibrated to 2mmHg interval and has a range between 0 to 200mmHg. Movement or change in position causes volume change in the pressure cells that is displayed on the gauge. Prior to data recording, the pressure cell was first inflated to a pressure of 40mmHg (orange band). The valve was then closed to stop air leakage [10]. To ensure accuracy of the PBU measurements, the device was pretested by loading a static weight of 4 kg for 24 hours. The PBU was only considered adequate if the device lost no more than 0.5mmHg during the 24-hour period

Background: Pressure biofeedback unit (PBU) is a widely used non-invasive device for the monitoring of transverse abdominals (TA) and multifidus (MF) muscles in patients with low back pain (LBP). There is little research that compared trunk muscle activities with pressure feedback during trunk muscle contraction in a seated position. The aim this study was to compare muscle activities between deep local trunk muscles (TA and MF) and different target pressures (50, 60 and 70mmHg) of PBU in individuals with and without cLBP.
Methods: Twenty-two patients with chronic LBP (cLBP) and 24 age matched healthy individuals were recruited. Electromyography (EMG) signals were recorded from the TA and MF muscles while the TA and MF were contracted to achieve PBU pressure value of 50, 60 and 70mmHg in random order. The average EMG amplitude (AEMG) of 3 replicate trials was used in the analysis after normalization to %MVIC. %MVIC is defined as the mean of the three AEMG divided by the AEMG of MVIC. Two-way ANOVA was performed to assess the effects of groups (Healthy and cLBP) and the three different target pressures of PBU. Spearman's correlation analysis was performed in the cLBP group to determine potential correlations between EMG activity, NPRS and ODI.
Results: The %MVIC of the TA and MF in the cLBP group were statistically higher than the control group at each pressure value (P 0.05). The slope of the %MVIC-pressure of the cLBP subjects was significantly steeper than the healthy subjects (TA: P=0.01, MF: P 0.001). During maximal voluntary isometric contraction (MVIC) of TA and MF, compared with pain-free group, cLBP subjects showed a decrease in EMG amplitudes (P≤0.001). The MVIC of MF was negatively correlated with Numerical Pain Rating Scale (r = -0.48, P=0.024) and Oswestry Disability Index (r = -0.59, P=0.004).
Conclusions: The study demonstrated the feasibility of using PBU to assess muscle contraction that corresponds with changes of muscle activity as measured by EMG. Clinicians may be able to confer PBU measurements with EMG recordings to estimate the level of muscle activities during MF and TA contraction exercise in patients with cLBP. Background Low back pain (LBP) is a major contributor to global burden of disability [1]. Animal experiment and mathematical model of the spine indicated that sequential injuries and deep muscles weakness resulted in spinal instability [2], which is defined as a significant decrease in the capacity of the spinal stabilizing system to maintain the intervertebral neutral zone within physiological limits [3]. Clinical guidelines suggested that spinal stabilization exercises (SSEs) may be as effective as other physiotherapy treatments in reducing disability and pain [4,5]. The principle of SSEs is to train the cocontraction pattern of the deep local trunk muscles of the transversus abdominis (TA) and the multifidus muscles (MF). These two muscles are widely accepted to be the primary segmental stabilization muscles which activate in advance of body movement to produce equilibrium force that counteracts the perturbation force generated during rapid movements of the arms and legs [6,7].
There is another theory that reflex dynamics, which is accounted for 42% of the total stabilizing trunk stiffness, are necessary in the control of spinal stability [8].
The quantitative measurement of the function and strength of MF and TA plays a crucial role in assessing muscle activation pattern and clinical effectiveness of SSEs. The gold-standard to measure the activity of deep local trunk muscles is by fine-wire electromyography [7,8]. However, factors such as pain, discomfort and risk of infection limit its clinical application as routine outcome measure.
Indirect measurements of MF and TA muscles functions rely on electromyography (EMG), ultrasound imaging or a pressure biofeedback unit (PBU). Surface EMG also has the limitation of cross-talk with other muscles that are in close proximity [8]. High cost and inconvenient hinders the common use of ultrasound imaging in clinical practice [9]. PBU has the potential to be a non-invasive, low-cost and convenient assessment tool to be used in clinical and research setting to estimate TA and MF activation [9]. This device measures the change of pressure (mmHg) that the abdominal wall or low back exerts on the lumbar spine during the voluntarily activation of the TA and MF muscles as they perform dedicated postures [10]. Furthermore, the inter-and intra-examiner reproducibility of PBU in measuring TA muscle activity in people with cLBP [11] and healthy [10] individuals was reported to be excellent.
However, few studies were conducted in seated position. In Australia, 70% of adults sit for more than 8 hours per day [13], and the time is likely to extend due to the increasing use of social media [14].
Prolonged sitting seems to be unavoidable in our modern daily life, work and study. Training the TA and MF muscles with PBU in the correct seated position could increase lumbar stability, which might be one of the measures to reduce the occurrence of LBP [10,15]. To improve the clinical application of PBU and facilitate interpretation of activation of deep local trunk muscles, further investigations are needed to establish the response between activity of trunk muscles and different pressures of PBU in seated position.
This study aimed to answer the research question of whether PBU measurements during MF and TA contraction in a seated position could induce corresponding changes in muscle activities as measured by EMG. The study also investigated if people with cLBP have different muscle activation pattern of TA and MF in a seated position, and whether changes in muscle activation pattern is associated with pain. The study hypothesized that PBU of different pressure induced corresponding changes in EMG measurements. Muscle activation patterns were significantly different between people with cLBP and heathy individuals, and the differences in activation pattern were associated with pain and disability.

Subjects
Subjects were recruited from the local rehabilitation ward and outpatient department of the hospital. The inclusion criteria of cLBP subjects were as follows: 1) experienced pain in the low back region with or without accompanying buttock pain over the past three months, and of sufficient intensity to have limited activities of daily living [11]; 2) pain score range between 3-7 on the Numerical Pain Rating Scale (NPRS) [15]; and 3) able to perform the experiment procedure without symptom aggravation. This was to minimize the variability in the level of pain during testing which may increase the variability of the data. Exclusion criteria were: 1) existence of respiratory, orthopaedic, circulatory or neurological conditions; 2) previous surgery to the abdomen or lower back; 3) female subjects who were pregnant or suffered from dysmenorrhea; 4) epilepsy or had family history of epilepsy. The exclusion criteria of NPRS higher than 7 was due to published literature indicated patients who have NPRS 7 or above were not able to perform maximum contraction [11,15].
As maximum contraction was necessary in this study to obtain the MVIC EMG data, this exclusion criteria was adopted. Age matched healthy individuals with no existing LBP and no LBP in the past year were recruited as control.

Ethics
The study protocol was reviewed and approved by the Human Subjects Ethics Sub-committee of the First Affiliated Hospital, Sun Yat-sen University (Approval number: [2017] C-034). All subjects gave written informed consent. The study was conducted in accordance to the Declaration of Helsinki.

Instruments
The Pressure Biofeedback Unit (PBU) (Chattanooga Group Inc., LLC Vista, California, USA) employed in this study is made up of a three-chamber air-filled pressure cell, a catheter and a sphygmomanometer gauge. The pressure cell of the PBU was made from latex-free rubber material, and the unfolded dimension of the cell was 16.7 × 24.0cm. The sphygmomanometer was calibrated to 2mmHg interval and has a range between 0 to 200mmHg. Movement or change in position causes volume change in the pressure cells that is displayed on the gauge. Prior to data recording, the pressure cell was first inflated to a pressure of 40mmHg (orange band). The valve was then closed to stop air leakage [10]. To ensure accuracy of the PBU measurements, the device was pretested by loading a static weight of 4 kg for 24 hours. The PBU was only considered adequate if the device lost no more than 0.5mmHg during the 24-hour period [16].
Muscle activity was assessed by the gold standard surface EMG (UMI-SE-I sEMG system, Shaoxing United Medical Instruments Co., LTD, China). EMG signals with common mode rejection ratio of 110db, bandwidth of 15-1000Hz and resolution of 0.1μV. Sampling frequency was set at 3000Hz and stored in a computer for offline analysis [15]. To reduce skin impedance, hair was removed from the measurement sites and the skin was deterged with alcohol before electrode placement.
Disposable Ag/AgCl surface electrodes were attached to the concerned muscle. The maximum space between the recording electrodes was 2cm. The locations of the EMG electrodes were determined in accordance to EMG placement guidelines [17] and published studies [6,7,11]. Electrodes to measure TA activity were placed at the center position that was 2cm cephalic to the pubic bone, just lateral to the midline, and parallel to the superior pubic ramus along either side of the course of the underlying muscle fibers [11,17]. For MF, electrodes were placed at the level of the L5 spinous process along the line joining the posterosuperior iliac spine (PSIS) and L1-L2 vertebral interspace [15,17]. To obtain the MVIC data of the TA, subjects were in sit-up styled movements with both legs fixated on the floor, using a belt or manually fixed. The spine then flexed to 30° and maintained the position for 5 seconds to record MVIC [17]. To obtain the MVIC data of MF, the subjects laid in prone position on a standard plinth with the upper limbs positioned overhead. Subjects then lifted their head and upper extremities with maximum effort and maintained the position for 5 seconds to record MVIC [15,17]. The maximal voluntary isometric contraction (MVIC) was measured to determine the level of voluntary contractions during testing. MVIC was repeated 3 times and the highest value was selected for analysis [10,17].
The average EMG amplitude (AEMG) at each pressure level was obtained from 3 repeated trials.
AEMG was used in the data analysis after normalization to %MVIC [10,17], which was calculated for the selected muscles using the formula: [See supplementary files to access the equation]. The slope of the %MVIC-pressure relation was calculated at 70mmHg and 50mmHg pressure and compared between the two groups. The rationale to choose AEMG and %MVIC as variables was that raw EMG readings were not comparable between individuals. By normalizing the MVC, the raw EMG reading was rescaled to a percentage of a reference value that was standardized across all individuals within the study. The standardized MVC values enabled quantitative comparisons of the EMG readings between individuals.

Experimental Procedures
Demographics and background clinical information were first collected. Participants were asked to score the pain intensity over the last 24 hours on the NPRS. The scale range between 0-10, with 0 being no pain and 10 being severe pain [18]. Pain related disability was assessed by the Chinese version of the Oswestry Disability Index (ODI) [19]. The total score of ODI ranges from 0 to 100, with 0 being no disability and 100 being maximum disability [19].
All subjects then received information about the anatomy, biomechanics and functions of the TA and MF muscles. All subjects were asked to fast for at least two hours, avoid performing any type of abdominal exercises, and empty their bladders prior to data collection [9]. The procedures of PBU tests were described in Table 1. They were performed in random order by selecting a single card with three cards marked with 1 (50 mmHg), 2 (60mmHg), or 3 (70mmHg). The examiner provided verbal instructions throughout the test. Subjects were asked to maintain within ±2mmHg from the target pressure [10].
All subjects in the pain-free group selected the muscles on the right side. Subjects in the cLBP group chose the more painful side as target muscles. Three trials were conducted for each target pressure and the averaged values were used for analysis. A resting interval of 30 seconds was provided between each trial to minimize fatigue. Participants were allowed to stand up and move during the interval.
[Please see the supplementary files section to access table 1.]

Statistical analysis
Statistical analysis was conducted in the statistical Package for the Social Sciences version 25.0 software (SPSS Inc., Chicago, IL, USA). All EMG data were expressed as mean and standard deviation.
The Shapiro-Wilk test was used to test the normality distribution of the data. Wilcoxon's signed-rank test was used for non-normal distribution parameters. Independent sample t-test was conducted to compare subject characteristics. Two-way ANOVA was performed to assess the effects of groups (Healthy and cLBP) and the three different target pressures (50, 60 and 70mmHg) of PBU. Spearman's correlation analysis was performed in the cLBP group to determine the correlations between EMG activity, NPRS and ODI. The relationships between EMG signal and pressure values were analyzed using the normalized EMG (%MVIC) and the average EMG (AEMG) of the three trials. A value of P<0.05 was considered to be statistically significant.

Results
Twenty-two right-handed individuals were recruited in cLBP group and 24 age matched healthy individuals were recruited in the control group. The sample characteristics of both cohorts are presented in Table 2. There were no group difference for gender, age, height, weight, BMI and educational level. Two cLBP individuals were unable to maintain 70mmHg of TA and were excluded from the analysis. Twenty cLBP participants were included in the analyses of TA at the pressure value of 70mmHg (Table 3). %MVIC-Pressure

The average EMG amplitudes (AEMG)
During MVIC of TA and MF, cLBP group were significantly less than the healthy group (P≤0.001).
No difference was observed in the AEMG at any target pressure value (P > 0.05). Table 3 illustrates the mean and standard deviation of AEMG and MVIC for the TA and the MF at each target pressure. Correlation between EMG activity, NPRS and ODI.

Discussion
The present study observed that PBU measurement during MF and TA contraction in a seated position induced corresponding changes in EMG muscle activities. Muscle activities of the TA and MF during MVIC were lower in the cLBP group than the control group. This indicated that people with cLBP were unable to generate as much force as healthy individuals. However, muscle activities were significantly higher at each pressure value in the cLBP group than the control group. This indicated a stronger contraction was required to maintain spinal stability in people with cLBP. The slope of %MVIC-pressure of the cLBP group was significantly steeper than the control group. The MVIC of MF was negatively correlated with NPRS and ODI. TA MVIC was not correlated with NPRS or ODI.

MF and TA Muscle Activities in Healthy Controls vs cLBP Subjects
Our results indicated that %MVIC of TA and MF were greater and the slope of %MVIC-pressure were steeper in subjects with cLBP than the control group. For MF, finding of our present study was consistent with previous research published by Ansari et al. [20] which explained the high %MVIC of MF. They suggested that muscle pain could be accompanied by hyperactivity in the back muscles during dynamic conditions, which was known as the pain adaptation model [20]. The significantly reduced MVIC of the MF muscle in cLBP group might also support this phenomenon. However, the mechanism of MVIC reduction is not totally understood, but may be related to pain inhibition which limits the ability to perform maximum muscle contraction. People with cLBP have a greater sensitivity to pain [21]. Thus, we speculate that cLBP alters spontaneous neuronal activity resulting in changes in EMG activity [15]. Additionally, atrophic changes of MF had been confirmed in around 77-80% of LBP cases, especially at the L5-S1 level [22] (the EMG site of MF in our study), which might contribute to lower MVIC in the cLBP group.
TA is an important deep muscle that plays a key role in the dynamic control of the lumbar spine [6,7]. The present study observed lower TA EMG amplitudes during MVIC in the LBP group than the healthy group. Hodge et al. reported a delay in TA muscle contraction relative to the agonist muscle that moved the limbs in people with LBP [6,7]. Gildea et al. [23] reported that under contraction status, the thickness of TA was higher in female dancers with cLBP than those without pain. These data support our finding of low TA muscle activity during MVIC.

Differences in muscle activation pattern
Our results showed that the TA %MVIC was more active than MF %MVIC at 50mmHg and 60mmHg in both groups. These findings are consistent with published studies that investigated the relationship between TA and MF in patients with cLBP [24]. The study reported that patients who had adequate contraction of multifidus were of 4.5 times likely to be able to contract TA. At the PBU pressure of 70mmHg, the cLBP group demonstrated almost equal muscle activity in both TA and MF muscles, whereas in the healthy group the %MVIC of MF had more activity than TA. The potential reason may be related to the fatigue of the multifidus muscle. According to published literature [22,25], sensorial factors influence the recruitment of TA and contributed to MF fatigue. The study by Ramos et al. [26] utilized surface EMG to assess fatigue of MF and PBU to detect activity of TA in patients with LBP.
They reported that patients with LBP had difficulties to depress the abdominal wall at the PBU pressure 70mmHg and higher. MF fatigue was also observed. Another possible reason was that cLBP patients had reduced flexibility and mobility in the frontal, transverse, and sagittal planes of motion [27]. When TA was contracted at 70mmHg, there was limited space in the anatomical position [27] that the low back required to complete the motion.

Correlation between muscle activities, pain and disability
The present study observed a negative correlation between MVIC of MF and NPRS and ODI which was consistent with previous studies [15]. There was no correlation between the MVIC of TA and NPRS or ODI. The potential reason was that TA and MF have different roles in maintaining lumbar stability due to their different anatomical structures, different muscle fiber size, different motor unit control properties [21,22]. Previous studies suggested that the TA was mainly involved in lumbar stability by contractile increase of abdominal pressure [9,11], and MF directly maintain lumbar stability through the thoracolumbar fascia [5,21,28]. Therefore, compared with the TA, MF might be more correlated with NPRS and ODI. Moreover, TA and MF had different neuromuscular and proprioceptive systems, along with varied changes in biomechanical alignment of the spine and developed different models of pain adaptation [29]. However, speculation on the connection between neuromuscular control mechanisms and pain was difficult as little is known about the underlying relationship between brain network and the TA and MF muscle activity [24].

Limitations
First, reliability and validity of the PBU and EMG in measuring MF muscle activity in patients with cLBP were not established. Further studies are needed to assess the reliability and validity of this method for evaluating MF. Second, cLBP may interfere with the person's ability to perform maximum muscle contraction. Individuals may therefore not able to perform their "maximum" ability during MVIC due to pain aggravation. The assessor provided detail instructions and verbal encouragements in the present study as an attempt to minimize the impact. The uneven distribution of gender in the sample population might be also a limitation.

Conclusions
The study demonstrated the feasibility of using PBU to assess muscle contraction which corresponded with changes of muscle activity as measured by EMG.

Ethics approval and consent to participate
The study protocol was reviewed and approved by the Human Subjects Ethics Sub-committee of the First Affiliated Hospital, Sun Yat-sen University (Approval number: [2017] C-034). All subjects gave written informed consent. The study was conducted in accordance to the Declaration of Helsinki.

Supplementary Files
This is a list of supplementary files associated with this preprint. Click to download.