Changes in the Intermuscular Coherence of the Multifidus and Its Relationship with Fatigue and Low Back Pain: A Case Series ^†

Abstract

Subjects with low back pain (LBP) have shown different responses to fatigue compared to healthy individuals. The aim of this study was to analyze and compare the connectivity of the multifidus muscles using intermuscular coherence (IMC) in a squat protocol between healthy subjects and a subject with a history of low back pain. The protocol consisted of three sets of squats, with the final set performed with as many repetitions as possible to induce muscular fatigue. The findings indicated that IMC decreased among healthy subjects during the last repetitions of the squat series, with percentage changes in connectivity (CC) ranging from approximately 43.3% to 67.3% when comparing the first and last 10 repetitions. In contrast, the subject with a history of low back pain (LBP) exhibited an opposite trend, showing a 41.8% increase in IMC. These results suggest a different response of IMC to fatigue related to LBP. This also establishes IMC as a potentially useful tool for characterizing fatigue and LBP during physical exercise.

1. Introduction

The functionality of the trunk musculature and the concept of central body stability (known as the core) are closely related. This relationship can be explained as a complex network of interrelationships among the ligamentous, neural, and muscular subsystems of control, particularly in the lumbo-pelvic region [1]. This system plays a fundamental role as a stabilizer of the lumbar spine by controlling position and movement while acting as a transmitter of forces between the limbs, with special attention to the prevention and treatment of low back pain (LBP) [2]. Deodato et al. [3] presented evidence that alterations in mechanical properties, such as reduced size, increased stiffness, and imbalance in contractile trunk muscles, are associated with the presence of LBP in young gymnasts. These alterations could affect the stability and motor control of the lumbar spine, thus contributing to the development and maintenance of low back pain.

LBP encompasses a wide range of pain types and intensities. Anatomically, it is typically located below the last rib and above the iliac crest, though it often coexists with pain in the gluteal area. Several etiological and risk factors for LBP have been identified, and a considerable proportion of individuals with LBP develop chronic or recurrent symptoms. The global burden of LBP is substantial: in the United States alone, treatment costs are estimated at around $100 billion annually, with approximately 75% attributed to productivity losses. These figures highlight the growing need for effective prevention and treatment strategies. Among them, physical exercise has demonstrated efficacy in primary prevention [4].

On the other hand, muscle fatigue is the inevitable consequence of sustained contractions and is generally defined as a reduction in a muscle’s ability to maintain strength or power levels and may be related to ineffective central stabilization [3]. There are some hypotheses that relate the prevalence of nonspecific LBP to mechanical dysfunction in central stability, particularly in relation to fatigue [5]. However, this relationship is still under investigation.

Surface electromyography (EMG) is a non-invasive technique used to assess muscle function in healthy and diseased muscles. Spatial and temporal variability in muscle activation is related to the rate at which fatigue develops. Fatigue and muscle activity variability have been associated with LBP [5]. While some studies have examined changes in the spectral power of electromyographic manifestations due to fatigue [6], more recent research has focused on intermuscular coherence (IMC) as a means to explore the neural control of muscle. In this regard, Ghazi et al. [7] provided evidence for the role of IMC in the beta band frequency spectrum (15 to 30 Hz) in relation to LBP. Recently, Yu et al. [8] reported a different behaviour of IMC between individuals with and without subacromial pain. In addition, there is increasing research on the behaviour of IMC in relation to fatigue [9,10,11].

Despite these findings, no studies to date have directly examined how fatigue modulates intermuscular coherence in trunk stabilizing muscles in individuals with a history of low back pain. The specific influence of fatigue on neural coordination patterns, as measured by IMC, remains poorly characterized in this population.

The hypothesis of the present work is based on the existence of an altered behaviour of the trunk stabilizing muscles under fatigue conditions in subjects with a history of nonspecific LBP, compared to healthy subjects. The aim of the present study was to analyze and compare the behaviour of the lumbar musculature using intermuscular connectivity and its relationship with muscle fatigue in a squat protocol between healthy subjects and a subject with a history of low back pain. This approach seeks to address the current lack of studies exploring how fatigue modulates neural coordination in trunk stabilizers among individuals with LBP.

2. Materials and Methods

2.1. Participants

Five adult male participants, aged between 21 and 28 years, voluntarily took part in this study. All were physically active and had at least two years of continuous strength training experience. Four participants reported no history of low back pain (LBP) in the three months prior to the study, while one participant reported nonspecific LBP during that period. Participants with diagnosed neuromuscular disorders, current musculoskeletal injuries, or pain at the time of the study were excluded to ensure valid neuromuscular assessments. Recruitment was conducted through digital platforms targeting university populations. Prior to participation, all individuals received detailed explanations of the experimental procedures and provided written informed consent.

2.2. Experimental Procedure

The participants performed a 10-min warm-up protocol consisting of joint mobility and pedalling on a stationary bicycle to prepare the body for the following tests. Three series of loaded squats (barbell and plates) were performed. The first two series were used as progressive conditioning for the third series. In the first series, a load equivalent to 40% of body weight (BW) was used. Each participant was instructed to maintain a rhythm for the movement, consisting of 3 s for the descending phase, 1 s in the transition phase, and 3 s for the ascending phase. This rhythm was presented on a timed videotape placed in front of the participant for feedback throughout the performance of the exercise. In the second series, a load corresponding to 60% of the BW was used, with the same rhythm as in the previous series. In the third series, 60% of the BW was again used, but the pacing was changed as follows: 3 s in the eccentric phase (descent), 1 s in the isometric phase (transition), and at the maximum possible voluntary speed in the concentric phase (ascent). In series 1 and 2, 20 repetitions were performed. In series 3, each participant was asked to perform, or voluntary limit.

2.3. Instruments

An RHA2000 acquisition system (Intan, Los Angeles, CA, USA) was used to collect muscle activity. EMG was recorded through the analog channels of the Intan system at a sampling frequency of 25 kHz. Sensors were placed in a bipolar configuration on the right and left multifidus muscles in the lumbar region (Figure 1). Surface preparation and electrode placement were performed according to SENIAM guidelines. A linear position transducer (WinLaborat, Buenos Aires, Argentina) with a sampling frequency of 1 kHz, attached to the bar, was used to capture motion. The position transducer signal was synchronized with the EMG using an auxiliary channel of the Intan system.

2.4. Signal Preprocessing

The analysis was performed offline using Matlab (MathWorks, Natick, MA, USA), version 2020b. The signal from the linear position transducer was processed through classical kinematics formulas to obtain displacement and velocity data. The phases of the exercise (i.e., rest, eccentric, isometric, and concentric) were identified using kinematic data. Peak velocity (PV) was defined as the maximum velocity achieved during the concentric phase. While the EMG signals were acquired at a sampling frequency of 25 kHz, they were downsampled to 1 kHz. The EMG signals underwent filtering using a zero-lag fifth-order Butterworth filter with a bandpass range of 10–100 Hz and a band stop between 49–51 Hz to eliminate noise. For subsequent connectivity analyses, we used the unrectified and non-normalized EMG signal to meet the theoretical demands and uphold the practical justifications for the computation [12].

2.5. Functional Connectivity Calculation

To determine the functional connectivity index between the multifidus muscle pair, the IMC was calculated using a method that has been used in recent years. The method presented by Bigot et al. [12] is a procedure that allows the calculation of the magnitude-squared coherence $R_{xy}^2\left(\omega ,u\right)$ between two time-frequency maps, as shown in Equation (1).

$$R_{xy}^2\left(\omega ,u\right)={\displaystyle \frac{{{|S}_{xy}\left(\omega ,u\right)|}^2}{S_x\left(\omega ,u\right)S_y\left(\omega ,u\right)}},$$

(1)

where $S_{xy}\left(\omega ,u\right)$ is the ‘Mean Cross-Spectrum’; $S_x\left(\omega ,u\right)$ and $S_y\left(\omega ,u\right)$ are the ‘Mean Auto-Spectrums’ of both signals. These maps are obtained using the continuous Morlet wavelet transform. A statistical test is then applied to determine the sectors where the mean cross-spectrum is significant. Finally, IMC is calculated as the mean value within a window of interest. In this work, the IMC was calculated only during the concentric phase of the proposed exercise, since this phase exhibits maximum dynamic stabilisation effort for the muscle pair under study. The window of interest was defined in the beta band (15 to 30 Hz) and with a time window of 500 ms centered on the PV (±250 ms), as depicted in Figure 2.

Two values for IMC were calculated using the first 10 repetitions (IMC_INITIAL) and the last 10 repetitions (IMC_FINAL) of the third series. The first two series were used for familiarization and preparation purposes and were therefore excluded from the analysis. Finally, the percentage change in connectivity (CC) was calculated for each subject according to Equation (2).

$$CC={\displaystyle \frac{{IMC}_{FINAL}-{IMC}_{INITIAL}}{{IMC}_{INITIAL}}}\times 100,$$

(2)

2.6. Statical Analysis

Parametric methods were used to compare the results obtained for the IMC variables at the level of each participant. Prior to the analysis, the normality of the differences between IMC_INITIAL and IMC_FINAL was assessed using the Shapiro–Wilk test. As the assumption of normality was met, a paired samples Student’s t-test was conducted to compare the change in connectivity status during the third series.

3. Results

The study sample comprised five adult male participants aged between 21 and 28 years, with a mean age of approximately 24.8 years. Their heights ranged from 1.70 to 1.80 m, and body weights varied between 70 and 80 kg.

Table 1. IMC values expressed as mean values and CC values expressed as percentages.

Participants	IMCINITIAL	IMCFINAL	DIF IMC	CC (%)
P1	0.03903	0.012742	p < 0.001	−67.3
P2	0.14721	0.070299	p < 0.001	−52.2
P3	0.11941	0.043518	p < 0.001	−63.5
P4	0.12218	0.069214	p < 0.001	−43.3
P5 ×	0.12716	0.180380	p < 0.001	+41.8

× Participant with a history of non-specific low back pain. DIF: differences.

shows the connectivity values calculated for each participant during the third series. A decrease in the magnitude of the connectivity value was observed in the four participants with no history of LBP. Participant 5 showed an opposite pattern.

4. Discussion

The main findings of this study indicate a distinct pattern of intermuscular coherence (IMC) modulation in response to fatigue during a squat protocol. Specifically, four healthy participants without a history of low back pain (LBP) exhibited a significant decrease in IMC values between the initial and final repetitions of the third series, with percentage changes in connectivity (CC) ranging from approximately 43.3% to 67.3%. In contrast, the participant with a history of non-specific LBP demonstrated an opposite response, showing a 41.8% increase in IMC over the same period. These divergent patterns suggest that neuromuscular adaptations to fatigue may differ fundamentally between healthy individuals and those with prior LBP, potentially reflecting altered motor control strategies in the latter group.

Contrasting the results obtained in this study with the available literature, some inconsistencies can be observed in the behaviour of IMC. Dos Santos et al. [9] showed in their study that fatigue tends to increase IMC in synergistic muscles in a group of healthy older adults, suggesting that this behaviour probably reflects an age-specific adaptation. In the same study, the young healthy subjects tested showed no significant fatigue-related changes in IMC. Semmler et al. [10], who studied healthy subjects using isometric contractions, obtained similar results, showing that IMC increases as a function of fatigue. However, in the study by Charissou et al. [11], which employed the same methodology to calculate IMC as used in this study, a slight decrease in fatigue-related ‘beta intermuscular interactions’ was reported in healthy subjects without strength training experience, whereas these interactions remained stable in subjects with strength training experience. These mixed findings suggest that fatigue may have variable effects on IMC depending on age, training status, and task characteristics.

In the present study, one participant—who had a history of low back pain (LBP)—exhibited a different trend compared to the others, showing an increase rather than a decrease in IMC. A recent study by Yu et al. [8] assessed IMC across shoulder muscles during sustained isometric contractions in individuals with and without subacromial pain. They found that asymptomatic participants exhibited increases in delta-, alpha-, and beta-band coherence with fatigue, whereas symptomatic participants failed to increase coherence in several frequency bands. These results suggest that not only fatigue but also pain may be associated with altered neuromuscular control strategies, potentially reflecting a specific adaptation aimed at maintaining motor performance under adverse conditions.

Moreover, previous studies have shown that individuals with LBP are more susceptible to fatigue, a characteristic that has been proposed as a potential indicator of the condition [13,14,15]. In parallel, research on individuals with chronic neck pain has suggested that IMC may serve as a biomarker to differentiate between healthy and symptomatic individuals, based on modifications in muscle synergy patterns [16]. Collectively, these findings highlight the potential utility of IMC as a tool to elucidate the complex interplay between fatigue and pain in musculoskeletal disorders.

This study presents several limitations that must be acknowledged. First, the small sample size limits the generalizability of the findings and restricts the ability to draw definitive conclusions, particularly regarding the differences observed in the participant with a history of low back pain. Increasing the number of participants—especially those with LBP—would strengthen the statistical power and allow for more robust comparisons. Second, the limited characterization of the sample, including lack of detailed information on training routines, physical activity levels, or pain chronicity, may have influenced the variability of the responses observed. Lastly, this study focused exclusively on the third squat series to induce fatigue, without analyzing potential changes in IMC across the earlier sets. Including such comparisons could have provided a clearer baseline and a more comprehensive understanding of the progression of neuromuscular adaptations during the protocol.

Future research should aim to address the limitations outlined above by incorporating larger and more diverse samples, particularly individuals with different LBP profiles. Exploring the modulation of IMC across different phases of the exercise protocol, including non-fatiguing sets, would contribute to a more nuanced understanding of muscle coordination and control throughout effort progression.

5. Conclusions

Evidence has been presented that the subject with a history of LBP showed a different behaviour between multifudus muscles compared to the rest of the participants. However, due to the variability of methods and protocols, it was not possible to make direct comparisons of the results obtained with those available in the scientific literature. It is recommended that future lines of research focus on the behaviour of IMC as a potential biomarker in both healthy subjects and those with LBP.