Muscle coordination involves the harmonious interplay of various muscle groups, encompassing coactivation, synergistic, and antagonistic relationships [1]. Coordinated actions between the jaw and neck muscles often exhibit functional coupling, where their activities synchronize to facilitate various oral motor tasks [2,3]. The functional relationship between these muscles is important in the coordination of intricate movements associated with mastication and head positioning [4,5].

The masseter is a powerful jaw-closing muscle that elevates the mandible during biting and chewing. It facilitates the initial stages of closing the jaw and exerting force during the power stroke of chewing. During clenching or chewing, the temporalis and masseter muscles exhibit synergistic relationships, facilitating jaw closure [6,7]. Alternatively, the digastric muscle assists in the initial phase of jaw opening, particularly in controlling the descent speed and direction of the mandible. While it is not a direct antagonist to the masseter, the digastric muscle has opposing functions in its primary actions during chewing as it mainly involves jaw opening rather than closing. The sternocleidomastoid (SM) muscle is a neck muscle with various functions, such as flexing, rotating, and tilting the head and neck. In addition, it plays a role in the chewing process as it helps stabilize the head and neck during jaw movements [8-10].

The coordination pattern of the masticatory muscles was evaluated using various indices calculated as the ratio of muscle activity intensity of muscle pairs [11-13]. Recent studies have employed coherence analysis to reveal complex coordination patterns between the jaw and neck muscles [14,15]. This analytical approach reveals the subtle interactions that control orofacial motor tasks by analyzing the synchrony and functional relationships between the involved muscles during various oral and mandibular movements.

Coherence serves as a measure of synchronization between two signals [16]. Coherence analysis involves computation of the average estimates of the cross-power spectrum and power spectra within segments, a method first introduced by Welch [17]. The coherence value reaches one when two signals exhibit complete synchrony at a specific frequency. Conversely, diminishing synchrony between signals causes the coherence values to approach zero. Transfer function analysis evaluates how an input signal interacts with the resulting output signal by considering its relative strengths and temporal alignment. Within this analysis, the gain value quantifies the relative strength, indicating how alterations in the input signal correspond to changes in the variability of the output signal [16,18]. Meanwhile, the phase value reveals the time lag between the signals [16,19]. For example, a phase of 180° at a specific frequency indicates complete inversion of the frequency components between the signals, elucidating their temporal relationship.

Against this background, the present study aimed to explore the coherence, gain, and phase interactions between the jaw and neck muscles during mastication in healthy adults. Specifically, surface electromyography (EMG) was employed to evaluate the relationship between the major jaw and neck muscles to understand their synchrony, interaction strength, and temporal alignment during chewing at a frequency of 1 Hz. For preliminary exploration for expanded follow-up research, this study first attempted to investigate the coordination between the jaw and neck muscles on the ipsilateral side under a unilateral mastication condition. The test hypothesis was that no coherence, gain, or phase interactions exist between the jaw and neck muscles during chewing in healthy adults.

1. Ethical Approval

Ethical approval was obtained from the Institutional Review Board of Chonnam National University Dental Hospital (CNUDH-2021-002). All experimental procedures were performed in accordance with the principles of the latest revision of the Declaration of Helsinki. Each participant was informed about the study procedure and provided written informed consent before participation.

2. Subjects

The included subjects had no tooth loss except for third molars, no crossbite, an anterior overbite and overjet <4 mm, a dental midline discrepancy <3 mm, and no horizontal tooth mobility. The exclusion criteria were subjects with temporomandibular disorders (TMDs), neck pain, neurological disorders, prolonged medication exceeding 1 month, current orthodontic treatment, and any known history of trauma or surgery in the head and neck regions.

Volunteers were recruited through a bulletin board posting at the Chonnam National University Dental Hospital. Among the 26 volunteers, 8 who did not meet the inclusion criteria and 6 who had errors in the EMG measurement or problems with the quality of the EMG recordings were excluded, resulting in the inclusion of 12 subjects in the final analysis. The volunteers and participants received differential financial support depending on their participation level. The researcher who led this study took on multiple roles, including overseeing recruitment, conducting screening evaluations, selecting participants, and performing subsequent EMG measurements.

3. Surface EMG Detection

A pair of surface electrodes was placed on the skin of the left and right sides of the anterior portion of the temporalis (TA), superficial masseter (MS), the anterior belly of the digastric (DA), and SM muscles. The positioning of the surface electrodes for the TA and MS muscles was described in a previous study [20]. The electrodes were positioned approximately midway between the mental and hyoid attachments for the DA muscle and halfway between the mastoid process and sternal notch for the SM muscle. Before electrode application, the skin was meticulously cleaned with 83% v/v ethanol. Disposable, monopolar, disc-type, Ag/AgCl surface electrodes with a 10-mm diameter were affixed to the skin using adhesive gel (T246H; Bio Protech). The electrode pairs were spaced at a fixed distance of 20 mm and secured with adhesive tape. As a reference electrode, a single monopolar electrode was placed over the skin over the spinous process of the seventh cervical vertebra. Each circular electrode had a conducting surface diameter of 14 mm.

Myoelectric signals from each electrode setup were amplified using an eight-channel single differential EMG system (WEMG-8 [LXM5308]; Laxtha). The signal was amplified within a bandwidth of 13 to 430 Hz (Butterworth filter; −3-dB response; with slopes of the cutoffs: low-pass 48 dB/octave, high-pass 12 dB/octave), sampled at a rate of 1,024 Hz, and digitized through a 10-bit A/D converter using a differential amplifier. This amplifier had an input impedance exceeding 100 MΩ, a high common-mode rejection ratio (>90 dB), and an input signal amplitude ranging from 312.5 μVp to 5 mVp. The resultant EMG signals were either monitored or stored as data files on the hard disk drive of a personal computer, facilitated by a dedicated software (TeleScan ver. 3.01; Laxtha).

4. Experimental Protocol

The task and EMG recordings for the study subjects were performed in a controlled environment—an isolated, quiet room equipped with a dental chair as well as regulated lighting and temperature. Each participant assumed an upright position in a comfortable chair without a headrest, maintaining a natural and unsupported posture. With their gaze fixed on a wall positioned 2 m in front of them, they were given precise instructions regarding the experimental protocol.

Resting muscle activity was initially recorded for a duration of 10 s, assessing both the relaxation of the monitored muscles and any noise within the recorded myoelectric signals. For the chewing task, the participants were instructed to begin by chewing two pieces of gum (Xylitol Original Gum; Lotte Wellfood). Then, with their eyes closed, they were prompted to chew the gum with their right molars at a frequency of 60 strokes per minute (equivalent to one stroke per second, synchronized with a metronome set at ♩=60) while focusing on the beat of the metronome. Concurrently, EMG signals were continuously recorded for 20 s throughout the chewing task.

5. Signal Processing

Of the eight channels capturing EMG signals during the chewing task, only the four channels on the right side were used for analysis. The raw EMG signals obtained from the recordings underwent digital filtering within a low-pass range of 0-500 Hz and a band-stop frequency of 60 Hz. Subsequently, the filtered signals were converted into text files using a dedicated software (TeleScan version 3.01; Laxtha).

Thereafter, signal analyses were conducted on the EMG datasets using Python 3.9.0 (https://www.python.org/) along with the SciPy (https://scipy.org/) and NumPy (https://numpy.org/) modules. The EMG signals were processed as absolute values, ensuring that all data points were converted to positive numbers (Fig. 1A, B). The power spectrum of the EMG signal within the range of 0-500 Hz was computed using the fast Fourier transform (FFT) algorithm, implemented using the scipy.fft function (Fig. 1C, D).

As the segmentation of the EMG record dataset increases, the confidence level in the coherence function analysis diminishes, as observed in a previous study [14]. The low- and high-coherence values become statistically significant due to the reduction in data points per segment when the number of segments increases. Such a reduction results in decreased frequency resolution, as calculated using the following formula:

Frequency resolution=Sampling frequencyNumber of data points

Here, sampling frequency refers to the number of samples taken per unit time, and the number of data points denotes the total number of samples used in the FFT algorithm. Increasing the number of data points (i.e., length of the signal) enhances the frequency resolution. Consequently, data gathering over an extended period yields more data points, improving the frequency resolution of the FFT.

In this study, the 20-s EMG recordings were divided into four segments. As the sampling frequency was 1,024 Hz, 20,480 data points were obtained for 20 s of recordings. Given the segmentation into four parts, each segment comprised 5,120 data points (acquired over 5 s). Consequently, the resulting frequency resolution was 0.2 Hz (1/5 s).

Coherence analysis was employed to evaluate the relationship between the two sets of EMG data in the frequency domain. The coherence function is mathematically expressed as follows and was obtained using the following scipy.signal.coherence function:

Coherence=Sxy(f)2Sxx(f)Syy(f)

Here, S_xx(f) and S_yy(f) denote the power spectral densities obtained using the scipy.signal.welch function, whereas S_xy(f) denotes the cross-spectral density derived from the scipy.signal.csd function. In this study, the reference signal was MS (input signal for the transfer function), whereas TA, DA, and SM were considered as the signals under examination (output signal for the transfer function). Conceptually, TA acts as a synergist with MS, whereas DA functions as an antagonist to MS.

The transfer function H(f) was initially derived using the following equation:

Transfer function H(f)=Sxy(f)Sxx(f)

From this transfer function, both the gain and phase values were obtained. With H(f) being a complex number, the numpy.abs function was used to calculate the gain (|H(f)|) and the numpy.angle function to determine the phase (∅(f)):

Gain H(f)=HR(f)2+HI(f)2Sxx(f)

Phase ∅(f)=tan−1HI(f)HR(f)

where H_R(f) and H_I(f) denote the real and imaginary parts of H(f), respectively.

In the power spectrum where MS was the reference signal, the primary peak distinctly emerged at 1 Hz (Fig. 1C), corresponding to the chewing rhythm frequency. Hence, all the coherence, gain, and phase (unit: rad) values were determined at a frequency of 1 Hz within the power spectrum. Fig. 2 presents an example of the coherence, gain, and phase relationship between the EMG signals of the MS and DA muscles. The analysis encompassed coherence, gain, and phase evaluations across the following muscle pairs: MS and TA, MS and DA, and MS and SM. Notably, the phase values were converted from radians to degrees before statistical analysis.

6. Statistical Analysis

The normality assessment using the Shapiro–Wilk test showed that not all data followed a normal distribution. Consequently, medians and interquartile ranges were used as measures of central tendency and dispersion. Given the nonnormal distribution, nonparametric statistical tests were employed for all subsequent analyses.

The pairwise median differences between the experimental conditions were evaluated using the Friedman test along with post hoc Wilcoxon signed-rank tests, incorporating the Bonferroni correction for multiple comparisons. The Friedman test was explicitly used to detect significant differences in coherence, gain, and phase among the following muscle pairs: MS and TA, MS and DA, and MS and SM. In the Friedman ANOVA tests, p<0.05 was considered to indicate statistical significance. Furthermore, for post hoc analyses with Bonferroni correction, the significance threshold was set at p<0.017. Statistical analysis was conducted using IBM SPSS Statistics for Windows, Version 21.0 (IBM Co.).

Following the Friedman test conducted across three muscle pairs for coherence with 12 participants, post-analysis evaluation using G*Power 3.1 [21] revealed a significant effect size (partial η²=0.617, Cohen’s d=1.269), and the actual power of 0.941 achieved exceeded the desired threshold (1-β) of 0.9. These results confirm the adequacy of the sample size for detecting significant effects within this experimental setup.

1. Participants

A total of 12 healthy young adult subjects (eight female participants and four male participants; mean age, 26.6±2.4 years) were included in the final analysis after screening. The mean±standard deviation (SD) height and weight of the subjects were 168±6 cm and 61.3±10.0 kg, respectively. Furthermore, their mean±SD body mass index (BMI) was 21.5±2.4 (range, 19.0 to 27.3) kg/m².

2. Coherence of the Jaw and Neck Muscles

The median coherence value for the MS and TA muscle pair was 0.992, almost reaching 1.0. Contrarily, the median values for the MS and DA muscle pairs and MS and SM muscle pairs were 0.848 and 0.957, respectively, which were lower than those of the MS and TA muscle pair (Table 1, Fig. 3). Significant differences were observed in coherence values among the muscle pairs (χ²(2)=17.167, p<0.001). Post hoc analysis revealed that both MS and DA (Z=−3.059, p=0.002) and MS and SM (Z=−2.981, p=0.003) muscle pairs exhibited significantly lower coherence than the MS and TA muscle pair. However, no statistically significant difference was observed between the MS and DA or between the MS and SM muscle pair (Z=−1.726, p=0.084) (Fig. 3).

3. Gain of the Jaw and Neck Muscles

The MS and TA muscle pair exhibited a median gain of 0.458, followed by the MS and DA muscle pair with a median gain of 0.201. The MS and SM muscle pair had the smallest median gain at 0.025, nearly approaching 0 (Table 1, Fig. 4). Significant differences in gain values were observed across muscle pairs (χ²(2)=20.667, p<0.001). Post hoc analysis revealed that both the MS and DA (Z=−2.824, p=0.005) and MS and SM (Z=−3.059, p=0.002) muscle pairs had significantly lower gain values than the MS and TA pair. In addition, the MS and SM pair exhibited significantly lower gain values than the MS and DA pair (Z=−3.059, p=0.002) (Fig. 4).

4. Phase of the Jaw and Neck Muscles

The median phase values for the MS and TA and MS and SM muscle pairs were 1.40° and −0.05°, respectively, approaching 0°. Contrarily, the median phase for the MS and DA muscle pair was −192.85°, close to −180° (Table 1, Fig. 5). A significant discrepancy in phase values was observed among muscle pairs (χ²(2)=18.000, p<0.001). The phase values of the MS and DA muscle pair were significantly lower than those of the MS and TA muscle pair (Z=−3.059, p=0.002) and MS and SM muscle pair (Z=−3.059, p=0.002). However, no statistically significant difference was observed between the MS and TA muscle pair and the MS and SM muscle pair (Z=−0.628, p=0.530) (Fig. 5).

The present study investigated the intricate coordination between the jaw and neck muscles during chewing in healthy adults, examining the coherence, gain, and phase relationships between specific muscle pairs.

The results indicated a closely intertwined functional relationship between the TA and MS muscle pair, suggesting that their synchronized and coordinated activity is crucial for efficient chewing movements. High-coherence values between these muscles imply robust synchronized activity, a strong functional relationship that facilitates efficient and coordinated jaw movements. Gain values possibly reflect the influence or dominance of one muscle over the other, where a higher gain from the MS to the TA muscle may indicate a significant impact or control during chewing. Furthermore, consistent phase alignment between the two muscles implies coordinated temporal relationships, potentially facilitating smooth and synchronous muscle contractions important for effective chewing.

The phase difference between the activity of the DA muscle and that of the MS muscle was −193°, close to −180°. This suggests that the DA muscle exhibits an antiphase to the MS muscle, potentially antagonizing the activity of the MS muscle and preceding it in initiating the opening motion during each stroke. In this study, the negative phase values for the MS and DA muscle pair are contrary to the positive values reported in other studies [3,14,22]. As the phase repeats in a 360° cycle, both −180° and 180° indicate an antiphase relationship. Lower coherence levels may imply insufficient simultaneous or coordinated action between the DA and MS muscles. Lower gain values can indicate a balanced or less-dominant influence of the MS muscle over the DA muscle. These findings may be partially due to the characteristics of surface EMG signal obtained from the attachment site of the DA muscle. Due to the crosstalk phenomenon [23], signals from adjacent suprahyoid muscles possibly influence the signals from the DA muscle; as a result, signals obtained from this location may not purely reflect the activity of the DA muscle.

As regards the SM and MS muscle pair, higher coherence values between these muscles suggest synchronized or coordinated activity during chewing. Low gain values indicate that the SM muscle facilitates the action of the MS muscle during chewing, whereas consistent phase alignment signifies synchronized temporal relationships, contributing to coordinated jaw movements important for efficient chewing. Comparable studies conducted under similar experimental conditions reported coherence, gain, and phase values of the jaw and neck muscles on the chewing side that were closely consistent with the findings of this study [3,14].

Chewing rhythm defines an individual’s chewing pattern and speed and is often observed to fall within the range of 1.0-1.25 Hz [24-26]. Coherence analysis revealed that the frequency of the chewing rhythm closely aligns with that in previous studies using other existing methods. In adults with normal teeth, coherence analysis revealed an average first peak frequency of the jaw and neck muscles during gum chewing of 1.15 to 1.25 Hz [3,14,22]. For denture wearers, this peak was observed at 1.37 Hz [22]. Ishii et al. [14] reported that this initial peak in the power spectrum is consistent with the chewing rhythm, providing insights into the power, temporal, and synergistic relationships between the jaw and neck muscles in healthy subjects.

However, in our study, the subjects exclusively chewed gum at a frequency of one stroke per second, synchronized with a metronome set at 1 Hz. Comparison of our findings with those of studies conducted under habitual chewing conditions may not be directly feasible. The coherence patterns observed might differ for several reasons. Habitual gum chewing involves natural muscle coordination, whereas chewing at a specified frequency imposes a rhythm that may adjust muscle synchrony or desynchrony, potentially altering coherence levels [27]. Furthermore, chewing at a frequency of 1 Hz may prompt specific muscle responses, leading to detectable changes in synchronization patterns. In this study, the first peak at 1.0 Hz was evident in the power spectra of all three muscle pairs. Moreover, habitual chewing engages a wider range of muscles due to varying food textures and intensities [28]. Chewing at a controlled frequency can limit muscle recruitment, influencing coherence patterns because of altered muscle engagement [29]. Additionally, rhythmic chewing might induce motor control adaptations [30,31], potentially affecting the observed muscle coordination dynamics in coherence analysis. Further studies should delve deeper into these inferences to elucidate the nuances.

Coherence analysis finds diverse applications across various conditions. Ishii et al. [3] reported that chewing rhythm, quantitative and temporal coordination, and functional coordination in jaw and neck muscle activities remained unchanged during soft and hard gum chewing or increased jaw and neck muscle activities. Narita et al. [22] demonstrated that gain values in individuals wearing dentures for jaw closing and opening decreased compared with those without dentures. Denture wearers exhibited comparable gain values for these activities to young subjects but had higher coherence values for neck muscle activities on the chewing and nonchewing sides. Additionally, Fassicollo et al. [15] investigated muscle coordination in patients with TMDs. Their findings indicated prolonged chewing stroke duration as well as impaired differential recruitment and coherence among the TA, MS, and SM muscles. Furthermore, they observed peak timing delays and temporal asynchrony between the jaw and neck muscles.

According to the BMI standards of the World Health Organization [32], 1 subject in this study who had a BMI of 27.3 kg/m² was classified as overweight whereas the remaining 11 subjects were classified as normal. In general, individuals with higher BMI values tend to have a higher proportion of body fat, including both subcutaneous and visceral fats. Subcutaneous fat can influence the amplitude and mean frequency estimates of surface EMG signals [33,34]. Although BMI can be a contributing factor, its direct influence on EMG signals might not follow a linear or easily quantifiable pattern. Other individual variations, such as muscle structure, hydration levels, and electrode placement precision, also contribute to signal variability [35-38]. Similarly, the coherence, gain, and phase of surface EMG signals are believed to be indirectly affected by BMI or subcutaneous fat thickness and influenced by multiple factors. Upon reviewing the coherence, gain, and phase data of the subject with an overweight BMI in this study, the gain value of the MS and SM muscle pair was an outlier, whereas the others were not.

This study has several limitations. First, the sample size was relatively small, limiting the ability to discern potential gender-based differences in coherence between women and men within jaw and neck muscle coordination. Second, the study employed a single chewing task to explore the coordination between the jaw and neck muscles. Examination of various mastication movements would have provided a more comprehensive understanding of the complexity of muscle coordination. In addition, analysis of the coordination of muscles on only one side of the body associated with chewing on the right side limited an expanded assessment of bilateral muscle interactions. Third, the surface EMG device used for the measurements exhibited inherent limitations. For example, crosstalk effects could have reduced the accuracy of the signal from the muscles to be monitored. This is particularly true for the DA muscle, as it has a small volume and the suprahyoid muscles are nearby. Finally, the inclusion of only healthy adults in this study made it difficult to generalize the results to populations with disease conditions such as TMDs or oral motor disorders.

Future studies should have a larger sample size and include healthy individuals and populations with various diseases. Furthermore, future investigations should explore various oral motor tasks using improved EMG equipment and techniques to understand the broader dynamics of muscle coordination and increase the accuracy of muscle activity assessment.

In conclusion, this study provides valuable insights into the coherence, gain, and phase relationships between the jaw and neck muscles during chewing movements in healthy adults. The TA muscle synergistically coordinates with the MS muscle, whereas the DA muscle exhibits antagonistic characteristics. Additionally, the SM muscle facilitates the activity of the MS muscle. The observed patterns of synchrony and coordination among specific muscle pairs highlight the intricate interplay involved in oral motor tasks. Understanding the nuanced interactions between these muscle groups holds promise for elucidating normal oral motor function and potential implications for therapeutic interventions in orofacial movement disorders.

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

Participants in this study did not provide written consent for the public sharing of their data. Because of the sensitive nature of the research, supporting data are not available for public access.

This study was financially supported by the Chonnam National University (Grant number 2021-2506).

Conceptualization: YGI. Formal analysis: HJS, SHH, YGI. Funding acquisition: YGI. Methodology: HJS, SHH, YGI. Project administration: YGI. Writing - original draft: HJS, YGI. Writing - review & editing: HJS, SHH, JYK, YGI.