Analysis of the average peak sEMG activity of masseter and temporalis muscles during chewing has been accomplished many decades ago to determine the timing sequence using one of the four muscles as a reference. Hannam et al early on discovered that the elevator muscles to not contract in unison during mastication but follow a consistent sequence of firing.1 They also revealed the extent that the masseter and temporalis muscles, which have multiple segments, function in a complex manner. Van Eijden et al more recently expanded on the previous work revealing that in normal subjects, different segments of the same muscle peak at different times that are specific to the activity being performed.2 Using a complicated array of subcutaneous wire electrodes they showed differences in anterior, medial and posterior segments of the muscles as well as significant differences between working and balancing side activity levels. However, their method of recording is not very suitable for routine use in a clinical setting.

The limitation of using one muscle as a reference is that it creates an artifactual error by transferring the variability of the reference muscle to all of the other muscles recorded. Incisor-point tracking was developed in the 1970s as a method of jaw motion recording that did not interfere with the process of mastication.3,4 Subsequently, sEMG activity has been recorded simultaneously along with mandibular incisor-point motion serving as the reference. Nielson et al also recorded from normal subjects using an incisor-point tracker to record motions and positions along with the sEMG muscle activity from bilateral masseter, anterior temporalis and suprahyoid areas.5 They revealed that the requirements of a particular position or motion determine the recruitment pattern in normal subjects.

Kuwahara et al developed the Averaged Chewing Pattern (ACP) as a method of differentiating normal chewing motions from abnormal chewing motions.6–8 They concluded that there are four commonly observed distinct frontal patterns, two distinct sagittal patterns and two distinct horizontal patterns. More recently Kerstein et al recorded the incisor-point motions along with surface sEMG of the bilateral masseter and anterior temporalis muscles’ activity from a group of symptomatic and dysfunctional TMD patients masticating prior to and post occlusal adjustment.9 The study revealed that after Immediate Complete Anterior Guidance Development occlusal coronoplasty, the patients’ Average Chewing Patterns (ACP) improved towards more normal patterns. The sEMG data from that study revealed a significant reduction in the time from the onset of opening to the peak of the muscle activity in closing.

Therefore, all “TMD patients” cannot be expected to consistently relate to any one particular chewing feature. The reality is that different dysfunctions affect the masticatory process in different ways. Also, some patients with specific disordered conditions, such as a TMJ internal derangement, eventually adapt to the condition and are able to produce relatively normal masticatory function. It is the poorly adapted patient that continues to produce anomalous patterns of both movement and muscle function. The simple act of identifying normal masticatory function in a patient with an identified but well adapted disorder could restrain any impulse toward unnecessary treatment.

While the exact timing of the peak activity of each elevator muscle varies with each masticatory closure, dependent upon the instantaneous status of the bolus, it is possible to average a sequence of cycles and establish a mean pattern of sEMG activity for each muscle.8–10 See Figure 1.

Figure 1
Figure 1.The averaged chewing cycle (ACC) of muscle activity from a normal control subject. The peaks of the sEMG activities of the masseter and anterior temporalis muscles occur prior to the end of closure (maximum bolus crush). The highest activity from the ipsilateral masseter is followed by the ipsilateral temporalis, contralateral temporalis and contralateral masseter. Note that the opening motion starts to the contralateral but returns to the ipsilateral for closing.

By simultaneously recording incisor-point motion as the reference, it is possible to compare the mean peak sEMG timings between elevator muscles as a bolus is masticated. In normal control subjects, the averaged masseter and anterior temporalis sEMG activities all peak before the end of closure (maximum bolus crush). Guo et al recorded the surface sEMG activity from the bilateral masseter and sternocleidomastoid (SCM) muscles while healthy asymptomatic subjects chewed gum for 15 cycles.10 They measured and averaged the time from the beginning of opening to the peak of the activity in each muscle. Their main conclusion was that the balancing side SCM peak occurred significantly before the other muscles in relation to the Terminal Chewing Position (TCP) of maximum bolus crush.

Other authors after reviewing the literature,11 or analyzing small groups,12 have suggested that despite all the many studies revealing significant differences between TMD and control subjects with respect to masticatory motion, “There is no strong evidence that any particular chewing feature is characteristic of TMD-patients.” This conclusion is the result of treating TMD as if it were a homogenous diagnosis. In fact, TMD is not even a diagnosis, but a category of at least 40 different diagnoses.13

It has been previously established that the amplitude of sEMG activity is nearly linearly related to the force applied by the muscle, which suggests that the peak of the sEMG activity represents the maximum effort of the individual muscle within a chewing burst.14,15 A delay in the timing of the peak sEMG activity during mastication until after the bolus is fully crushed suggests some degree of hesitancy and uncertainty on the part of the subject. The reasons for hesitancy may be multiple; a malocclusion, a TMJ internal derangement, a maxillo-mandibular mal-relationship, a sensitive tooth or any condition that can interfere with mastication.


Calculate the average timing and variability of the sEMG peak muscle activity from the bilateral masseter and anterior temporalis muscles of patients with dysfunctional TMJs during mastication. Compare the dysfunctional group’s timings to those of an age and gender matched asymptomatic control group.


Inclusion Criteria

  • TMD patients from a private practice previously diagnosed from history, clinical examination, jaw tracking, sEMG, Joint Vibration Analysis (JVA), CBCT and/or MRI that were identified as having bilateral internal derangements not well adapted.

  • Evidence of masticatory dysfunction in the ACP and the ACC, as slow, small, highly variable.16

  • For the control group; asymptomatic volunteers with no history or current indications of any masticatory dysfunction

Exclusion criteria

  • Unilateral masticatory dysfunction, but with good function on the contralateral side

  • The presence of any systemic disease or trauma limiting masticatory function

  • Bilateral TMD patients that were so well adapted as to function within normal limits on at least one side with fast, large and consistent ACP & ACC

Previously recorded gum chewing records from 57 TMD patients (33 F, 24 M) identified with bilateral TMJ internal derangements were extracted from a database of magnetic incisor-point tracking together with their simultaneously recorded electromyographic records that were routinely used to evaluate the masticatory function of TMD patients (JT-3D and BioEMG III, BioResearch Associates, Inc. Milwaukee, WI USA). The electrode placements are shown in Figure 2. The time resolution of the data, which were sampled at 2000 samples/second, was 0.5 milliseconds.

Figure 2
Figure 2.The electrode placements used for the bilateral anterior temporalis and superficial masseter muscles.

Each subject’s masticatory dysfunction was confirmed from:

  1. history of TMD symptoms,

  2. clinical signs of TMD,

  3. bilateral CBCT and/or MR images of the TMJs,

  4. slow, small and/or highly variable Averaged Chewing Patterns (ACP)16–19

  5. abnormal sEMG Averaged Chewing Cycles (ACC)16,18

Factors 4 & 5 while masticating a gum on the left and right sides produced 114 records. See Figure 3. 60 asymptomatic volunteers (39 F, 21 M), all with normal appearing ACP and ACC mean chewing patterns (n = 120) were previously recorded with the identical protocol. The control group database was originally designed to match an average TMD group in age at 35 (+/- 15) years, but this ID group turned out to be older, mean age 42 (+/- 16) years (p = 0.01), probably because actual masticatory dysfunction increases with age although the other painful TMD symptoms do not. A typical example of a JVA record showing a left TMJ Piper 4a reducing disk displacement is shown in Figure 4.

Figure 3
Figure 3.History, clinical exam, JVA, CBCT (top) and MRI (bottom) all providing evidence of poor adaptation bilaterally.
Figure 4
Figure 4.JVA of dysfunctional subject with a left TMJ reducing anterior disk displacement with no degenerative indication.

Objective measurements were used to select all the subjects and controls, eliminating any need for blinding, calibration of examiners or Kappa statistics as required for subjective approaches such as the DC/TMD. As is usual, the End of Closure was established at a point 0.3 mm vertically below maximum bolus crush at the end of each closing movement, which provided a consistent reference position for the timing. See Figure 1. This also partially compensates for the known fact that the recorded motions are slightly delayed a few milliseconds in relation to the sEMG activity as recorded. We were most interested in the relative timing between groups and since the small delay caused by the jaw tracker is constant and exactly same for both groups, it was deemed irrelevant to these timing measurements.

The positions of the muscle activity peaks were recorded in milliseconds before or after the reference time-point. The sEMG activity of 15 cycles for the bilateral masseter and anterior temporalis muscles (4) were recorded for each subject’s left and right chewing and then the mean and standard deviation were calculated for each muscle in each group according to working masseter, working temporalis, balancing masseter and balancing temporalis. Using this approach allowed combining the left and right data because we found no significant differences between chewing sides. The Mann-Whitney U Test was used to compare the means of each of the four muscles between the two groups.

All patients were previously informed of all procedures and consented to all the diagnostic procedures. Inasmuch as sEMG measurements themselves are perfectly blind, no additional blinding of operators was required. An IRB exemption was sought and approved # BIRB/952Z/020.


A significant delay was found in the timing of the peaks of all four muscles within the dysfunctional group compared to the control group (p = 0.00001). See Table 1 and Figure 5.

Table 1.The timing of the peak of the sEMG activity of the dysfunctional subjects Vs normal control subjects.
Mann-Whitney U test Working TA Working MM Non-working MM Non-working TA
Dysfunctional Male 4.2 12 2.4 17.7
Dysfunctional All 4.7 13.8 1.3 15.4
Dysfunctional Female 5.1 15.1 0.5 13.8
p = 0.00001 0.00001 0.00001 0.00001
Normal Male -54.6 -45.9 -61.8 -41.1
Normal All -55 -49.3 -67.3 -43.3
Normal Female -55.3 -51.2 -70.2 -44.5

A negative value (milliseconds) indicates the peak occurred before the End of Closure.

Figure 5
Figure 5.Within the control group the mean sEMG peak values were all located well before the end of closure. The dysfunctional groups’ mean sEMG peak values occurred after the end of closure, but in the same order as for the normal control group. No significant gender differences were found in either group.

Note: The relative relationships between the four muscles within each group were the same;

  • non-working masseter first to peak

  • working temporalis second to peak

  • working masseter third to peak

  • non-working temporalis the last to peak

The variability of the dysfunctional group was also significantly greater than the control group (p < 0.05). No significant differences were found between genders in the timing of the sEMG activity peaks for any muscle or either group (p >> 0.05). See Figure 5.


The location of the sEMG peak activity is only one of many possible indications of a TMD condition. It is certainly not expected that this one parameter should be used to diagnose all TMD. However, a delay in the sEMG activity peak could contribute to estimating the degree of hesitancy of the patient in mastication and appears to be common in patients not well adapted to their TMD condition. It also concurs with a diminished amplitude of chewing, a slower rate of chewing and an increase in the variability.16–19 Figure 6.

Figure 6
Figure 6.An example of a dysfunctional chewing pattern with a slow, small, highly variable and distorted ACP. The ACC shows weak contractions with the sEMG peaks all late, well after the End of Closure.

The non-simultaneous muscle activity peaking patterns found in this study are in agreement with Hannam, Van Eijden and Nielson.1,2,5 The consistent mean pattern of the order of the peak sEMG activity between the four muscles found did not agree with the Soboļeva et al’s claim of a simultaneity of action in closing of the contralateral temporalis and both masseters.11 See Figure 5. However, the mean time difference from first to last sEMG peak in our dysfunctional group was only 14.2 milliseconds and this could have easily been overlooked. Also, at least within this group of normal asymptomatic subjects, the masticatory activity from the anterior digastric area continued to produce activity throughout the closing phase and did not cease with closure as has been suggested by some previously.11 See electrode placement for anterior digastric muscles in Figure 1.

The primary objective of this study was to measure the time from the peak of the closing muscle activity to the end of closure (maximum bolus crush) in TMD patients, review the same measurements in subjects with good masticatory function and compare the two groups. The significant difference found infers a hesitancy on the part of the dysfunctional patients that likely contributes to their slower chewing rate as well. Coincidentally, the relative pattern of peak activity between the four elevator muscles appeared to be rather consistent between both the asymptomatic control group and the TMD patients, suggesting that the averaged pattern is probably the most efficient one even in the presence of TMJ dysfunction. This fact of mastication was not found to have been previously disclosed in a search of the dental literature.

The standard deviations of the timing of the sEMG peaks were quite large in relation to the means for both the normal subjects (coefficients of variation from 0.433 to 0.648) and even greater for the dysfunctional group (all coefficients of variation > 1.0). This indicates that the rate of chewing and the individual cycle times vary greatly from cycle to cycle as the bolus is consumed even within the normal control subject group, no two cycles are the same. See Figure 7. Consequently, the mean values of the normal group, while they may be ideal, are not required for good masticatory function. Three of the control subjects exhibited one muscle’s sEMG peaking a few milliseconds after the End of Closure and one control subject exhibited two muscles peaking slightly after the End of Closure. However, it can be expected that as a general rule, sEMG muscle peaking prior to the End of Closure should certainly be considered the most normal pattern.

Figure 7
Figure 7.The variability of muscle function is greatest for the poorly adapted internal derangement cases. Well adapted cases are still more variable than normal control subjects but are able to function well enough. Note the abnormally high non-working masseter activity of the well adapted case. This is often seen as one means of adaptation. It does suggest that the in this case the masseters to never rest when “not working.”

In the ideal scenario the non-working masseter peaks first during the closure but provides very little effort, thus it is mostly resting. The working temporalis peaks second as the closure progresses until sufficient resistance from the bolus occurs, at which time the working masseter peaks nearer to the End of Closure. The non-working temporalis peaks last as it and the non-working masseter are normally acting only as stabilizers rather than as bolus crushers. Although this study did not find any significant differences between genders for the patterns of sEMG peaking, others have detected timing differences in chewing movements.20

It is not an absolute requirement for the muscles peaks to occur in the common order seen here or for the intensities to follow the normal hierarchy (1–2–3–4) of the working masseter, the working temporalis, the non-working temporalis and non-working masseter. Often the patterns are altered in one or more of the muscles, even in an asymptomatic control subject. The differences are random in the control group, causing them to be canceled, but in the dysfunctional group it was noted that the non-working masseter was most often altered in a subject’s attempt to adapt effectively. The greater variability of the dysfunctional group resulted in much smaller differences in the spacing between the mean peaks of the four muscles and the extent that the non-working. See figure 5. When the resting masseter is not allowed to ever rest, the masseter muscles can become tired during chewing, especially when the bolus is tough.

The finding of high variability even among normal control subjects, calls into question the concept of a central pattern generator that provides a chewing rhythm.21 The term rhythm suggests a more or less steady timing of opening and closing, but clearly the duration of a cycle is continuously variable as the bolus is consumed during mastication. The masticatory system controller (CNS) is required to evaluate the status of the bolus between closures and adjust the motion to maximize the consumption of it. Since successive cycles very substantially, it is unlikely that a predetermined pattern of motion could be very useful to the process of mastication. It is more likely that the mastication of food is a learned skill and that it requires updating whenever there are changes to the system.22,23 This concept would also concur with the previous empirical finding that patients receiving new complete dentures are rather quickly able to adapt to them.24


The study group was limited to patients with masticatory dysfunction that was bilateral to avoid any suggestion that a preferred side might be different than a non-preferred side. Ultimately, this also gave us twice the number of data points for a given number of patients, since no significant differences were found between left and right chewing. A patient dysfunctional on only one side could avoid chewing on that side, prefer to use only the better functioning side, exhibit fewer symptoms and claim to have a preferred side.

The determination of what is normal for the control group is always partially subjective. Some subjects are more inclined to complain about a particular condition that others may not complain about. It is very rare to find a perfect control subject with no hint of any dysfunction, occlusal wear, misaligned teeth, etc. Being asymptomatic is a matter in part of the patient’s tolerance level. While the control group was without complaints, they were not perfect. However, the imperfections within the control group being random, it is likely that most of those effects were effectively canceled by averaging the data of the whole group. By avoiding subjective assessments and relying on objective measurements to select the subject and control groups, investigator bias was reduced.

Surface EMG records from an extended area beyond the immediate surface of the patient’s skin and the possibility that some of the recorded data came from another muscle besides the one specified is real. However, for the masseter and anterior temporalis during function, the facial muscles cannot override them and the possibility that other masticatory muscles (e.g., internal pterygoid, buccinator, etc.) may have contributed is possible, but not likely to have been detrimental because they should be comparable for both the subjects and controls. The locations of the masseter and anterior temporalis electrodes are standardized today and can easily be replicated by others.


The mean peak sEMG activities of the masseter and anterior temporalis muscles were significantly delayed in the group of bilaterally dysfunctional TMD patients compared to a group of asymptomatic subjects. The large variability of the activity suggests a cycle-to-cycle adjustment of muscular requirements that are not likely to be pre-programmed in a central pattern generator. It was observed that the relative pattern of the peaks of the sEMG activity between the means of the four elevator muscles were the same for the dysfunctional patient group and the asymptomatic subject group, suggesting some mechanical advantage may be present within this pattern. Further study of the significance of this pattern is warranted.

Clinical Significance

Late peaking of the masseter and anterior temporalis muscles’ sEMG activity in mastication appears to be a characteristic of dysfunction that can be interpreted as a hesitancy during the bolus crush stage. The reasons for the delay may not be limited to either a TMJ disturbance, a mal-occlusion or a maxillo-mandibular mal-relationship, but each of these potential conditions can easily be evaluated. No scientific evidence has yet been established to support a significant causative emotional factor for any masticatory dysfunction.

Conflict of Interest Statement

John Radke is currently the Chairman of the Board of Directors and Parvathi Kadamati is a biomedical engineer for BioResearch Associates, Inc., the manufacturer of the JT-3D and BioEMG III. They receive no monetary value based upon sales and the concepts presented here are not limited to these specific instruments.

Funding statement

No funding was received from any source for any aspect of this study.