Manual Therapy 12 (2007) 372379

Original article
The use of surface electromyography as a tool in differentiating
temporomandibular disorders from neck disorders
Virgilio F. Ferrario

, Gianluca M. Tartaglia, Francesca E. Luraghi, Chiarella Sforza

Dipartimento di Morfologia Umana, Via Mangiagalli 31, I-20133 Milano, Italy
Received 4 July 2005; received in revised form 6 March 2006; accepted 3 July 2006
Abstract
The aim of this study was to assess the electromyographic characteristics of the masticatory muscles (masseter and temporalis) of
patients with either temporomandibular joint disorder or neck pain. Surface electromyography of the right and left masseter
and temporalis muscles was performed during maximum teeth clenching in 38 patients aged 2167 years who had either (a)
temporomandibular joint disorder (24 patients); (b) neck pain (13 patients). Ninety-ve control, healthy subjects were also
examined. During clenching, standardized total muscle activities (electromyographic potentials over time) were signicantly
different in the three groups: 75 mV/mVs % in the temporomandibular joint disorder patients, 124 mV/mVs % in the neck pain
patients, and 95 mV/mVs % in the control subjects (analysis of variance, Po0:001). The temporomandibular joint disorder patients
also had signicantly (Po0:001) more asymmetric muscle potentials (78%) than either neck pain patients (87%) or control subjects
(92%). A linear discriminant function analysis allowed a signicant separation between the two patient groups, with a single patient
error of 18.2%. Surface electromyographic analysis during clenching allowed to differentiate between patients with a
temporomandibular joint disorder and patients with a neck pain problem.
r 2006 Elsevier Ltd. All rights reserved.
Keywords: Temporomandibular disorders; Head pain; Neck pain; Electromyography
1. Introduction
Pain, either in the face, head or neck, is the almost
universal complaint of the patients seeking dental care.
On some occasions, the origin of this pain remains
uncertain, even after the execution of all the conven-
tional clinical and instrumental diagnostic procedures.
Diagnosis is therefore of a generic cranio-facio-
cervical dysfunction, and therapeutic procedures are
focused only on the solution of the symptom, namely
pain relief (Ash and Ramfjord, 1995; DeVocht et al.,
2003).
Some of these cranio-facio-cervical dysfunctions
could be categorized as temporomandibular disorders
(TMD). TMD is a complex disease, and its nature is not
completely understood yet (Ash and Ramfjord, 1995;
Visser et al., 1995; De Wijer et al., 1996; Gross et al.,
1996; Liu et al., 1999; Alcantara et al., 2002; Landulpho
et al., 2004). A large number of TMD patients report
pain in the masticatory muscles, and present symptoms
and signs of muscular alteration (Visser et al., 1995; De
Wijer et al., 1996; Sato et al., 1998; Liu et al., 1999;
Pinho et al., 2000; John et al., 2003; Suvinen et al., 2003;
Landulpho et al., 2004). Among the several treatments
used in the TMD patients, occlusal splints, a conserva-
tive and reversible therapy, can reduce pain in most
cases (Nemcovsky et al., 1992; Ash and Ramfjord, 1995;
Ferrario et al., 2002; Landulpho et al., 2004).
In contrast, in other patients the occlusal splints are
not benecial, and pain remains a major problem. These
patients often present cervical pain as their main
complaint. Indeed, signicant associations between neck
ARTICLE IN PRESS
www.elsevier.com/locate/math
1356-689X/$ - see front matter r 2006 Elsevier Ltd. All rights reserved.
doi:10.1016/j.math.2006.07.013

Corresponding author. Tel.: +39 02 503 15407;

fax: +39 02 503 15387.
E-mail address: farc@unimi.it (V.F. Ferrario).
pain and temporomandibular joint (TMJ) dysfunction
have been demonstrated (Ciancaglini et al., 1999;
Alcantara et al., 2002; Visscher et al., 2002), with some
overlap between stomatognathic and cervical symp-
toms in both TMJ and neck disorders (De Wijer et al.,
1996). The stomatognathic apparatus comprises all the
head and upper cervical structures pertaining to the
digestive apparatus, including the oral cavity, teeth with
supporting bones and gengivae, the tongue, salivatory
glands, the pharynx, the masticatory muscles and TMJ.
Cervical muscles are also involved in TMJ dysfunctions,
and patients can exhibit a more exed head position
than normal individuals of the same age and sex (Lee et
al., 1995), even if non-signicant differences have also
been reported (Visscher et al., 2002). In healthy
individuals mandibular and neck muscles act coordi-
nately in an integrated structure, and alterations in one
part can derange also the other one (Eriksson et al.,
2004).
The assessment of the electromyographic (EMG)
characteristics of cervical and head muscles may allow
a deeper insight into this coupling in both health and
disease (Pinho et al., 2000; Ferrario et al., 2002; Suvinen
et al., 2003). Previous investigations analysed mastica-
tory function in patients with TMJ alterations, and
found EMG a useful diagnostic tool (Sato et al., 1998;
Pinho et al., 2000; Landulpho et al., 2004). EMG has
also provided valuable information in patients with
chronic neck pain (Falla, 2004). On no occasion was
stomatognathic function of patients with TMJ altera-
tions compared to that of patients with cervical pain as
the principal symptom.
In the present study, the EMG characteristics of two
groups of patients with either TMJ alteration or neck
pain have been analysed. The null hypothesis was that
the subjects in the two analysed groups had no
differences in the EMG characteristics of their masseter
and temporalis muscles during standardized teeth
clenching. Patient data were also compared to those
collected in healthy young subjects with a good
occlusion, a group used to provide an external reference
of normality.
2. Materials and methods
2.1. Patients
One hundred and thirty-three subjects aged 1867
years were examined. Thirty-eight were patients refer-
ring to a dental clinic for the treatment of craniocervical
pain, who reported subjective symptoms of pain in the
orofacial and neck regions. After physical and radio-
graphic examination of the neck and stomatognathic
apparatus (De Wijer et al., 1996; Bogduk, 1999) and
considering the subjective information gathered during
the clinical history, the patients were subdivided into
two non-overlapping groups: (a) TMJ disorder (17
women and seven men, age range 2166 years, mean 35,
SD 14); (b) neck pain (11 women and three men, age
range 3067 years, mean 48, SD 17).
Physical examination included palpation of the
masticatory and neck muscles, palpation of the TMJ,
examination of the oral cavity with teeth and supporting
structures, assessment of the TMJ and neck movements.
The TMJ disorder patients had internal derangement
with or without reductions, capsulitis, synovitis, masti-
catory muscle myalgia with articular dysfunction,
arthrosis, arthritis with or without arthralgia. Pain
(either in the TMJ area or over the masticatory muscles,
both spontaneous and during palpation) ranged from
moderate to severe, with limited active mouth opening
and mandible laterodeviations on mouth opening. In all
patients pain duration was inferior to 6 months.
Clicking and other TMJ sounds were also detected. In
contrast, the neck pain patients had craniocervical
pain without TMJ dysfunction (Ash and Ramfjord,
1995; De Wijer et al., 1996; Bogduk, 1999; Visscher
et al., 2002). The neck pain patients reported a general-
ized neck pain (moderate to severe in intensity,
exacerbated by palpation), which was indicative of
myalgia, together with a limitation in the amplitude of
the active head and neck movements. Anamnesis was
negative for whiplash injuries; pain duration was
inferior to 6 months. Radiographic examination of the
cervical spine was negative for anatomical alterations.
The patients entered the study consecutively, but all
patients who fall in both categories were excluded, and
replaced by other subjects. None of them had systemic
diseases involving the locomotor apparatus, vestibular or
neurological problems, or anatomical alterations of the
cervical spine. All patients had a complete dentition, with
at least 24 elements (all natural teeth or partial xed
prostheses), with no crossbite. They had no periodontal
problems, or acute inammatory oral diseases.
Ninety-ve control subjects (28 women and 67 men,
age range 1822 years, mean 20, SD 2) were also
examined. All of them had a complete natural dentition
(28 teeth at least), no crossbite, and were free from
periodontal problems and acute inammatory oral
diseases. Exclusion criteria in the control group were
past or present signs or symptoms of TMD (muscular or
TMJ pain, TMJ noises, limitations on mouth opening,
mandible laterodeviations on mouth opening), neck
problems (including whiplash injuries), systemic dis-
eases. These subjects were used to provide an external
reference of normality, with healthy occlusal conditions.
All the procedures were non-invasive, and performed
with minimal disturbance to the subjects. The subjects
were previously informed about all the adopted proce-
dures, and they all signed an informed consent form
approved by the local ethic committee.
ARTICLE IN PRESS
V.F. Ferrario et al. / Manual Therapy 12 (2007) 372379 373
2.2. EMG recordings and measurements
2.2.1. Instrumentation
The masseter and anterior temporalis muscles of both
sides (left and right) were examined. Disposable silver/
silver chloride bipolar surface electrodes with a diameter
of 10 mm and an interelectrode distance of 2171 mm
(Duo-Trode; Myo-Tronics Inc., Seattle, WA, USA)
were used, while a disposable reference electrode was
applied to the forehead.
Bipolar surface electrodes were positioned on the
muscular bellies parallel to muscular bres as previously
described (Ferrario et al., 2002): temporalis anterior:
vertically along the anterior margin of the muscle (about
on the coronal suture); masseter: parallel to the
muscular bres, with the upper pole of the electrode at
the intersection between the tragus-labial commissura
and the exocanthion-gonion lines. To reduce skin
impedance, the skin was carefully cleaned prior to
electrode placement, and recordings were performed 5
6 min later, allowing the conductive paste to adequately
moisten the skin surface.
EMG activity was recorded using a computerized
instrument (Freely, De Go tzen srl; Legnano, Milano,
Italy). The analogue EMG signal was amplied (gain
150, bandwidth 010 KHz, peak-to-peak input range
from 0 to 2000 mV) using a differential amplier with a
high common mode rejection ratio (CMRR 105 dB in
the range 060 Hz, input impedance 10 GO), digitized
(12 b resolution, 2230 Hz A/D sampling frequency), and
digitally ltered (high-pass lter set at 30 Hz, low-pass
lter set at 400 Hz, band-stop for common 5060 Hz
noise). The signals were averaged over 25 ms, with
muscle activity assessed as the root mean square (r.m.s.)
of the amplitude (unit: mV). EMG signals were recorded
for further analysis.
2.2.2. Standardization recording (clenching on cotton
rolls)
At rst, a standardization recording was performed.
This would provide reference EMG values for a
subsequent normalization.
Two 10-mm thick cotton rolls were positioned on the
mandibular second premolar/ rst molars of each
patient, and a 5-s maximum voluntary clench (MVC)
was recorded. For each of the four analysed muscles
(right and left masseter and temporalis), the mean EMG
potential (r.m.s. of the amplitude) was set at 100%, and
all EMG potentials obtained during MVC directly
performed on the occlusal surfaces (see below) were
expressed as a percentage of this value (unit: mV/
mV100). During all recordings, the patients sat with
their head unsupported and were asked to maintain a
natural erect position. They were invited to clench their
teeth as hard as possible.
2.2.3. MVC directly on occlusal surfaces
EMG activity was recorded during a 5-s MVC test in
intercuspal position: the patient was invited to clench as
hard as possible with the maxillary and mandibular
teeth in maximum contact, and to maintain the same
level of contraction for all the recording. For each
patient, the best (those with the most constant r.m.s.
EMG signal) 3 s of the MVC test were then auto-
matically selected by the EMG software, and the EMG
potential were normalized as detailed before (EMG
amplitude on occlusal surfaces divided for the mean
EMG amplitude of the normalization record on the
cotton rolls). Subsequently, the mean (left and right
masseter and temporalis) total muscle activities were
computed as the areas of the standardized EMG
potentials (normalized r.m.s. amplitude) over time (unit:
mV/mVs %) (Ferrario et al., 2004). Muscle activity has
already been used as a global index of the masticatory
muscle work performed during both dynamic and static
tasks (Ferrario et al., 2004, 2006).
The EMG waves of paired muscles were compared by
computing a percentage overlapping coefcient (POC,
unit: %) (Ferrario et al., 2000). POC is an index of the
symmetric distribution of the muscular activity as
determined by occlusion; it is computed by super-
imposing the left and right side normalized EMG
amplitudes of a muscle (masseter or temporalis) over
time: the area of superimposition is assessed as a
percentage of the total EMG amplitudes. The index
ranges between 0% and 100%: when two paired muscles
contract with perfect symmetry, a POC of 100% is
obtained (EMG amplitudes completely superimposed).
Mean (masseter and temporalis) POCs were obtained
for each patient.
Considering the direction in the lateral plane of the
muscular bres relative to the cranium (xed bone) and
the mandible (mobile bone), the two muscles assessed in
the current study act in an opposite direction: the
supercial part of the masseter muscle goes cranio-
caudally and anterior-to-posterior, while the temporalis
anterior goes cranio-caudally and posterior-to-anterior.
A muscular couple is therefore generated when the
contralateral masseter and temporalis muscles (for
instance, right temporalis and left masseter) contract
(Ferrario et al., 2000). If only one muscular couple is
activated, unbalanced by a similar contraction of the
other muscular couple, a potential lateral displacing
component acting on the mandible might occur. The
Torque coefcient (TC, unit %) (Ferrario et al., 2000) is
calculated by superimposing the right temporalis plus
left masseter normalized EMG amplitudes over the left
temporalis plus right masseter normalized EMG ampli-
tudes: the area of superimposition is assessed as a
percentage of the total EMG amplitudes. TC ranges
between 0% (complete presence of lateral displacing
force) to 100% (no lateral displacing force).
ARTICLE IN PRESS
V.F. Ferrario et al. / Manual Therapy 12 (2007) 372379 374
Reproducibility of surface EMG measurements of the
same muscles has already been tested in our laboratory
by repeated analyses of seven subjects chosen at random
(Ferrario et al., 2006). For all EMG variables the intra-
class correlation coefcients were larger than 0.63,
showing a good accuracy of the measurements, without
random errors (paired Students t test, P40:05).
2.3. Data analysis
Descriptive statistics were computed for the control
group, and for each of the two patient groups. Age and
sex distribution in the three groups were compared by
one-way analysis of variance (age), followed by post-hoc
tests (Tukeys honestly signicant difference), and
w-square test (sex distribution). Subsequently, the EMG
variables (POC, TC and activity) were compared among
the three groups by using a generalized linear model
(GLM) that used age, sex, and the age sex interaction
as factors in an analysis of variance. The model could
separate the effects of age and sex from those of the
actual differences in the EMG variables among the three
groups. The statistical package SAS was used.
A linear discriminant analysis was run between the
two patient groups using the individual values of muscle
activity and POC. The discriminant function analysis
(Lison, 1961; Walker and Kowalski, 1974) allows to
differentiate between two populations by the calculation
of a function L L
x
x L
y
y, where x and y are two
independent variables measured in the populations, and
L
x
, L
y
, are the relevant discriminant coefcients. The
analysis also supplies a threshold value L
0
for the
discrimination between the two populations, together
with the probability error for the classication of a new
single individual according to L
0
. The threshold L
0
is
chosen to minimize the number of mis-classied
individuals. The signicance of the discriminant analysis
was calculated with an analysis of variance (Lison,
1961). Signicance was set at 5% (Pp0:05) for all
statistical tests.
3. Results
On average, the control group was signicantly
younger than the two patient groups (analysis of
variance, Po0:001). The post hoc tests found signicant
differences among all three groups. Also, the sex
distribution among the three groups was signicantly
different (w-square test, Po0:001). To take into con-
sideration these differences, the EMG variables were
subsequently compared among the three groups using a
GLM analysis of variance.
The TMJ disorder patients had a standardized
muscular activity (four muscles pooled) during MVC
of 75 mV/mVs %, that is, their muscular electric poten-
tials were approximately one-fourth smaller when
clenching in intercuspal position (directly on the
occlusal surfaces) than on the cotton rolls (Fig. 1). In
the neck pain patient group, the MVC in intercuspal
position was made with larger muscle potentials than
the MVC on the cotton rolls, and the mean standardized
muscular activity was larger than 100% (namely,
124 mV/mVs %) (see Table 1).
In the control subjects, clenching on the cotton rolls
or clenching in intercuspal position was performed with
similar muscular activities (standardized muscular work
95 mV/mVs %). The mean activities of the three groups
were signicantly different (P 0:002) as assessed by
the GLM analysis of variance. The model found that the
effects of sex (P 0:656), age (P 0:237), and the
sex age interaction (P 0:092) were all not signicant,
while the effect of group was highly signicant
(Po0:001).
Signicant differences among the three groups were
found also for muscular symmetry (POC index, mean of
temporalis anterior and masseter muscles, Po0:001 at
the analysis of variance) and torque (TC index,
Po0:001): on average, the control subjects were the
most symmetric and those with the larger absence of
torque, while the TMJ disorder patients were the least
symmetric and those with the larger presence of torque.
The signicant effects of group (Po0:001) remained for
both variables even when the factors sex (P 0:981 for
POC, P 0:589 for TC, both not signicant), age
(P 0:004 for POC, P 0:005 for TC, both signi-
cant), and the age sex interaction (P 0:008 for POC,
signicant; P 0:945 for TC, not signicant), were
factored out by the GLM analysis of variance.
The linear discriminant analysis performed between
the two patient groups was highly signicant (F 28,
2;35 degrees of freedom, Po0:001). Using the function
L 0:1002 POC 0:0478 muscular activity,
and the threshold value L
0
13:173 for the discrimina-
tion, a single, new patient could be allocated either to
the TMJ disorder patients or neck pain patients with an
ARTICLE IN PRESS
Fig. 1. Activity standardized, symmetry (POC) and torque (TC)
indices in patients with TMJ alteration, neck pain, and control subjects
(mean+1 SD).

Po0:001 (GLM analysis of variance).
V.F. Ferrario et al. / Manual Therapy 12 (2007) 372379 375
error of 18.2%. In the present 38 patients, sensitivity of
the discrimination was 0.86, with a 0.92 specicity.
4. Discussion
Surface EMG of masticatory muscles is currently a
part of the quantitative assessment of patients in
dentistry. Among the jaw elevator muscles, the masseter
and temporalis muscles are those most often assessed in
clinical evaluations because they are the most super-
cial, and they are the only accessible to surface EMG
examination. In contrast, the medial and lateral
pterygoid muscles can be evaluated only with needle
EMG. Indeed, in the assessment of stomatognathic
dysfunction and several head disorders, the analysis of
masseter and temporalis muscles can provide quantita-
tive functional data with minimal discomfort to the
patient and without invasive or dangerous procedures
(Visser et al., 1995; Sato et al., 1998; Liu et al., 1999;
Burnett et al., 2000; Pinho et al., 2000; Ferrario et al.,
2002, 2004; Suvinen et al, 2003; Landulpho et al., 2004).
Unfortunately, as underlined by several researchers, this
simple, low cost, and fast exam also has many
limitations that must be carefully considered and
eventually removed (De Luca, 1997). For instance,
technical artifacts (the instrumental noise), the thickness
of the skin fat layer, crosstalk from different muscles.
Therefore, a correct EMG assessment should be
performed only with standardized (normalized) poten-
tials, thus removing most of biological and technical
noise (De Luca, 1997).
In the current study, to reduce patient variability, the
EMG protocol comprised a normalization record
(a MVC on cotton rolls performed just before the
recording of the actual test, i.e. with the same electrodes,
cables, and EMG apparatus, and on the same cutaneous
area) that should limit biologic and technical noise
(De Luca, 1997; Burnett et al., 2000). Indeed, the height
of the cotton roll might slightly modify the vertical
dimension (and consequently the length of muscular
bres and the interelectrode distance), but, when
clenched, it becomes so thin to make the effect
negligible. The resulting standardized EMG potentials
should therefore be determined only by the muscular
contraction as it correlates to the occlusal surfaces
(Ferrario et al., 2000, 2002).
Standardized EMG potentials can allow the measure-
ment of the actual impact of morphology on stomatog-
nathic function (Visser et al., 1995; Sato et al., 1998; Liu
et al., 1999; Burnett et al., 2000; Pinho et al., 2000;
Ferrario et al., 2002, 2006; Landulpho et al., 2004).
From the standardized electric potentials produced by
the single masticatory muscles, the muscular activity
(integrated value in time) can be calculated to assess the
actual effort made by the muscles (Sato et al., 1998;
Burnett et al., 2000; Ferrario et al., 2004, 2006).
Several previous investigations analysed the EMG
characteristic of patients with TMJ alterations. The
masticatory muscles of symptomatic TMD patients were
more hypertonic at rest, less efcient and become more
easily fatigued when compared to those of healthy
subjects matched for sex and age (Liu et al., 1999; Pinho
et al., 2000). Overall, the contraction of masticatory
muscles elicited reduced electric potentials (Visser et al.,
1995; Sato et al., 1998; Pinho et al., 2000), the
masticatory efciency was lessened, and the maximum
bite force was signicantly reduced (Sato et al., 1999). In
the current study, bite force was not measured, but
EMG activity during MVC may be considered as a
useful approximation (Fogle and Glaros, 1995; van
Kampen et al., 2002).
Chronic musculoskeletal disorders of the head and
neck are often without specic anatomical modications
that can be used for an objective differential diagnosis
(Bogduk, 1999; Visscher et al., 2002), and functional
assessments may provide useful information (Falla,
2004). A recent investigation found that patients with
craniomandibular disorders and patients with cervical
spine disorders had no differences in their head posture
ARTICLE IN PRESS
Table 1
Maximum voluntary teeth clenching in TMJ disorders and neck pain patients and control subjects (mean and standard deviation)
Unit TMJ Neck pain Control P Sex Age Interaction Group
Number 24 14 91
Age Year 35.25
a
47.67
b
20.01
c
0.001
14.06 16.71 2.00
Activity standardized mV/mVs % 74.96 124.00 94.68 0.002 0.656 0.237 0.092 0.001
22.08 36.97 19.44
Symmetry (POC) % 77.75 86.86 91.53 0.001 0.981 0.004 0.008 0.001
11.21 3.34 2.48
Torque (TC) % 84.89 91.25 94.80 0.001 0.589 0.005 0.945 0.001
10.36 1.60 1.14
P, probability value: Age, one-way analysis of variance (2;125 degrees of freedom), means with different superscript (a, b, c) differ at post-hoc test
(Tukeys honestly signicant difference); EMG variables, GLM analysis of variance factored out for sex, age, sex age interaction, actual group
difference.
V.F. Ferrario et al. / Manual Therapy 12 (2007) 372379 376
(Visscher et al., 2002). In contrast, no EMG data on
dental patients with cervical pain as the principal
symptom have been reported so far.
In the two groups of patients analysed in the present
study, the mean standardized muscular activity during
the MVC test was signicantly different from that found
in healthy controls. In the TMJ disorder patients, MVC
on cotton rolls (the standardization recording) was
made with signicantly larger EMG potentials than
MVC performed directly on the occlusal surfaces. Also,
the standardized activity of the masseter and temporalis
muscles of these patients was signicantly unbalanced,
both between sides (asymmetry) and muscular couples
(torque). Asymmetric normalized muscular activity and
unbalanced muscular couples might potentially dislo-
cate the mandible on one side, and produce more force
on one dental hemiarch and TMJ than on the
contralateral structures (Ferrario and Sforza, 1994).
Asymmetry (sides) and instability (muscular couples) in
normalized muscular activity often result from a
functionally unstable occlusion when the maxillary and
mandibular teeth contact during clenching and swallow-
ing (Ferrario et al., 1999). According to literature
references (Ferrario et al., 2000), and in comparison to
the analysed control subjects, TMJ disorder patients had
a functionally unstable occlusion (Landulpho et al.,
2004). Clenching on the cotton rolls reduced the
proprioceptive inputs from this unstable occlusion,
and allowed the patients to contract more efciently
their masticatory muscles. Even if the actual role of
occlusion in the development of signs and symptoms in
patients with TMD is still controversial, in some
patients altered occlusal conditions may be a factor in
triggering abnormal muscular activity (Ferrario et al.,
2002).
In contrast, in the neck pain patients the cotton rolls
inhibited the muscular contraction, and a larger activity
was found in intercuspal position. A possible explana-
tion of this inhibition may be that the cotton rolls
introduced a further alteration, which incremented the
actual non-occlusal and non-TMJ problem. Indeed,
their POC and TC coefcients, even signicantly lower
than that calculated in the control subjects, were inside
normal ranges (larger than 85% for the POC index, and
larger than 90% for the TC index, Ferrario et al., 2006).
To make the discrimination between the two patient
groups as simple as possible, a linear discriminant
analysis was performed. This analysis not only allows
the recognition of differences between two populations
using a linear combination of variables (two in the
current investigation) (Walker and Kowalski, 1974), but
it also supplies the probability error for the classication
of a new single individual where the same variables are
measured (Lison, 1961). The surface EMG assessment
of masseter and temporalis muscles during standardized
teeth clenching and the calculation of muscular activity
and index POC allows to allocate the new patient either
to the TMJ disorder group or neck pain group with an
error of 18.2%, i.e. the number of mis-classied
individuals would be less than 2 out of 10. Also,
sensitivity and specicity of the test, as assessed from the
current group of patients, were good.
It has to be mentioned that the analysed individuals
represent a convenience sample, and the extrapolation
of the present results to a wider population should be
done with caution. Furthermore, the control group was
younger than the two patient groups, and the male:
female ratio was different. The GLM analysis of
variance used to compare the EMG variables of the
three groups took these differences into consideration,
and found signicant effects of age for POC and TC
indices, and a signicant age sex interaction for POC.
In contrast, the effect of sex was never signicant. Even
considering these effects, the differences in EMG indices
remained highly signicant among the three groups.
Indeed, the use of standardized potentials should reduce
inter-subject variability (De Luca, 1997; Burnett et al.,
2000; Ferrario et al., 2000).
Literature reports on the effects of sex and age on
EMG variables are scanty. In accord with the current
ndings, no signicant sex-related differences in normal-
ized EMG potentials recorded during MVC were
reported by Ferrario et al. (2000, 2006) for young
healthy subjects. In subjects with a good occlusion,
aging seems to have minor effects on the normalized
EMG indices: in control subjects aged 53 years on
average, a recent study reported a mean standardized
activity of 104.9 mV/mVs % (SD 28.9) mean POC of
87% (SD 0.9) and mean TC of 90.8% (SD 0.4) (Ferrario
et al., 2004). In contrast, Ueda et al. (2002) found that
the masticatory muscles were more fatigable in women
than in men, and some effect of age was found on
muscle recruitment during incisal biting at different bite
forces by Fogle and Glaros (1995). Indeed, in view of the
different experimental settings and analysed variables,
it is difcult to make conclusive assertions on this
problem.
5. Conclusion
Surface EMG of the right and left masseter and
temporalis anterior muscles, performed with a well-
dened protocol, provided standardized data that were
used for a quantitative assessment of two patient
groups. The test was simple, low cost, fast and non-
invasive, it provoked no discomfort to the patients and
it had no side effects.
A discriminant analysis of the standardized data
obtained from the EMG test allowed a possible
differentiation between patients with a TMJ disorder
and patients with a neck pain problem. The two groups
ARTICLE IN PRESS
V.F. Ferrario et al. / Manual Therapy 12 (2007) 372379 377
of patients need different treatments, and a quick
decision between the two diagnoses may allow a prompt
solution of their disturbance. Patients with a low
standardized muscular activity and a reduced rightleft
muscular symmetry during MVC had more probably a
TMJ disorder, and the rst, immediate step of their
treatment included the preparation of a stabilization
splint (Ferrario et al., 2002). Indeed, among the several
treatments used in the TMD patients, occlusal splints
are believed to be benecial in a large part of cases: they
are a conservative and reversible therapy, and can
reduce pain in most cases (Nemcovsky et al., 1992; Ash
and Ramfjord, 1995; Ferrario et al., 2002).
In contrast, patients with a high standardized
muscular activity and a nearly normal rightleft
muscular symmetry during MVC had most likely a
neck problem, and an occlusal splint will probably be
not benecial for them. Their assessment should be
performed in collaboration with a physiotherapist and/
or a chiropractor, and a manual treatment together with
an appropriate exercise programme may be benecial
(Alcantara et al., 2002; DeVocht et al., 2003; Falla,
2004). An occlusal splint may be used only to avoid the
biomechanical inuences of occlusal stress on the neck.
Acknowledgements
The precious statistical assistance of Mr. Bruno
Lovecchio, B.Sc, and the expert secretarial assistance
of Ms Cinzia Lozio are gratefully acknowledged.
References
Alcantara J, Plaugher G, Klemp DD, Salem C. Chiropractic care of a
patient with temporomandibular disorder and atlas subluxation.
Journal of Manipulative and Physiological Therapeutics
2002;25:6370.
Ash MM, Ramfjord S. Occlusion, 4th Ed. Philadelphia: WB Saunders
Co; 1995. p. 2334.
Bogduk N. The neck. Baillie` res of Clinical Rheumatology
1999;13:26185.
Burnett CA, Fartash L, Murray B, Lamey P-J. Masseter and
temporalis muscle EMG levels and bite force in migraineurs.
Headache 2000;40:8137.
Ciancaglini R, Testa M, Radaelli G. Association of neck pain with
symptoms of temporomandibular dysfunction in the general adult
population. Scandinavian Journal of Rehabilitation Medicine
1999;31:1722.
De Luca CJ. The use of surface electromyography in biomechanics.
Journal of Applied Biomechanics 1997;13:13563.
DeVocht JW, Long CR, Zeitler DL, Schaeffer W. Chiropractic
treatment of temporomandibular disorders using the activator
adjusting instrument: a prospective case series. Journal of
Manipulative and Physiological Therapeutics 2003;26:4215.
De Wijer A, Steenks MH, Bosnian F, Helders PJM, Faber J.
Symptoms of the stomatognathic system in temporomandibular
and cervical spine disorders. Journal of Oral Rehabilitation
1996;23:73341.
Eriksson P-O, Zafar H, Haggman-Henrikson B. Deranged jawneck
motor control in whiplash-associated disorders. European Journal
of Oral Sciences 2004;112:2532.
Falla D. Unravelling the complexity of muscle impairment in chronic
neck pain. Manual Therapy 2004;9:12533.
Ferrario VF, Sforza C. Biomechanical model of the human mandible
in unilateral clench: distribution of the TMJ reaction forces
between the working and balancing side. Journal of Prosthetic
Dentistry 1994;72:16976.
Ferrario VF, Sforza C, Serrao G, Colombo A, Schmitz JH. The effects
of a single intercuspal interference on EMG characteristics of
human masticatory muscles during maximal voluntary teeth
clenching. Cranio-Journal of Craniomandibular Practice
1999;17:1848.
Ferrario VF, Sforza C, Colombo A, Ciusa V. An electromyographic
investigation of masticatory muscles symmetry in normo-occlusion
subjects. Journal of Oral Rehabilitation 2000;27:3340.
Ferrario VF, Sforza C, Tartaglia GM, Dellavia C. Immediate effect of
a stabilization splint on masticatory muscle activity in tempor-
omandibular disorder patients. Journal of Oral Rehabilitation
2002;29:8105.
Ferrario VF, Tartaglia GM, Maglione M, Simion M, Sforza C.
Neuromuscular coordination of masticatory muscles in subjects
with two types of implant supported prostheses. Clinical Oral
Implants Research 2004;15:21925.
Ferrario VF, Tartaglia GM, Galletta A, Grassi GP, Sforza C. The
inuence of occlusion on jaw and neck muscle activity: a surface
EMG study in healthy young adults. Journal of Oral Rehabilita-
tion 2006;33:3418.
Fogle LL, Glaros AG. Contributions of facial morphology, age, and
gender to EMG activity under biting and resting conditions: a
canonical correlation analysis. Journal of Dental Research
1995;74:1496500.
Gross AR, Haines T, Thomson MA, Goldsmith C, Mcintosh J.
Diagnostic tests for temporomandibular disorders: an assessment
of the methodologic quality of research reviews. Manual Therapy
1996;1:2507.
John MT, Miglioretti DL, LeResche L, Von Korff M, Critchlow CW.
Widespread pain as a risk factor for dysfunctional temporoman-
dibular disorder pain. Pain 2003;102:25763.
Landulpho AB, E Silva WAB, E Silva FA, Vitti M. Electromyographic
evaluation of masseter and anterior temporalis muscles in patients
with temporomandibular disorders following interocclusal appli-
ance treatment. Journal of Oral Rehabilitation 2004;31:958.
Lee WY, Okeson JP, Lindroth J. The relationship between forward
head posture and temporomandibular disorders. Journal of
Orofacial Pain 1995;9:1617.
Lison L. Statistica applicata alla biologia sperimentale. Milano, Italy:
CEA; 1961. p. 31329.
Liu ZJ, Yamagata K, Kasahara Y, Ito G. Electromyographic
examination of jaw muscles in relation to symptoms and occlusion
of patients with temporomandibular joint disorders. Journal of
Oral Rehabilitation 1999;26:3347.
Nemcovsky CE, Gazit E, Serfati V, Gross M. A comparative study of
three therapeutic modalities in a temporomandibular disorder
(TMD) population. Journal of Craniomandibular Practice
1992;10:14855.
Pinho JC, Caldas FM, Mora MJ, Santana-Penin U. Electromyo-
graphic activity in patients with temporomandibular disorders.
Journal of Oral Rehabilitation 2000;27:98590.
Sato S, Ohta M, Goto S, Kawamura H, Motegi K. Electromyography
during chewing movement in patients with anterior disc displace-
ment of the temporomandibular joint. International Journal of
Oral and Maxillofacial Surgery 1998;27:2747.
Sato S, Ohta M, Sawatari M, Kawamura H, Motegi K. Occlusal
contact area, occlusal pressure, bite force, and masticatory
ARTICLE IN PRESS
V.F. Ferrario et al. / Manual Therapy 12 (2007) 372379 378
efciency in patients with anterior disc displacement of the
temporomandibular joint. Journal of Oral Rehabilitation
1999;26:90611.
Suvinen TI, Reade PC, Kononen M, Kemppainen P. Vertical jaw
separation and masseter muscle electromyographic activity: a
comparative study between asymptomatic controls & patients with
temporomandibular pain & dysfunction. Journal of Oral Rehabi-
litation 2003;30:76572.
Ueda HM, Kato M, Saifuddin M, Tabe H, Yamaguchi K, Tanne K.
Differences in the fatigue of masticatory and neck muscles between
male and female. Journal of Oral Rehabilitation 2002;29:57582.
van Kampen FMC, van der Bilt A, Cune MS, Bosman F. The
inuence of various attachment types in mandibular implant-
retained overdentures on maximum bite force and EMG. Journal
of Dental Research 2002;81:1703.
Visscher CM, De Boer W, Lobbezoo F, Habets LLMH. Is there a
relationship between head posture and craniomandibular pain?
Journal of Oral Rehabilitation 2002;29:10306.
Visser A, Kroon GW, Naeije M, Hansson TL. EMG differences
between weak and strong myogenous CMD patients and healthy
controls. Journal of Oral Rehabilitation 1995;22:42934.
Walker GF, Kowalski CJ. Computer aided diagnosis of craniofacial
abnormalities. Medinfo 1974;74:5537.
ARTICLE IN PRESS
V.F. Ferrario et al. / Manual Therapy 12 (2007) 372379 379