Biofeedback-based Cognitive-Behavioral Treatment
Compared With Occlusal Splint for
A Randomized Controlled Trial
Meike C. Shedden Mora, PhD,* Daniel Weber, DDS,w Andreas Neff, PhD,z
and Winfried Rief, PhD*
Objectives: Cognitive-behavioral treatment has proven efficacy
for chronic temporomandibular disorder (TMD). However, most
patients receive dental treatment that may not address psycho-
logical comorbidities often present in TMD. The aim of the present
study was to evaluate the efficacy of biofeedback-based cognitive-
behavioral treatment (BFB-CBT) versus dental treatment with oc-
clusal splint (OS). Moreover, changes in nocturnal masseter muscle
activity (NMMA) were investigated.
Methods: Fifty-eight patients with chronic TMD were randomly
assigned to receive either 8 weekly sessions of BFB-CBT or 8 weeks
of OS treatment. Diagnoses were established using Research
Diagnostic Criteria for TMD. Pain intensity and disability were
defined as primary outcomes. Secondary outcomes included emo-
tional functioning, pain coping, somatoform symptoms, treatment
satisfaction, and adverse events. NMMA was assessed during 3
nights pretreatment and posttreatment with portable devices. Fol-
low-up assessment took place 6 months after the treatment.
Results: Both treatments resulted in significant reductions in pain
intensity and disability, with similar amounts of clinically mean-
ingful improvement (45% for BFB-CBT and 48% for OS). Patients
receiving BFB-CBT showed significantly larger improvements
in pain coping skills. Satisfaction with treatment and ratings
of improvement were higher for BFB-CBT. Effects were stable over
6 months, and tended to be larger in the BFB-CBT group for all
outcomes. No significant changes were observed in NMMA.
Discussion: The fact that BFB-CBT resulted in larger improve-
ments in pain coping skills, and was well accepted by the patients,
underlines the importance and feasibility of psychological treat-
ments in the clinical management of TMD.
Key Words: temporomandibular disorders, randomized controlled
trial, biofeedback, cognitive-behavioral treatment, occlusal splint
(Clin J Pain 2013;29:1057–1065)
cles and joint, limited mandibular motion, and joint sounds.1
emporomandibular disorders (TMD) are characterized
by pain or dysfunction in the temporomandibular mus-
Cognitive-behavioral therapy (CBT) for TMD successfully
reduces pain, pain-related disability and increases pain cop-
ing skills by enhancing adaptive cognitions and behaviors.2–6
In order to address pathophysiologic mechanisms, CBT has
been successfully combined with biofeedback, which leads to
substantial long-term improvements in pain and symp-
toms.7,8Biofeedback for TMD aims to improve proprio-
ceptive awareness of oral parafunctions and to re-
duce mandibular muscle hyperactivity and sleep bruxism.9
Patients with TMD show muscular hyperactivity,10and the
reduction of oral parafunctions leads to symptom improve-
ments.7,9Nocturnal training with portable electromyography
(EMG) alarm systems successfully reduces sleep bruxism,
which might also improve TMD symptoms.11
Despite the success of CBT, dental treatment with
occlusal splint (OS) remains far more common in the clinical
management of TMD.12,13A recent meta-analysis has con-
firmed a modest effect in reducing TMD pain compared with
nonoccluding appliances and no treatment.14Rationales for
OS use include reducing oral parafunctions and load on
temporomandibular structures, and modifying occlusal force
distribution.13,14Several studies show a decrease in sleep
bruxism during OS treatment, whereas others report no
change or even an increase.11,15,16
Few studies have compared psychological approaches
with dental treatment. In terms of pain reduction, CBT or
biofeedback seem to be equally effective or superior com-
pared with OS appliances or medical treatment.17–19In
clinical practice, patients with TMD mostly seek dental
treatment in the first place, and are referred to other spe-
cialists with symptoms being heterogenic. Psychological
distress often remains unaddressed that might contribute
to chronic development of the condition.20Therefore, the
comparison of dental and psychological treatment is highly
relevant. However, former studies have a number of impor-
tant shortcomings. The majority of studies have assessed
treatment success solely in terms of pain reduction. However,
following recommendations of the Initiative on Methods,
Measurement, and Pain Assessment in Clinical Trials
(IMMPACT),21other important outcomes such as activity
interference, emotional functioning, pain coping, treatment
satisfaction, and possible adverse effects should be considered.
Moreover, the lack of standardized diagnostic criteria, such as
the Research Diagnostic Criteria for TMD (RDC/TMD)22
constitutes a major shortcoming of former studies.17,19
The aim of the present study was to evaluate a short
biofeedback-based cognitive-behavioral treatment (BFB-
CBT) compared with dental treatment with OS. First, it
was hypothesized that (1) improvements in pain intensity
Received for publication April 22, 2012; revised December 22, 2012;
accepted December 27, 2012.
From the Departments of *Clinical Psychology and Psychotherapy;
wProsthetic Dentistry; and zOral and Maxillofacial Surgery,
Philipps University of Marburg, Marburg, Germany.
The authors declare no conflict of interest.
A.N. and W.R. contributed equally.
Reprints: Meike C. Shedden Mora, PhD, Department of Clinical
Psychology and Psychotherapy, Philipps University of Marburg,
Gutenbergstr. 18, 35032 Marburg, Germany (e-mail: m.shedden@
Copyrightr2013 by Lippincott Williams & Wilkins
Clin J Pain?Volume 29, Number 12, December 2013www.clinicalpain.com|1057
and functional limitations would occur in both treatment
conditions, but (2) BFB-CBT would be more effective in
improving secondary outcomes such as emotional func-
tioning, pain coping, and somatization. Second, we aimed
to investigate whether reductions in nocturnal masseter
muscle activity (NMMA) occurred during treatment, and
whether clinical improvements are also reflected in reduced
nocturnal masseter muscle activity.
MATERIALS AND METHODS
Patients seeking treatment for chronic TMD at the
Marburg Dental School, Philipps University of Marburg,
Germany were recruited for this study by the Department
of Prosthetic Dentistry and the Department of Oral and
Maxillofacial Surgery. In addition, dental clinics in Mar-
burg were informed about the study and referred eligible
patients to the Department of Prosthetic Dentistry. Patient
recruitment and follow-up assessment took place from
August 2008 to April 2011.
Study inclusion criteria were as follows: (1) a painful
axis I TMD diagnosis according to the Research Diagnostic
Criteria for Temporomandibular Disorders (RDC/TMD),22
in other words group I (myofascial pain), or group III
(arthralgia, arthritis, arthrosis) or both; patients could also
have a group II diagnosis (disk displacement), but a painless
group II diagnosis was not sufficient for study inclusion; (2)
pain present for at least 3 months; (3) age between 18 and 70.
Patients were excluded for any of the following: (1)
presence of an OS already matching to our standards, for
example an OS as described below (patients could be
included if they currently used a splint that did not meet our
standards, such as a non-OS); (2) need for further diagnostic
investigation or need for dental/maxillofacial treatment, as
judged by a specialized dentist; (3) other major chronic pain
conditions predominant in disability, for example chronic
low back pain or headache, as assessed in the diagnostic
interview; (4) major medical or psychiatric conditions that
would interfere with the ability to participate.
The flow of patients through the study is shown
in Figure 1. A total of 103 patients reported interest in
study participation. Thirty-three patients were not eligible
for study participation. Twelve patients refused to partic-
ipate. The main reasons for declining participation were the
unwillingness to be randomized, time commitment for
treatment, and distance to the clinic. Finally 58 patients
were enrolled in the study (Fig. 1).
The study was approved by the ethics committee for
medical research at the University of Marburg. The trial
was registered at Clinical Trials (http://www.clinicaltrials.
gov; trial code NCT00769561). We followed CONSORT
guidelines for reporting randomized trials.23
Patients were informed about the study and gave their
written informed consent. RDC/TMD axis I diagnoses
were established by 2 dentists who were trained and cali-
brated in RDC/TMD diagnostic process. The presence of
psychiatric disorders was verified by a trained clinical psy-
chologist with structured clinical interview for DSM-IV.24
Patients were asked to complete questionnaires. Some of
the outcome measures chosen for this study differed from
the original RDC/TMD axis II recommendations (see
below). The proposed measures are more useful for screening
purposes, however, in the context of this randomized con-
trolled trial it was important to apply dimensional outcome
measures sensitive to change.25,26NMMA was recorded at
home during 3 consecutive nights (see below).
Patients were reassessed after completing treatment.
They were asked to complete questionnaires, to record
NMMA, and underwent examination according to RDC/
TMD by another trained and calibrated dentist/maxillofacial
surgeon blind to subject status. Follow-up assessment took
place 6 months after the end of treatment. Patients were
asked to complete questionnaires and to return them by
Primary Outcomes: Pain and Disability
Characteristic pain intensity was calculated by aver-
aging ratings of current pain, average pain, and worst pain
in the past month on a numeric rating scale from 0 to 10, as
recommended by RDC/TMD.22Following IMMPACT
recommendations we compared the 2 groups in terms of
clinically meaningful improvement, defined as reduction in
characteristic pain intensity of at least 30% from baseline to
posttreatment (responder). An increase of pain intensity of
at least 30% was defined as an aggravation of pain.27
Pain-related disability was assessed using the Pain
Disability Index (PDI).28The PDI is a brief self-rating scale
that assesses the level of pain-related disability in 7 areas of
daily life (eg, social activity, self-care) on a 0 to 10 numeric
rating scale. The Cronbach a was 0.87 in the current
Jaw use limitations were measured using the Jaw Dis-
ability List (JDL) from the RDC/TMD. The JDL asks the
patient to rate interference with 11 oral activities, for
example chewing or talking. We used an 11-point numeric
rating scale instead of ratings of “yes” and “no”. The
Cronbach a was 0.86 in the current sample.
Depressive symptoms were measured using the Centre
for Epidemiological Studies Depression scale (CES-D).29
The CES-D asks for the frequency of 20 symptoms of
depression during the past week on a scale ranging from 0
(less than 1d) to 3 (5 to 7d). It is suitable for use in chronic
pain patients as it relies less on physical symptoms of
depression than do other measures. The Cronbach a level in
the current sample was 0.89.
General anxiety symptoms were assessed using the
7-item scale from the Patient Health Questionnaire (GAD-
7).30The GAD-7 asks for anxiety symptoms during the past
month on a 1 (not at all) to 2 (more than half of the days)
rating scale. The Cronbach a level in the current sample
Cognitive and behavioral pain coping strategies were
assessed with Coping Strategies Scale from the German
Pain Coping Questionnaire (FESV).31The scale asks for the
use of 24 cognitive (eg, cognitive restructuring) and
behavioral (eg, use of relaxation techniques) strategies for
coping with pain on a scale ranging from 1 to 6. A sum
score was used with higher scores indicating more adaptive
coping. The Cronbach a level was 0.80.
Shedden Mora et alClin J Pain?Volume 29, Number 12, December 2013
r2013 Lippincott Williams & Wilkins
Somatoform Symptoms and TMD-related Symptoms
Somatoform complaints during the past week were
assessed using the Screening for Somatoform Symptoms
(SOMS-7).32Thirty-two medically unexplained symptoms
(29 for male patients) representing DSM-IV criteria for
somatization disorder were rated on a 0 to 4 scale. A sum
score was calculated with higher scores indicating higher
intensity and burden of somatoform complaints. In the
current sample, the Cronbach a level was 0.88.
TMD-related symptoms, such as jaw pain, toothache,
or dizziness, were measured using a 41-item TMD symptom
list.33Following the SOMS-7 scale, intensity of symptoms
experienced during the past week was rated from 0 (not at
all) to 4 (very high intensity). A sum score was built with
higher scores indicating higher intensity of TMD-related
symptoms. The TMD symptom list has not been evaluated
previously; however, large bivariate correlations with
somatization (Pearson r=0.79), medium to large correla-
tions with pain intensity (r=0.48), and medium correla-
tions with depression (r=0.37) and anxiety (r=0.27)
provide evidence of good convergent and divergent validity.
The Cronbach a level in the current sample was excellent
Participant ratings of global improvement were meas-
ured with the Patient Global Impression of Change scale
(PGIC).34The PGIC is a single-item rating of improvement
with treatment on a 7-point scale ranging from “very much
improved” to “very much worse” with “no change” at the
Satisfaction with treatment was measured on a 13-item
rating scale adapted from a randomized controlled trial for
chronic tinnitus.35A mean score ranging from 0 to 5 was
built with higher scores indicating higher treatment sat-
isfaction. The Cronbach a level for the scale was 0.90.
Adverse events: Patients were asked with the use of an
open question whether they had experienced any adverse
events or aggravation of symptoms.
Demographic and health history measures: Demo-
graphic variables, general health state (on a scale from 1
“excellent” to 5 “very poor”), the number of treatments
received for TMD, the number of healthcare providers
consulted for TMD during the past 6 months, and medi-
cation intake were assessed.
Assessment of NMMA
NMMA was recorded during 3 consecutive nights with
single-channel EMG ambulatory recording devices (Loguva
Brux, Haynl Elektronik GmbH, Scho ¨ nebeck, Germany). A
detailed description of data assessment and analyses has
been published elsewhere.36Participants were precisely
instructed to use the devices. They were asked to note time
periods of wakefulness or any problems in a protocol and to
contact the study investigator in case of difficulties.
The analyses of NMMA were performed using a cus-
tom program (LabVIEW 9.0, National Instruments Cor-
poration, Austin, TX). Mean number and duration of
bursts per hour sleep were calculated as indicators of
masseter muscle activity for each nocturnal recording.
A burst was defined as a supra-threshold EMG signal
(above 10mV) of 0.25- to 2.0-second duration according
FIGURE 1. Patient flow (CONSORT diagram).
Clin J Pain?Volume 29, Number 12, December 2013 Cognitive-behavioral Treatment for TMD
r2013 Lippincott Williams & Wilkins www.clinicalpain.com|1059
to Lavigne et al.37NMMA variables were normalized with
a logarithmic transformation. The stability of EMG record-
ings between the 3 nights was good (Pearson correlation
r=0.68 to 0.83).
After baseline assessment, patients were randomly
assigned to receive BFB-CBT or OS treatment. Random
assignment to conditions was generated by a researcher not
involved in the study with the use of randomization soft-
ware (GraphPad Software Inc., La Jolla, CA), and assign-
ment was concealed in closed envelopes. The treatment
duration was approximately 8 weeks for both conditions.
The BFB-CBT consisted of 8 weekly sessions with 50-
minute duration. Each session contained complementary
cognitive-behavioral and biofeedback elements, with com-
parable time and emphasis on both elements. Treatment
was provided by 1 of 3 trained Master’s level psychologists
who had at least 1 year of clinical experience in biofeedback
and CBT for chronic pain conditions. Therapists received a
thorough training by the first author of this article. Treat-
ment fidelity was controlled through regular supervision
sessions (every 2 to 4 sessions), provided by a specialized
clinical psychologist with more than 15-year experience in
biofeedback and CBT for chronic pain conditions including
TMD. A treatment manual was provided to the therapist
that contained detailed instructions on the treatment
rationale, materials to use, and assignment of exercises (the
manual in German language can be requested from the
authors). Session content was largely the same for all
patients, but exact specification depended on patients’
individual needs, their own causal model, and level of
knowledge. Patients received a booklet with the informa-
tion discussed in the session, home assignments, as well as a
relaxation audio CD and a video with jaw exercises.
The biofeedback device used during the sessions was
a 10-channel FlexComp Infinity (Thought Technology,
Montreal, QC, Canada). Activity in masseter, temporalis,
frontalis, and trapezius muscles was assessed by surface
EMG biofeedback, as these sites are involved in oral par-
afunctions and TMD pain.9Feedback of skin conductance
level, finger temperature and breathing activity was used as
indicators of autonomic arousal.
The main aim of the BFB-CBT was to reduce pain and
to improve pain-related disability. Biofeedback elements
targeted to reduce oral parafunctions by improving pro-
prioceptive awareness, to reduce jaw muscle tension in
specific muscle sites, to reduce NMMA, and to improve
general relaxation skills. Cognitive-behavioral elements
aimed to improve pain coping skills and feelings of control.
The intervention was based on treatment protocols for
TMD pain that have shown effectiveness in previous stud-
ies.4,19In the first 2 sessions, patients were educated about
symptoms and causes of their TMD, and a personal
etiological model was established. The importance of psy-
chophysiological relations was demonstrated with a stress-
relaxation-test using biofeedback. In the third session, a
habit reversal technique was used to increase awareness of
tooth contacting habits and to implement a relaxed jaw
posture.38In the fourth session, patients received an EMG
biofeedback home trainer (Loguva Brux) for diurnal and
nocturnal biofeedback training.7,11Diurnal biofeedback
training consisted in daily exercises over a period of 2
weeks. Patients were asked to practice in stressful situations
associated with increased jaw muscle tension. The device
was programmed to emit an alarm tone when EMG levels
exceeded a preset threshold. For nocturnal training, they
were asked to use the home trainer every night for 2 weeks.
At night, individual threshold was established as 20% of
a strong biting force, and the device emitted an alarm tone
if EMG levels exceeded the threshold for more than
4-second duration. Patients were asked to fill in systematic
reports of their training. Session 5 aimed to improve
relaxation skills by teaching biofeedback-based progressive
muscle relaxation. Furthermore, the BFB-CBT protocol
focused on training patients in pain management using
problem solving skills training (Session 6), thought chal-
lenging (Session 7), as well as relapse prevention (Session 8).
Dental Treatment With OS
Flat plane acrylic OS were individually fabricated for
each patient. Their proposed therapeutic action consisted
of eliminating centric or eccentric premature contacts,
restoring missing posterior support and, when indicated to
reduce abrasion, guiding the mandible into a more vertical
movement pattern. The treatment comprised 5 sessions
of approximately 20 to 40 minutes. First, maxillary and
mandibular dental impressions (session 1) as well as an
arbitrary face bow registration (session 2) were taken. Stone
casts were mounted on a semiadjustable articulator in
therapeutic position after appropriate occlusal registration.
Fabricated to fit either the maxilla or the mandible, the OS
covered all anterior and posterior teeth with 1.5mm
thickness between posterior teeth, and provided simulta-
neous uniform contacts in the premolar and molar region.
Anterior teeth were in light point contact or were discluded
slightly; a slight anterior and canine rise was achieved in
excursive or protrusive functions. After fabrication and ad-
justment (Session 3), all patients were instructed to use the
splint every night over a period of 8 weeks. In case of
present diurnal parafunctional activities patients were also
asked to wear the splint during daytime, which was the case
in 10 patients. If muscular etiology was predominant, which
was the case for most patients, they were given instructions
to avoid contact with the splint and tooth contact during
daytime. All patients were instructed to return to the clinic
for reexamination and possible adjustment of the splint
after 1 week (Session 4) and 5 weeks (Session 5). They were
treated by 2 dentists specialized in functional dental treat-
ment. Adherence to splint use was assessed after treatment.
All statistical analyses were performed using SPSS
Statistics 19.0.0 for Windows. Multiple imputation using
NORM (version 2.03) was used to replace single missing
values of questionnaire data with imputed values for
existing time points. For missing time points, an intent-to-
treat approach with the last value carried forward was used
that included all randomized patients with available base-
line data starting treatment.
As recommended by IMMPACT, characteristic pain
intensity and disability were defined as primary outcomes.
Secondary outcomes included depression, anxiety, pain cop-
ing, somatoform symptoms, and TMD-related symptoms.
To determine general efficacy and differences between
groups on primary and secondary outcomes, 2 repeated-
measures multivariate analyses of variance (MANOVAs)
were calculated, with treatment condition (BFB-CBT and
Shedden Mora et alClin J Pain?Volume 29, Number 12, December 2013
r2013 Lippincott Williams & Wilkins
OS) and time (pre, post, follow-up) as factors. In case
of a significant main effect, repeated-measures analyses of
variance (ANOVAs) were calculated for each of the 8
outcome variables, correcting for multiple comparisons
using Bonferroni-Holm procedure. In ANOVA, planned
contrasts (pre vs. post, post vs. follow-up) were calculated
to determine when changes took place. Effect sizes and
confidence intervals were computed for both groups for
changes from pretreatment to posttreatment and from
pretreatment to follow-up, using the formula by McGaw
The ratio of patients clinically meaningful improved
was compared between groups using w2test. Participants’
ratings of global improvement with the PGIC were split up
into those who reported at least some improvement (cor-
responds to rating of 5 to 7), no change (rating of 4), and
feeling worse (rating of 1 to 3). Ratios were compared with
w2test. Treatment satisfaction was compared using t tests.
Sample Size Estimation
Required sample size was estimated using G*Power
(version 3.0.10) for MANOVA. To detect a minimum effect
of f=0.25 for global treatment effects for primary out-
comes with conventional a and power levels (a=0.05 and
b=0.80, respectively), a minimum total sample size of 48
was required. Considering discontinuation and dropout,
our aim was to recruit a sample of at least 60 participants.
Characteristics of the study sample are shown in
Table 1. Participants were predominantly characterized by
long illness duration and high levels of healthcare utilization.
The groups were comparable on all sociodemographic vari-
ables except that patients in the BFB-CBT group had
received more treatments for their condition (T=2.32,
P<0.05) and had consulted more healthcare professionals in
the past 6 months (T=2.12, P<0.05; data only available
for 50 patients). No baseline differences in primary and sec-
ondary outcome variables were found. A distribution of
RDC/TMD diagnoses is shown in Table 2, with the majority
of patients diagnosed with myofascial pain.
General Efficacy and Maintenance of Results
Table 3 displays mean, SDs, effect sizes, and confidence
intervals for primary and secondary outcome measures.
Repeated-measures MANOVAs were calculated with the
number of received treatments as covariate. As information
on healthcare use was only available for 50 patients, it was
not included as a covariate. For the primary outcomes pain
intensity, pain-related disability and jaw use limitations,
repeated-measures MANOVA showed a significant time
effect (Pillai’s trace V=0.30, F6,49=3.43, P<0.01, partial
Z2=0.30), but no significant effect for treatment condition
(Pillai’s trace V=0.03, F3,52=0.46, P=0.71, partial
Z2=0.03). The condition-time interaction failed levels of
significance (Pillai’s trace V=0.20, F6,49=2.02, P=0.08,
partial Z2=0.20). Thus, BFB-CBT and OS did not differ
significantly in the main outcomes. Subsequent repeated-
measures ANOVAs revealed significant improvements in
both groups on the primary outcomes characteristic pain
intensity (F2,53=19.61, P<0.001, partial Z2=0.27), pain-
related disability (F2,53=6.56, P<0.01, partial Z2=0.11),
and jaw use limitations (F2,53=7.13, P<0.01, partial
Z2=0.12). Repeated-measures contrasts showed that all
significant changes occurred from pretreatment to post-
treatment and remained stable to follow-up.
For secondary outcomes, repeated-measures MAN-
OVA was calculated with TMD-related symptoms, soma-
toform symptoms, depressive symptoms, anxiety, and pain
coping. Using Pillai’s trace, there was a significant time
effect (Pillai’s trace V=0.34, F10,45=2.22, P<0.05, par-
tial Z2=0.34), but no significant effect for treatment con-
dition (V=0.16; F5,50=1.85, P=0.12, partial Z2=0.21).
The condition-time interaction failed to reach level of sig-
nificance (V=0.30; F10,45=1.87, P=0.08, partial Z2=
0.30). Subsequent repeated-measures ANOVAs revealed
significant improvements in both groups for all secondary
outcomes: TMD-related symptoms (F2,53=18.71, P<0.001,
partial Z2=0.26), somatoform symptoms (F2,53=4.96,
P<0.01, partial Z2=0.08), depressive symptoms (F2,53=
4.43, P<0.05, partial Z2=0.08), anxiety (F2,53=14.96,
P<0.001, partial Z2=0.22), and pain coping (F2,53=16.10,
P<0.001, partial Z2=0.23). All significant changes occur-
red from pretreatment to posttreatment and remained stable
to follow-up, as indicated by repeated-measures contrasts.
There was a significant condition-time interaction and
group effect for pain coping (F2,53=6.89, P<0.01, partial
Z2=0.11; and F2,53=8.41, P<0.01, partial Z2=0.14,
respectively), indicating that patients in the BFB-CBT group
reported larger improvements in pain coping skills than
TABLE 1. Sample Characteristics
Age (±SD)36.3 (±13.4) 34.3 (±12.5)T=0.54,
Education at least 13y62.1%44.4%
General Health State
(1 “excellent” to 5
“very poor”) (±SD)
Diagnosis of psychiatric
Pain duration in
Mean number of
for TMD (±SD)
Number of healthcare
during past 6mo
Number of patients taking medication (%)
3.1 (±1.0)3.5 (±0.8)
41.9 (±55.6) 40.3 (±43.1)
4.1 (±2.6)2.6 (±2.2)
6.4 (±6.4) 3.2 (±3.6)T=2.12,
7 (24.1%)7 (25.9%)
Nonopioid analgesics5 (17.2%)3 (11.1%)
3 (10.3%)3 (11.5%)
BFB-CBT indicates biofeedback-based cognitive-behavioral treatment;
OS, occlusal splint.
Clin J Pain?Volume 29, Number 12, December 2013Cognitive-behavioral Treatment for TMD
r2013 Lippincott Williams & Wilkinswww.clinicalpain.com|1061
patients receiving OS. The groups did not differ significantly
on the other outcomes.
Pretreatment to posttreatment and pretreatment to fol-
low-up effect sizes revealed medium to large effects in
the BFB-CBT group and small to medium effects in the OS
group, according to convention.40Effect sizes were larger for
BFB-CBT compared with OS for all outcome variables,
though a significant difference was only found for pain coping
In terms of clinically meaningful improvement, 13
patients (44.8%) in the BFB-CBT group and 13 patients
(48.1%) in the OS group were classified as responders.
Accordingly, 2 patients (6.9%) in the BFB-CBT group and
3 (11.1%) patients in the OS group reported an aggravation
of pain of at least 30%. The rates did not differ significantly
between groups (w2=0.49, P=0.78), suggesting that
patients responded well to both treatments (Table 3).
Ratings of Global Improvement
For participants’ ratings of global improvement, more
patients in the BFB-CBT group reported some improve-
ment (26 patients [88.9%] in the BFB-CBT group, 16
patients [59.3%] in the OS group; w2=6.93, P<0.05).
Two patients (7.4%) in the BFB-CBT group reported no
change, compared with 10 patients (37%) in the OS group.
In each group, 1 patient reported a change for the worse.
Acceptability of Treatments and Adverse Events
Satisfaction with treatment was significantly higher in
the BFB-CBT group (mean rating 4.0±SD=0.55 vs.
3.07±SD=0.70 on a scale from 0 to 5) (T=5.23,
P<0.001). Three patients in the BFB-CBT condition
reported an aggravation of symptoms, namely pain in teeth,
jaw pain, and limited jaw motion. In the OS condition,
7 patients reported change for the worse, namely higher
muscle tension, pain in head, back, ears and teeth, jaw
limitations at chewing, and tinnitus. The ratios did not
differ significantly (w2=1.62, P=0.20).
Adherence to splint use in the OS group was assessed
at the end of treatment, and patients were asked how often
they used the splint since the beginning of treatment.
Twelve patients reported that they had used the splint every
night, 14 patients reported the use during 3 to 6 nights
TABLE 3. Results of the Randomized Controlled Trial: Primary and Secondary Outcome Measures
Effect Size (g) (ITT)
Effect Size (g) (ITT)
Pain intensity (CPI) (range 0-10)
Pain disability (PDI) (range 0-70)
Jaw use limitations (JDL) (range 0-110)
TMD-related symptoms (RDC) (range 0-164)
OS 37.44 (17.28)
BFB-CBT 47.69 (24.12)
Somatoform symptoms (SOMS) (range 0-128)
OS 11.81 (7.57)
Depression (ADS-L) (range 0-60)
Anxiety (GAD-7) (range 7-21)
OS 14.15 (2.35)
Pain Coping (FESV) (range 24-144)
0.21 (?0.18 to 0.59)
0.16 (?0.22 to 0.53)
0.38 (?0.01 to 0.77)
0.67 (?26 to 1.07)
0.06 (?0.32 to 0.44)
0.28 (?0.10 to 0.67)
0.34 (?0.04 to 0.71)
0.18 (?0.20 to 0.56)
0.08 (?0.30 to 0.46)
0.29 (?0.09 to 1.04)
0.26 (?0.12 to 0.65)
0.07 (?0.31 to 0.45)
CI indicates confidence interval; ITT, intention-to-treat.
TABLE 2. TMD Diagnoses According to the Research Diagnostic
Criteria for Temporomandibular Disorders (RDC/TMD)
Ia: myofascial pain 93.1%81.5%
Ib: myofascial pain
with limited opening
Any disk displacement:
Any disk displacement:
IIIa: arthralgia left,
IIIb: osteoarthritis left,
IIIc: osteoarthrosis left,
3.4%, 0% 0%, 0%
0%, 0%0%, 0%
Multiple diagnoses are possible.
Shedden Mora et alClin J Pain?Volume 29, Number 12, December 2013
r2013 Lippincott Williams & Wilkins
a week, and 1 patient used the splint 2 nights or less a week.
Of the 10 patients who were instructed to use the splint
during daytime, 2 patients reported daily splint use, 5
patients used it 1 to 3 nights a week, and 2 patients stated
that they had used the splint 2 or less days a week.
In the BFB-CBT group, all patients reported the use of
the portable biofeedback devices during daytime and for
nocturnal training. Eighteen of 27 patients filled in sys-
tematic reports about home BFB training, and these reports
confirmed good adherence (mean number of nocturnal
training sessions=11.1, SD=5.5; mean number of train-
ing sessions during daytime=9.6, SD=3.2). Moreover,
participants in the BFB-CBT group were asked to rate the
usefulness of 4 treatment elements on a 0 (not useful at all)
to 5 (very useful) rating scale. Ratings were generally
positive,with the highest
conversations” (mean=4.56, SD=0.51), followed by
“exercises” (mean=4.11, SD=0.70), and “biofeedback”
Changes in NMMA
Forty complete data sets (21 patients in the BFB-CBT
group and 19 patients in the OS group) were available for
analysis of pre-post changes in NMMA. The main reasons
for missing data were loosened electrodes during nocturnal
recording, failure to handle the device, or failure to charge
batteries. At baseline, no significant differences were found
in NMMA in terms of mean number and mean duration of
EMG bursts (T=0.77, P=0.45 and T=0.61, P=0.55,
Repeated-measures ANOVAs did not reveal a sig-
nificant time or interaction effect for the mean number of
EMG bursts per hour sleep (F1,38=2.57, P=0.12 and
F1,38=0.01, P=0.94, respectively) and the mean duration
of EMG bursts per hour sleep (F1,38=2.61, P=0.11 and
F1,38=0.00, P=0.97, respectively). In other words, the
amount of NMMA did not change significantly from pre-
treatment to posttreatment in either of the groups.
The present study aimed to investigate the efficacy of
BFB-CBT compared with dental treatment with OS for
patients with chronic TMD. The results of general treat-
ment efficacy showed statistically and clinically significant
improvements in characteristic pain intensity, pain-related
disability, and jaw use limitations, with mostly medium
effect sizes, for patients in both conditions. In other words,
treatments were highly comparable in terms of primary
outcomes. Forty-five percent of the patients receiving BFB-
CBT and 48% of patients receiving OS treatment showed
clinically meaningful improvement, indicating that patients
responded well to both treatments. Patients receiving BFB-
CBT were more satisfied with the treatment, and rated their
overall improvement higher compared with patients in the
OS group. Slightly more adverse events were reported in the
OS group, though the difference was not statistically
For secondary outcomes, BFB-CBT resulted in sig-
nificantly larger improvements in cognitive and behavioral
pain coping skills. Although multivariate analyses only
approached significance, short-term and long-term effect
sizes for BFB-CBT were larger for all secondary outcomes.
In line with our hypothesis, BFB-CBT tended to be
superior though the difference was not as pronounced as
expected. Effect sizes did not change substantially between
posttreatment to follow-up, showing that improvements
were maintained over the 6-month follow-up period.
Our findings join a number of previous studies that
have provided evidence of the general effectiveness of CBT
for chronic TMD.2–6In line with the few studies comparing
psychological with dental treatment, we found BFB-CBT
and OS to be equally effective in terms of pain reduction.
Turk et al19concluded that OS treatment results in larger
pain reductions on short term, whereas biofeedback and
stress management lead to larger reductions in pain and
depression on long term. Similarly, Carlson et al18showed
that a behavioral self-regulation training improved pain
slightly more than dental treatment. A recent trial reported
that a short education program was slightly more effective
than OS treatment.41
Fricton et al14
“behavioral therapy has equal efficacy to appliances short
term but may have more positive effects than appliances in
the long term, perhaps due to its addressing of the under-
lying central etiologies more effectively than appliances” (p.
249). Given the highly chronic condition and psychological
comorbidityof our patient sample, both
approaches can be regarded as successful. The fact that
BFB-CBT resulted in larger improvements in pain coping
skills, but also was well accepted by the patients, underlines
the importance and feasibility of psychological treatments
in the clinical management of TMD.
Contrary to our expectations, no significant changes
occurred in NMMA. Although most patients in our study
showed clinical signs and reported sleep bruxism, it is
possible that not all patients with TMD actually exhibited
elevated levels of NMMA. Previous studies have shown
clear relations between self-reported sleep bruxism and
TMD, but findings on associations of directly assessed
NMMA with TMD pain are inconclusive (for an overview,
see Manfredini and Lobbezoo42). In fact, a recent pub-
lication by our group did not show elevated levels of
NMMA in TMD compared with pain-free controls.36
Longitudinal changes in pain do not show clear relations to
physical changes in jaw function.43Although both bio-
feedback and OS may have at least a temporary effect on
NMMA, physiological changes do not necessarily corre-
spond to symptom changes. Apart from sleep bruxism,
other pathophysiologic mechanisms such as peripheral and
central sensitization might contribute to the maintenance of
chronic TMD.15Moreover, awake oral parafunctions seem
to play a more important role in the development and
maintenance of TMD pain.9Thus, increasing awareness
and reducing oral parafunctions may be more important
than the reduction of sleep bruxism.
Several limitations should be considered: first of all,
since our aim was to compare psychological with dental
treatment, we did not include a combined BFB-CBT and
OS condition, thus we are not able to estimate the benefits
of combining the approaches. Second, the BFB-CBT group
received more treatment sessions, thus we cannot rule out
the possibility that advantages of BFB-CBT could be
attributed to longer treatment duration or higher amount
of attention. An attention control group would have been
helpful to determine specificity of CBT elements as reported
by Turner et al.4Furthermore, the follow-up period of 6
months was rather short; with longer follow-up periods
differential treatment effects might have been more pro-
nounced, as shown in previous studies.44Moreover, with
Clin J Pain?Volume 29, Number 12, December 2013 Cognitive-behavioral Treatment for TMD
r2013 Lippincott Williams & Wilkinswww.clinicalpain.com|1063
regard to length of the questionnaire, we relied on a single
item to assess adverse events. For OS, undesired events
such as tooth pain, occlusal changes or inflammation have
rarely been considered.14Nocturnal training, on the other
hand, might interfere with sleep quality and change sleep
stages. Future studies should assess adverse events in a
more systematic way. Although we also targeted oral par-
afunctions in our BFB-CBT treatment protocol, we did not
systematically assess physiological changes during wake
state. Future studies should consider methods to assess oral
parafunctions such as experience sampling methods or
Although adherence rates were acceptable in both
treatment conditions compared with those reported in
former studies,5we cannot rule out the possibility that
higher adherence rates might be associated with better
treatment outcome. Adherence to biofeedback training was
assessed in the BFB-CBT condition, but we did not sys-
tematically assess adherence to other homework such as jaw
exercises. Future TMD studies should focus more closely
on the systematic assessment of adherence and its relation
to treatment outcome.
Finally it is necessary to critically reflect our study
design. During recruitment, several participants did not
meet inclusion criteria because they already disposed of an
OS sufficient to our standards, though it had not provided
the desired improvement. Other participants clearly favored
one of the two quite different treatment approaches, and
were unwilling to be randomized. In fact, both patients and
health care providers may possess a certain etiological ill-
ness concept (eg, medical, psychosocial), depending on
previous experience and expertise.45A treatment matching
to the patient’s concept might be more convincing and
successful for the patient. In future, patient’s etiological
concepts should be taken into account, and treatment
should be matched to address the patient’s needs.46
Our findings support the importance of CBT for
chronic TMD pain, and suggest directions for future
research. Patients suffering from TMD most probably seek
dental treatment in the first place.47However, our results
and former evidence suggest that short CBT might be
sufficient and even more cost-effective to improve pain
and functioning, and prevent chronic manifestation of
TMD.3,6,20Psychological variables are closely associated
with pain intensity, physical functioning, and chronic
manifestation.48Therefore, they should be assessed and, if
necessary, targeted in TMD treatment. Tu ¨ rp et al46claim
that patients with high levels of psychological distress
benefit from multimodal treatment. Similarly, 2 studies by
Dworkin et al3,49have evaluated different treatment
options depending on the level of chronicity. In contrast,
several studies have shown that CBT works equally well for
all patients with TMD.5,50Clearly more research is needed
to provide a treatment matched to address the patient’s
TMD is a common and cost-intensive condition
affecting approximately 5% to 10% of the adult pop-
ulation.1The fact that patients suffer from TMD for several
years before they receive effective therapy, indicates a lack
of availability of adequate treatment options. Patients at
high risk of chronic manifestation can be reliably identi-
fied.20Especially, those patients might benefit from early
TMD-related symptom burden and disability.
The authors are extremely grateful to Prof. Dr U.
Lotzmann and his staff of the Department of Prosthetic
Dentistry, Philipps University of Marburg, for supporting this
1. Scrivani SJ, Keith DA, Kaban LB. Temporomandibular
disorders. N Engl J Med. 2008;359:2693–2705.
2. Gardea MA, Gatchel RJ, Mishra KD. Long-term efficacy
of biobehavioral treatment of temporomandibular disorders.
J Behav Med. 2001;24:341–359.
3. Dworkin SF, Turner JA, Mancl L, et al. A randomized clinical
trial of a tailored comprehensive care treatment program
for temporomandibular disorders. J Orofac Pain. 2002;16:
4. Turner JA, Mancl L, Aaron LA. Short- and long-term efficacy
of brief cognitive-behavioral therapy for patients with chronic
temporomandibular disorder pain: a randomized, controlled
trial. Pain. 2006;121:181–194.
5. Litt MD, Shafer DM, Kreutzer DL. Brief cognitive-behavioral
treatment for TMD pain: long-term outcomes and moderators
of treatment. Pain. 2010;151:110–116.
6. Aggarwal VR, Glenny AM, Javidi H, et al. Psychosocial
interventions for the management of chronic orofacial pain.
Cochrane Database Syst Rev. 2011;11.
7. Crider A, Glaros AG. A meta-analysis of EMG biofeedback
treatment of temporomandibular disorders. J Orofac Pain.
8. Crider A, Glaros A, Gevirtz R. Efficacy of biofeedback-based
treatments for temporomandibular disorders. Appl Psychophy-
siol Biofeedback. 2005;30:333–345.
9. Glaros AG. Temporomandibular disorders and facial pain:
a psychophysiological perspective. Appl Psychophysiol Biofeed-
10. Flor H, Birbaumer N, Schugens MM, et al. Symptom-specific
11. Ommerborn MA, Schneider C, Giraki M, et al. Effects of an
occlusal splint compared with cognitive-behavioral treatment
on sleep bruxism activity. Eur J Oral Sci. 2007;115:7–14.
12. Al-Ani MZ, Davies SJ, Gray RJM, et al. Stabilisation splint
therapy for temporomandibular pain dysfunction syndrome.
Cochrane Database Syst Rev. 2004.
13. Kreiner M, Betancor E, Clark GT. Occlusal stabilization
appliances: evidence of their efficacy. J Am Dent Assoc. 2001;
14. Fricton J, Look JO, Wright E, et al. Systematic review and
meta-analysis of randomized controlled trials evaluating
intraoral orthopedic appliances for temporomandibular dis-
orders. J Orofac Pain. 2010;24:237–254.
15. Svensson P, Jadidi F, Arima T, et al. Relationships between
craniofacial pain and bruxism. J Oral Rehabil. 2008;35:
16. van der Zaag J, Lobbezoo F, Wicks DJ, et al. Controlled
assessment of the efficacy of occlusal stabilization splints on
sleep bruxism. J Orofac Pain. 2005;19:151–158.
17. Flor H, Birbaumer N. Comparison of the efficacy of electro-
myographic biofeedback, cognitive-behavioral therapy, and
conservative medical interventions in the treatment of chronic
musculoskeletal pain. J Consult Clin Psychol. 1993;61:653–658.
18. Carlson CR, Bertrand PM, Ehrlich AD, et al. Physical self-
regulation training for the management of temporomandibular
disorders. J Orofac Pain. 2001;15:47–55.
19. Turk D, Zaki H, Rudy T. Effects of intraoral appliance and
biofeedback/stress management alone and in combination in
Shedden Mora et al Clin J Pain?Volume 29, Number 12, December 2013
r2013 Lippincott Williams & Wilkins
treating pain dibular disorders. J Prosthet Dent. 1993;70: Download full-text
20. Stowell AW, Gatchel RJ, Wildenstein L. Cost-effectiveness of
treatments for temporomandibular disorders. Biopsychosocial
intervention versus treatment as usual. JADA. 2007;138:202.
21. Turk DC, Dworkin RH, Allen RR, et al. Core outcomes
domains for chronic pain clinical trials: IMMPACT recom-
mendations. Pain. 2003;106:337–345.
22. Dworkin SF, Le Resche L. Research diagnostic criteria for
temporomandibular disorders: review, criteria, examinations and
specifications, critique. J Craniomandib Disord. 1992;6:301–355.
23. Schulz K, Altman D, Moher D. The Consort Group.
CONSORT 2010 Statement: updated guidelines for reporting
parallel group randomised trials. BMC Med. 2010;8:18–26.
24. Wittchen HU, Wunderlich U, Gruschwitz S, et al. Structured
Clinical Interview for DSM-IV, Axis I (SCID). Go ¨ ttingen:
25. Rief W, Hiller W. A new approach to the assessment of the
treatment effects of somatoform disorders. Psychosomatics.
26. Haythornthwaite JA. IMMPACT recommendations for clin-
ical trials: opportunities for the RDC/TMD. J Oral Rehabil.
27. Farrar JT, Young JP Jr, LaMoreaux L, et al. Clinical
importance of changes in chronic pain intensity measured on
an 11-point numerical pain rating scale. Pain. 2001;94:149–158.
28. Tait RC, Chibnall JT, Krause S. The Pain Disability Index:
psychometric properties. Pain. 1990;40:171–182.
29. Hautzinger M, Bailer M. Centre for Epidemiological Studies
Depression scale (CES-D). Weinheim: Beltz; 1993.
30. Lo ¨ we B, Zipfel S, Herzog W. PRIME MD Patient Health
Questionnaire—German Version. Karlsruhe: Pfizer; 2002.
31. Geissner E. German Pain Coping Questionnaire (FESV).
Go ¨ ttingen: Hogrefe; 2001.
32. Rief W, Hiller W, Heuser J. Screening for Somatoform
Disorders. Huber: SOMS; 2008.
33. Kares H, Schindler HJ, Scho ¨ ttl R. Temporomandibular
Disorders. Hannover: Schlu ¨ tersche; 2006.
34. Guy W. ECDEU Assessment Manual for Psychopharmacology
(DHEW Publication No. ADM 76-338). Washington, DC: US
Government Printing Office; 1976.
35. Weise C, Heinecke K, Rief W. Biofeedback-based behavioral
treatment for chronic tinnitus: results of a randomized
controlled trial. J Consult Clin Psychol. 2008;76:1046–1057.
36. Shedden Mora M, Weber D, Borkowski S, et al. Nocturnal
masseter muscle activity is related to symptoms and somatiza-
tion in temporomandibular disorders. J Psychosom Res.
37. Lavigne GJ, Rompre PH, Montplaisir JY. Sleep bruxism:
validity of clinical research diagnostic criteria in a controlled
polysomnographic study. J Dent Res. 1996;75:546–552.
38. Glaros AG, Kim-Weroha N, Lausten L, et al. Comparison
of habit reversal and a behaviorally-modified dental treatment
for temporomandibular disorders: a pilot investigation. Appl
Psychophysiol Biofeedback. 2007;32:149–154.
39. McGaw B, Glass GV. Choice of the metric for effect size in
meta-analysis. Am Educ Res J. 1980;17:325–337.
40. Cohen J. A power primer. Psychol Bull. 1992;112:155–159.
41. Michelotti A, Iodice G, Vollaro S, et al. Evaluation of the
short-term effectiveness of education versus an occlusal splint
for the treatment of myofascial pain of the jaw muscles. JADA.
42. Manfredini D, Lobbezoo F. Relationship between bruxism and
temporomandibular disorders: a systematic review of literature
from 1998 to 2008. Oral Surg Oral Med Oral Pathol Oral
Radiol Endod. 2010;109:e26–e50.
43. Ohrbach R, Dworkin SF. Five-year outcomes in TMD:
relationship of changes in pain to changes in physical and
psychological variables. Pain. 1998;74:315–326.
44. Dworkin SF, Turner JA, Wilson L, et al. Brief group cognitive-
behavioral intervention for temporomandibular disorders.
45. Suvinen TI, Reade PC, Kemppainen P, et al. Review of
aetiological concepts of temporomandibular pain disorders:
towards a biopsychosocial model for integration of physical
disorder factors with psychological and psychosocial illness
impact factors. Eur J Pain. 2005;9:613–633.
46. Tu ¨ rp JC, Jokstad A, Motschall E, et al. Is there a superiority
of multimodal as opposed to simple therapy in patients
with temporomandibular disorders? A qualitative systematic
review of the literature. Clin Oral Implants Res. 2007;18:
47. Dao TT, Lavigne GJ. Oral splints: the crutches for tempor-
omandibular disorders and bruxism? Crit Rev Oral Biol Med.
48. Turner JA, Dworkin SF, Mancl L, et al. The roles of beliefs,
catastrophizing, and coping in the functioning of patients with
temporomandibular disorders. Pain. 2001;92:41–51.
49. Dworkin SF, Huggins KH, Wilson L, et al. A randomized
clinical trial using research diagnostic criteria for temporo-
mandibular disorders-axis II to target clinic cases for a tailored
self-care TMD treatment program. J Orofac Pain. 2002;16:
50. Turner JA, Holtzman S, Mancl L. Mediators, moderators, and
predictors of therapeutic change in cognitive-behavioral
therapy for chronic pain. Pain. 2007;127:276–286.
Clin J Pain?Volume 29, Number 12, December 2013 Cognitive-behavioral Treatment for TMD
r2013 Lippincott Williams & Wilkinswww.clinicalpain.com|1065