Biofeedback and Self-Regulation, VoL 7, No. 4, 1982
Nocturnal Biofeedback for Nocturnal Bruxism 1 Anthony Piccione 2 San Francisco VA Medical Center
Thomas J. Coates The Johns Hopkins School of Medicine
June M. George Stanford University
David Rosenthal Stanford University
Peter Karzmark San Francisco VA Medical Center
Reports have appeared recently describing the successful reduction o f nocturnal bruxism through nocturnal biofeedback. These claims o f effective treatments rest mainly on the use o f a single index o f integrated masseter E M G levels as a measure o f bruxism and are based only on short-term effects. The present study was conducted to provide a more rigorous evaluation o f the effectiveness o f nocturnal biofeedback f o r nocturnal bruxism through the use o f all-night polysomnographic recordings. The results from multiple indices o f bruxism are internally consistent and indicate that simple nocturnal biofeedback does not appear to be effective in reducing nocturnal bruxing. Recommendations are made f o r a more comprehensive approach to the treatment o f nocturnal bruxism.
Bruxism is an oral habit involving grinding a n d / o r clenching o f the teeth which can lead to serious dental damage and is often associated with facial and neck pain and muscle contraction headaches. The habit can occur nocturnally or diurnally but it is generally believed that this distinction 1This study was supported in part by Grant No. MH-27551 from the Clinical Research Branch of NIMH to Carl E. Thoresen, Ph.D. The authors thank William Whitehead, Ph.D. for his helpful comments on an earlier version of this manuscript. 2Address all correspondence to Anthony Piccione, Ph.D., VA Medical Center (116-B), 4150 Clement Street, San Francisco, California 94121. 405 0363-3586/82/1200-0405503.00/0 © 1982PlenumPublishingCorporation
406
Piecione, Coates, George, Rosenthal, and Karzmark
respresents two separate phenomena (Reding, Zepelin, Robinson, Zimmerman, & Smith, 1968). Dental treatment aimed at eliminating tooth grinding while asleep traditionally has involved occlusal adjustment or the use of occlusal appliances (bite-plates or splints). These remedial approaches, however, have limited effectiveness and have not received systematic controlled demonstration of efficacy (Glaros & Rao, 1977). A number of reports have appeared recently describing the successful, though short-term reduction of nocturnal bruxism through nocturnal biofeedback (Rugh & Solberg, 1975; Kardachi & Clarke, 1977; Beemsterboer, Clark, & Rugh, 1978; Rugh & Johnson, 1981). These studies have typically used a portable electromyographic recording device containing an adjustable threshold detector and with audible feedback capability. Treatment has involved monitoring subjects' masseter muscle activity throughout the night and exposing them to an audio signal whenever the masseter tension levels exceed the preset thresholds. An arousal task has also been included in the treatment regimen in an attempt to facilitate maintenance of the results (Beemsterboer, Clark, & Rugh, 1978). Generally, a single index of integrated masseter EMG activity has been used as the primary dependent variable. The validity of this single index as a measure of bruxism has been questioned (Reding et al., 1968; Glaros & Rao, 1977). In recognition of the necessity for a more direct and multidimensional definition of bruxing activity, several recent studies have recorded both number of bruxes and duration of bruxing. Funch and Gale (1980) employed a portable EMG unit with a tape recording of the feedback tone to assess the number and duration of bruxes with a single subject. Using two sequential ABA designs with a single subject, they found a complex pattern of results. Mean duration of bruxing increased from initial baseline during the first treatment period and continued to increase during reversal. In the second phase, a reduction in duration of bruxing occurred during feedback but successive increases occurred in the reversal and follow-up assessment. In both phases a rebound effect occurred after treatment such that duration of bruxing in the second reversal and follow-up periods occurred at approximately five times the initial baseline. A similar pattern of results was found with number of bruxes, except that a reduction in incidence of bruxing occurred during the first treatment phase. With both measures, when biofeedback was discontinued, bruxing demonstrated a highly significant increase over the treatment period. Although this study provides some evidence for efficacy of biofeedback with bruxism, these results are overshadowed by the generally progressive increases in bruxing over the course of the study. This raises the question o f whether unanticipated treatment effects were increasing bruxing behavior during the
Nocturnal Bruxism
407
study. In any event, unanticipated increases of such magnitude indicate that the results of this study should be viewed cautiously. Rugh and Johnson (1981) also assessed the effect of biofeedback on both duration and number of bruxes. The study used five subjects and recording was done at home with a portable EMG unit and a battery operated chart recorder. All subjects showed a significant reduction in duration of bruxes during biofeedback but no reduction in number of bruxes. This suggests that biofeedback appears to inhibit bruxing after it begins but does not reduce the probability of its occurrence. Findings of this study are more interpretable than those of Funch and Gale and bring into question whether learning does occur during nocturnal biofeedback training. However, the study is limited by lack of reported reversal or follow-up data. It remains difficult to draw firm conclusions from these two studies. They do demonstrate, however, the need for a more comprehensive definition and a more complete assessment of bruxing in evaluating the efficacy of nocturnal biofeedback treatment. The purpose of the present study was to conduct a more rigorous evaluation of the effectiveness of nocturnal biofeedback for nocturnal bruxism through the use of several methodological refinements. These refinements involved the use of amplitude, duration, and rhythmicity criteria in defining bruxism (Reding et al., 1968), use of all night polysomnographic recording, and use of a time series design which incorporated treatment reversal within subjects and time lagged treatment across subjects.
METHOD
Subjects The subjects were two females who each reported a history of nocturnal bruxism without diurnal bruxing. Subject 1 was a 50-year old, single airline hostess with a 29-year history of nocturnal teeth grinding. She first became aware of the problem when her college roommate complained that she made "funny grinding noises" while asleep. Dental examination revealed vertical bony pockets about the molar teeth (periodontis), hypermic dental pulps (sensitive teeth), muscle spasm, joint pain, crepitus and popping, arid ter/derness (tendinitis) about the insertions of the masseter, medial pterygoid, and temporalis muscles. No obvious causative factors were found. Malocclusion was not indicated. She had been wearing a night guard regularly for the last two years to prevent further wear on her teeth. Subject had at one time attempted
408
Piccione, Coates, George, Rosenthal, and Karzmark
self-hypnosis as a treatment for her bruxism but terminated when she experienced herself as a "bad subject." Four root canals were performed in 1975 and complete gum surgery was performed in 1976. Subject 1 reported no sleep problems and, although she experienced occasional morning jaw ache, she was not using any medications. She reported having experienced chronic back pain until one year ago when she initiated an exercise program for herself. She presented no history of psychiatric problems or any obvious current life stressors. Subject 2 was a 44 year old divorced, registered nurse who had been experiencing morning jaw fatigue on the average of three to four times per week for the last three years. She attributed feelings of tension in her neck to the jaw fatigue. Her dental status, other than a slight malocclusion, was similar to Subject 1. Although she had never been told that she made nocturnal grinding noises, subject reported that she had often awakened with her teeth clenched tightly. In 1977, her dentist indicated that she was making excessive tempero-mandibular joint noise. Due to the malocclusion, she was considering dental reconstruction. Also, she was considering the use of a night guard as suggested by her dentist. Subject 2 reported no sleep problems or other medical disorders. She attributed nocturnal bruxism to stress due to her work and caring for her two children. No history of psychiatric problems was reported.
Design A time-series design was used. Experimental control was established using two design features: treatment reversals and time-lagged treatments. Each subject was observed first during baseline, followed by biofeedback, baseline, and biofeedback. Treatment for Subject 2 was lagged one week behind treatment for SUbject 1. Table I presents the experimental schedule for Subject 1 and Subject 2.
Measures All-night polysomnographic recordings were completed according to the schedule indicated in Table I at the Stanford Sleep Research Laboratory. Subjects slept on twin beds in private bedrooms. The EEH was recorded from the C3 and C4 placements referent to A1 and A2 (International 10-20 Jasper System), the EOG was recorded from the outer canthi of the eyes,
Nocturnal Bruxism
409
8~
~
crl ('-4
cr~O
.-a.a r~
t~¢'4 ("4 ('~ .,=
r~
r--I
88 g
S.a
i---I
410
Piccione, Coates, George, Rosenthal, and Karzmark
and the EMG for sleep recording purposes was recorded from the digastric chin muscles. EMG placements on the left and right masseter muscles were used to monitor bruxes. During all recordings, the high frequency filter for the EMG was set at 75 and the low frequency filter was set at 10. Standard prebedtime subject calibrations were used to provide baseline EEG, EOG, and EMG readings against which judgments regarding the presence or absence of indicators for the states and stages of sleep and degree of bruxing could be made (cf. Rosekind, Coates, & Thoresen, 1978). Masseter and digastric EMG activity was calibrated so that a 50 taVimpulse deflected the pen 1 cm. Paper speed was set at 10 ram/see. The data were recorded in an adjoining room on a Grass Model 7D polygraph. A technician was on hand throughout the night to monitor recording quality and assist subjects if necessary. All sleep stages were scored by an experienced technician and according to Rechtschaffen and Kales (1968) criteria. Bruxing incidents were defined by masseter EMG activity meeting amplitude, rhythmicity, and duration criteria (Reding et al., 1968). Amplitude: to separate contractions from baseline muscle activity, a brux was scored only if the EMG exceeded 20/~V; Duration: to separate contractions from clenching and holding, a brux was scored only when EMG activity exceeded 20/aV for no less than .5 seconds and no longer than 1.5 seconds; Rhythmicity: to separate momentary increases in EMG from the repeated contractions characteristic of bruxing, bruxing incidents were scored only when at least three bursts meeting amplitude and duration criteria occurred in sequence, separated by a return to baseline in the EMG of no more than 2.5 seconds. Figure 1 presents a characteristic tracing of bruxing from Subject 1. Each bruxing incident was scored for the number of bruxes and total duration. Number of bruxes was computed by totaling the number of distinct bruxes meeting amplitude criteria within the bruxing incident. The total duration of the bruxing incident was scored at the time from the beginning of the first brux to the end of the last brux in the series. The amplitude of each distinct brux.was categorized as either greater than 20 ~V but less than or equal to 50/aV or greater than 50/aV. A research assistant, trained to read the EMG printouts according to these criteria, scored all records. A second technician, experienced in scoring all-night sleep recordings, trained and supervised the research assistant in this task. Before beginning, each scored one record blind to the rating of the other. They reached or exceeded .90 interrater agreement on all variables (agreements - agreements + disagreements).
Nocturnal Bruxism
411
DIGASTRIC
50'u'vT
I I sec
lTig.1. A tracing from the polygraph showinga typicalbruxing responsefrom .Subject 1.
Biofeedback Apparatus and Procedure The portable E M G biofeedback devices previously described by Rugh and Solberg (1977) were used. The unit contains an adjustable threshold detector which could be set at any level from 10 to 100/aV. An audio signal in the form o f a constant tone was delivered by the feedback unit whenever masseter muscle activity exceeded the preset threshold. A speaker amplifier was used to amplify the audio signal presented to the subject. Except for the use of the all-night polysomnographic recordings, the procedure used in this study was a replication of the procedure used by
412
Piccione,
Coates, George,
Rosenthal,
and Karzmark
Beemsterboer, Clark, and Rugh (1978) and reported to be successful for the treatment of nocturnal bruxism. A rationale for using EMG biofeedback for treatment of nocturnal bruxism was presented first. Each subject was then issued a portable biofeedback unit and correct operation of the equipment and placement o f the electrodes on the masseter musd e was reyiewed and practiced. The threshold was set at a level whereby the tone was triggered only by tension beyond that produced by a moderate biting force (about 20 taV). Each subject was instructed to wear the biofeedback device nightly during the treatment phase of the study. Thus, while sleeping, subjects received a loud auditory signal whenever their masseter muscle activity exceeded the preset threshold. The need to wake up completely at the sound to the feedback signal was emphasized. To facilitate full arousal the subjects were told to use the biofeedback signal like an alarm clock. They were asked to arise from bed as soon as they were awakened by the
B-I 1.20
FB- I
B-2
FB-2
S-I
/
I.OO
_
_
_
.80
,o
oT fie
.40
:~
,20
AI )_~ I,,~
~
i
o
I
_
i
i
2
3
,
i i ,
8
I
_
9
i
i
14 s
2
,
.
.
16 24 303i'' 3
'
'
4'3 4'4 4'51
5
.
f_
6'4 6's
7
I0
i_ S-2
.80
lu
I
. . . . .
20
O DAYS WEEKS
Fig. 2.
, I
, 2
i 3 I
Total
seconds
, 8
i 9 2
of EMG Subjects
i tO
i , , 14 15 2 9 '3
, , 3 0 31 5
, 36
, i 37:38 6
43
i 44 7'
, i 45 57
, 58 9
a c t i v i t y > 20/~V d i v i d e d b y t o t a l s l e e p t i m e in h o u r s 1 and 2 during the treatment phases.
for
Nocturnal Bruxism
413
"alarm," cross the room, and record on a log sheet the time of night and the quality of their sleep on a scale from 0 to 10.
RESULTS To establish comparability with previous research (Rugh, 1978), the data were analyzed first in terms of seconds of masseter EMG activity > 20 /aV divided by total sleep time in hours. Results for Subjects 1 and 2 are presented in Figure 2. Biofeedback was associated with an initial suppressing effect in Subject 1, an increase in EMG activity on Day 16, and a return to baseline on Days 30 and 31. A dramatic rebound effect is noted in the second baseline period characterized by periods of EMG activity greater than demonstrated during the first baseline period. This was followed by a decrease to just above first baseline in the second feedback period. Subject 2 increased slightly from the first baseline to the first feedback period, increased EMG activity in the second baseline period, and decreased to original baseline levels in the second feedback period. Figures 3 and 4 present outcomes for Subjects 1 and 2 respectively in term of total bruxes, ~< 50/aV, and bruxes > 50/aV. Results are essentially the same as those presented in Figure 2. The all-night recording technician regularly monitored the subject through the intercom linking the equipment room with the subject's bedroom. For at least two hours per recording night, the technician marked on the record each time the feedback buzzer sounded. We were able to document that the feedback unit was working appropriately about 95% of the time in the laboratory. Subjects additionally were given cassette tape recorders and timers so that they could record feedback sound while sleeping at home. While it was not possible to document that the feedback unit was sounding only when EMG exceeded the preset level, the tape recordings did indicate that the units were sounding as subjects slept. Figures 5 and 6 present data on the effects of biofeedback on sleep. Total sleep time was suppressed initially for both subjects, but the effect was transitory. Feedback appeared to increase arousals during sleep. Number of arousals (to full wakefulness or to Stage 1) increased from first baseline to first feedback for Subject 1, remained at that level in second baseline, and then increased again during the second feedback phase. For Subject 2, number of arousals increased with feedback and then decreased during second baseline, increasing again with second feedback. In our laboratory observations, we found that subjects could turn off the biofeedback tone without awakening. Once the tone sounded, the subjects could reduce masseter EMG activity, turn off the tone, and maintain sleep. Over time, however, the subjects learned to ignore the tone and to maintain sleep.
414
Pieeione, Coates, George, Rosenthal, and Karzmark
B
FEEDBACK
?: ~,~
FB
illsEL,NE
r
~m 240
¢
180
}
_ 120 60
I
2 3 8 I
9 /4
15
2
3
16293031 4344456465 5 7 I0 636
35(1 ~ 30C Vl 250
'-U
~x 20(;
g ~
t50
z
IOG
_
5(;
I
2
3
8
I
9
14 15
2
16 2 9 3 0 3i
3
5
43 44 45 64 65 7
I0
S-~
o
A
== 2c m
,
DAYS WEEKS
I
,
,
2
,
,
,
,
,
,
,
,
3 8 9 14 r5 16 29 30 31 43 44 4 5 6 4 6 5 I
2
3
5
7
I0
Fig. 3. Subject 1: Number of bruxes per night, number of bruxes ~< 50/aV, number of bruxes > 50/aV, and total time bruxing.
Nocturnal Bruxism
415
s-_.zz BASELINE
FEEDBACK
B
rB
Jso~_ ~ IO0
~oIr ; ~ ; ~ ~ ,b r~,,;2'93'o3',3'6 3'7 3893 i495g~ 58 i
2
3
5
6
7
9
s-_2 ,50[
v,'2el ~ 90
~ 6o
3°I
.
.
.
I
2
3 I
.
.
.
.
.
........
8 9
, i J
I0 14 1,5 [9 3031 3 6 3 7 3 8 4 3 4 4 4 5 5 7 5 8
2
3
5
6
7
9
s-__z
120 IOO A
.~_
80
x
60 4o OD
2O DAYS WEEKS
i ~ ~ ~ ~ ,'o,~ J~2'93b 3', 3'63'~3'343 4'445 sT 5e I
2
3
5
6
?
9
Fig. 4. Subject 2: Number of bruxes per night, number of bruxes ~ 50/aV, number of bruxes > 50/aV, and total time bruxing.
416
Piccione, Coates, George, Rosenthal, and Karzmark
S-I
410
B-I
-
B-2
FB-2
-
m_j 350
y
0
.
.
.
.
.
.
.
300
z .~
250
20
a: LIJ
I0
z 0
IJJ
~
,o
~
0
o~
~
,
. . . . .
|
_ ,o
tl
i
i
DAYS 0
I
2 3 8 9
WEEKS
,
i
I
|
i
2
i
|
14 15 16 293031 43 44 45 64 615 3 5 7 I0
Fig. 5. Subject 1: Minutes of total sleep (Stages 1, 2, 3, 4, REM), number o f arousals to full wakefulness or to NREM Stage 1, percent of time awake while in bed, and percent of total sleep time spent in delta (Stages 3 and 4) sleep.
Nocturnal Bruxism
S-2
417
FB-I
500
B-2 II
450
400
350 -
I
30O
30
'
'
'
'
'
'
'
'
i
i
i
i
20 o
er I.i./
tO:"
Z ~:
0
z~ <
~w Z J
W~ O
.
.
.
.
b 0
I1~4~1P"
'
'
5
W
0 DAYS
WEEKS
I
2
3 I
8
9 2
I0 14 15 2 9 3 0 3 1 3
5
36 37 38 43 44 45 5758 6
7
9
Fig. 6. Subject 2: Minutes of total sleep (stages 1, 2, 3, 4, REM), number of arousals to full wakefulness or to NREM Stage 1, percent of time awake while in bed, and percent of total sleep in delta (Stages 3 and 4) sleep.
418
Piccione, Coates, George, Rosenthal, and Karzmark DISCUSSION
The present study represents one of the only attempts to study the effects of a commonly prescribed biofeedback technique on nocturnal bruxing using direct counts of bruxing incidents from a continuous all-night sleep recording. In contrast to two previous studies employing a similar approach, but without the sleep recordings, neither the number of bruxes nor the arhount of time spent bruxing were reduced. The lack of reduction in number ofbruxes supports Rugh and Johnson's findings of no reduction in number of bruxing incidents. Our finding of an initial suppressing effect of biofeedback in Subject 1 adds some support to the hypothesis that biofeedback may inhibit the behavior of bruxing after it commences but does not reduce long term occurrence. One explanatior~ for the observed increase in bruxing after the initial biofeedback period may be a treatment induced sleep deprivation. The procedure used resulted in an initial disruption of sleep with increased arousals to full wakefulness. While this was ultimately overcome by adaptation, the major impact of the initial training period could have been sleep deprivation. The resulting changes in depth and duration of sleep epochs could explain the associated increase in bruxing. This explanation has direct implications for the study of any biofeedback technique that can potentially disrupt sleep. If the disorder being studied is influenced by the nature of sleep, such a treatment approach would be clinically contraindicated. The finding of no reduction in time spent bruxing in either subject is in contrast to previous reports. It is unknown whether methodological differences from prior studies are responsible for this discrepancy. However, this study, together with those of Funch and Gale (1980) and Rugh and Johnson (1981), indicates that a more comprehensive examination of the effects of nocturnal biofeedback on nocturnal bruxing yields a more complex picture than that obtained by earlier, single dependent measure studies. Effective treatment of nocturnal bruxism may require more than simple feedback of masseter EMG activity. The addition of an arousal task alone also seems insufficient. One of the more successful treatments of nocturnal enuresis involves an elaborate procedure whereby subjects are awakened, must change the bed, and then practice the positive response of urinating in the toilet (Azrin, Sneed, & Fox, 1974). Nocturnal bruxism may require a similarly comprehensive procedure. Self-administered treatment may require a device that has subjects get up and out of bed and be fully awake before the machine turns off. Positive practice can then follow, perhaps using a programmed tape with subsequent automatic resetting of the biofeedback device as the subject returns to sleep.
Nocturnal Bruxism
419 REFERENCES
Azrin, N. H., Sneed, T. J., & Fox, R. N. Dry bed training, rapid elimination of childhood enuresis. Behavior Research and Therapy, 1974, 12, 147-156. Beemsterboer, G. T., Clark, G. T., & Rugh, J. D. The treatment of bruxism using nocturnal biofeedback with an arousal task. Paper presented at the International Association for Dental Research, Washington, D.C., 1978. Funch, D. P., & Gale, E. N. Factors associated with nocturnal bruxism and its treatment. Journal of Behavioral Medicine, 1980. 3, 385-397. Glaros, A. G., & Rao, S. M. A critical review. Psychological Bulletin, 1977, 84, 767-781. Kardachi, B. J., & Clarke, N. G. The use of biofeedback to control bruxism, Journal of Periodontology, 1977, 48, 639-642. Rechtschaffen, A., & Kales, A. (Eds.). A Manual of Standardized Terminology, Techniques, and Scoring Systems for Sleep Stages of Human Subjects. Washington, D.C.: U.S. Government Printing Office, 1968. Reding, G. R., Zepelin, H., Robinson, J. E., Zimmerman, S. O., & Smith, V. H. Nocturnal teeth-grinding: All night psychophysiologic studies. Journal of Dental Research, 1968, 47, 786-797. Rosekind, M. R., Coates, T. J., & Thoresen, C. E. Telephone transmission of all-night polysomnographic data from subjects' homes. Journal of Nervous and Mental Disease, 1978, 166, 438-441. Rugh, J. D. Electromyographic analysis of bruxism in the natural environment. In P. Weinstein (Ed.), Advances in Behavioral Research in Dentistry. Seattle, Washington: University of Washington Press, 1978. Pp. 67-83. Rugh, J. D., & Johnson, R. W. Temporal analysis of nocturnal bruxism during EMG biofeedback. Journal of Periodontology, 1981, 52, 263-265. Rugh, J. D., & Solberg, W. K. Electromyographic studies of bruxist behavior before and during treatment. Journal of the California Dental Association, 1975, 3, 56-59. Rugh, J. D., & Solberg, W. K. The identification of stressful stimuli in natural environments using a portable biofeedback unit. In J. D. Rugh, B. S. Perlis, & R. I. Disraeli (Eds.), Biofeedback and Dentistry: Research and Clinical Applications. Phoenix: Semantodonics, 1977. Pp. 41-43.