Introduction

Surface electromyographic biofeedback em­ploys noninvasive instrumentation to monitor subtle changes in muscle physiology and present that information to a patient to assist them in learning to modify muscular activity and resting tonus in a desired direction. Training goals may range from general relaxation of the masticatory and other muscles of the face, neck, and shoulders to functional muscle reeducation aimed at establishing a more normal bite pattern. In this learning environment, the clinician frequently assumes the role of a coach, helping the patient understand (discriminate) feedback information, sometimes suggesting strategies to employ to move in the desired direction, providing encouragement (reinforcement) for targeted behavior, and adjusting training goals (shaping) as needed. As can be inferred from these terms, the rational basis for biofeedback derives from the field of learning theory and behavior modification.

The biofeedback process can also be seen as a form of three-party social interaction in which the instrument not only provides information, but it also inserts another party (the biofeedback device) into the evaluative process. This can be useful in situations where the patient has difficulty learning the task; the clinician does not have to take on the role of “judging” the patient, telling them they have to relax, but instead the clinician can become an ally who assists the patient in trying to learn from the feedback provided by the machine. Between sessions, the patient is encouraged to rehearse/practice the desired behavior. At the next session, the instrumentation provides a means of obtaining objective information on how much progress the patient has made. The goal of this learning process is to produce a relatively permanent change in behavior that becomes internalized such that feedback from the instrumentation is no longer needed to produce or maintain the behavior.

Current surface electromyography (sEMG) monitoring and clinical biofeedback instrumentation provide healthcare professionals with tools that can be utilized in a range of assessment and treatment tasks. The manner in which these instruments are utilized will vary based on the perspective, training, and role a particular clinician plays in the overall care of the dental patient. Thus, in one practice, the primary practitioner may use the instrumentation directly, while in others, trained auxiliary staff or allied health professionals may employ them. Similarly, an instrument that monitors muscle activity may be used in one setting primarily as a biofeedback device to aid in relaxation training while it is utilized in another clinic largely as an assessment tool or in muscle reeducation.

Fundamental to a behavioral approach to health care is the concept of assessment. How can one propose to modify a system without first establishing the characteristics of the system? While this may seem like a statement of the obvious, it is an important perspective to keep in mind when developing a treatment plan, particularly in complex presentations as may be the case in myofascial pain disorder (MPD) or temporomandibular disorder (TMD). For example, careful assessment of a complaint of jaw pain may reveal heightened reactivity to emotional stimuli in one patient, suggesting possible roles for biofeedback-assisted relaxation training and stress management counseling. In another individual, the history may not indicate any significant anxiety or stress-related issues, and functional assessment may not reveal any remarkable reactivity to emotional stimuli. On the other hand, this patient may show clear patterns of dysfunctional muscle activation during normal jaw opening and clenching activities, suggesting a role for feedback-assisted retraining of muscle recruitment patterns. As these brief examples suggest, biofeedback instrumentation can play as important a role in assessment as it does in actual treatment.

This chapter is intended to provide an introduction to some of the established and emerging applications of sEMG instrumentation in dental practice, as well as discussing factors in the selection and effective utilization of this instrumentation. Some of the technical material presented, such as the discussion of electromyography (EMG) quantification and frequency characteristics, are not typically covered in introductory presentations and may appear a bit intimidating at first to some readers. However, it is the author’s opinion that the basic “take-away” messages in each of these sections are straightforward and can be made quite accessible. Mastery of the essentials of biofeedback is well within the capability of dentists and/or appropriately trained assistants who are effectively already involved in behavioral modification/educational activities with patients on a daily basis. (See Dworkin’s (2007) discussion of role of the dentist and the dental healthcare team in the delivery of biobehavioral modalities.) Opportunities to obtain “hands-on” experience with the concepts and techniques presented here are generally available through specialized short courses offered at various professional society meetings and other venues.

Historical Development

The roots of biofeedback can be traced back to a number of sources; however, one of the seminal papers in modern biofeedback was Budyznski and Stoyva’s (1969) publication on a method to produce deep muscle relaxation by means of “analog information feedback.” Muscle activity was recorded from the region of the frontalis muscles using sEMG, and subjects were instructed to close their eyes and relax as deeply as they could, particularly the muscles of the forehead. The experimental group was presented with an auditory tone that decreased in pitch as muscle activity decreased, and they were further instructed to keep the pitch of the tone low. One control group was presented with a constant low tone unrelated to their level of muscle activity and were told that the monotonous tone would help them relax. A second control group was not presented with a tone and simply instructed to relax. The active feedback group showed the greatest success at lowering muscle activity, supporting the author’s contention that providing individuals with feedback about their instantaneous level of muscle activity would be more successful in producing a relaxed state than simply asking an individual to relax.

Budyznski and Stoyva (1973) went on to extend their method to examine both auditory and visual feedback and the training of the masseter muscles. They outlined a shaping technique to produce gradual relaxation by successive approximations to the deeply relaxed state. They proposed that graduated training steps are likely to reduce the chance of error which may occur if a training goal is set too large. Eighty males were divided into 4 groups of 20. Two experimental groups received the shaping method with biofeedback using either an auditory tone or a visual signal. An irrelevant, steady, low-tone control group and a silent (no-tone) group were again used as control conditions. The active biofeedback groups showed a deeper level of muscle relaxation in the target muscles than the control groups. The level of muscle relaxation achieved by the auditory and visual feedback groups was not statistically different, indicating that both forms of informational feedback were effective. In this paper, Budyznski and Stoyva proposed that this methodology might be particularly useful in dentistry in the treatment of myofascial pain dysfunction syndrome where excessive muscle activity is present and also proposed potential applications of the instrumentation in the treatment of bruxism.

Dental Relevant Applications

Budyznski and Stoyva’s work was followed by clinical reports by Carlsson, Gale, and Ohman (1975) and by Gessel (1975) on the use of EMG biofeedback in TMDs and by Cannistraci (1976) on biofeedback training and bruxism. By 1979, Buonomano and Buonomano had published an expansive article in General Dentistry entitled “Biofeedback: The Emergence of a New Dental and Medical Perspective.” Most major texts on biofeedback now include discussions on applications with dental patients (e.g., Cannistraci & Fritz, 1989; Glaros & Lausten, 2003). A role for sEMG as a modality in the context of biofeedback-assisted relaxation training and in muscle reeducation is now generally recognized by most relevant professions. A 1982 American Dental Association (ADA) President’s Conference concluded that EMG biofeedback had reasonable scientific support and could be recommended for TMD cases involving the muscles of mastication; this position was reaffirmed in an ADA-affiliated follow-up review (Mohl et al., 1990). The status of sEMG as an assessment modality for diagnosis in dentistry is, however, still somewhat controversial and will be addressed later in this chapter.

The listing that follows provides a clinically oriented conceptual grouping of the ways in which biofeedback techniques are most commonly employed with dental patients:

• Biofeedback-assisted relaxation/desensitization training for dental anxiety/phobia (e.g., Berggren & Carlsson, 1984; Lundgren, Carlsson, & Berggren, 2006)
• Biofeedback-assisted relaxation training within the context of stress management/emotional reactivity training for stress-related TMD symptoms (see citations throughout this chapter, particularly the section reviewing efficacy studies)
• Biofeedback-assisted training to improve patterns of muscle activation/balance at rest and during masticatory activity (see citations throughout this chapter)
• Biofeedback-assisted muscle training to accommodate dental appliances/bite guards (Kasman, Cram, & Wolf, 1998; Mahony, 2010).

Some of the above-mentioned topic areas, such as dental fear and anxiety, are covered in depth in other chapters of this book. The use of sEMG as part of a nocturnal alarm system in the treatment of bruxism has shown some success but also has limitations (see Chapter 11 on bruxism by Glaros and Hanson).

Biofeedback in Clinical Practice

A feedback device can be as simple as a mirror that is used to visually demonstrate to a patient when one shoulder is being held higher than another and when they are in balance. But typically, biofeedback devices are electronic instruments that can detect extremely small physiological changes and provide direct feedback (information) about activity in the form of visual or auditory signals. Biofeedback devices may be relatively basic instruments that indicate physiological changes by means of a moving needle or a simple tone that changes in pitch as a physiological variable changes. Alternatively, computer-based systems are available that monitor multiple muscle sites, or a range of physiological parameters, at the same time and provide detailed visual and auditory display options. Computer-based systems may also offer extensive data reduction and report generation capabilities.

In some instances, the feedback may be indirect, where the clinician observes the instrument and provides the patient with cues indicating whether activity is moving in a desired or nondesired direction. More commonly in dental practice, the meaning of the feedback signal is simply explained (i.e., “when the line on the computer display goes down, this indicates lower muscle tension”), and suggestions are provided to start the feedback process (“try slowing your breathing, allow your jaw to drop, and observe what happens on the display”). If two muscles are being monitored, they may be represented as different colored lines or as bars. Biofeedback displays can be quite engaging, and some patients will respond to a visual display of unbalanced bilateral massetter or trapezius activity by immediately trying to bring the two signals into balance. Other patients will require more active guidance. For an extended consideration of biofeedback principles and clinical technique, the edited volumes by Basmajian (1989) and Schwartz and Andrasik (2003) are standard references. Chapters in edited texts specifically on biofeedback applications in dental patients include Cannistraci and Fritz (1989), Hudzinski and Lawrence (1990), and Glaros and Lausten (2003); see also chapter 9 in Kasman et al. (1998). Another source of information and professional support is the web site of the Association for Applied Psychophysiology and Biofeedback (http://www.aapb.org).

In situations where a major component of the treatment plan is to teach a patient to relax specific muscles, or to lower their arousal generally, it is often advantageous to combine some basic form of relaxation training with biofeedback. In fact, some professionals prefer to talk about the biofeedback work they engage in as “instrument-assisted relaxation training.” There are many approaches to relaxation training ranging from progressive muscle relaxation to techniques involving varying emphasis on the use of imagery, breathing, autogenic phrasing, or hypnotic suggestion. There are numerous sources for guidance on clinically oriented relaxation training; Lehrer, Woolfolk, and Sime (2007) and Turk, Meichenbaum, and Genest (1983) are excellent resources. An introduction to relaxation can be provided by a primary clinician or appropriately trained auxiliary staff member. Alternatively, there are a wide range of commercially available audio programs that can be successfully used as part of a relaxation training program (see Schwartz (2003) for a discussion on the use of audiotapes for relaxation training). However, an important caution should be kept in mind. Occasionally, a patient may be encountered with various psychological issues who may become uncomfortable if they attempt to relax. While it may ultimately be to the patient’s advantage that such anxiety is discovered and dealt with (either through appropriate support or referral), it is suggested as a general rule of practice that patients who appear unusually anxious or who otherwise seem uncomfortable with the idea of relaxation training should not be left alone with a training tape or a biofeedback device; an observer should be present to provide guidance as needed. See Schwartz, Schwartz, and Monastra (2003) for a discussion of possible negative reactions to relaxation and biofeedback-assisted relaxation training and guidelines for minimizing and managing such situations.

The question is sometimes posed as to whether a particular relaxation training technique or biofeedback is more effective as a treatment modality. Based on clinical experience, and a review of the research literature, it can be argued that in many situations, it is the combination of relaxation training and biofeedback that is the most effective. For many patients, biofeedback monitoring during (or before and after) a relaxation training session helps to demonstrate that the relaxation training procedure is producing a measurable change in their body. This may be the crucial step in helping a patient “buy into” the prescription that they practice the relaxation procedure outside of the clinic. In a similar fashion, knowing that their progress is going to be visible on the biofeedback device influences the motivation of some patients to practice at home between clinic sessions since their progress can be objectively monitored.

Budyznski and Stoyva’s (1973) emphasis on using a shaping procedure to train patients through a process of successive approximations to the eventual goal of muscle relaxation is a key aspect of applied learning theory. It is now well recognized that learning can be accelerated through positive reinforcement. In humans, simply meeting a goal of lowering the tone of a biofeedback instrument or dropping the signal representing muscle activity below a threshold line displayed on a computer screen can be an effective source of positive reinforcement. Positive reinforcement is most likely to occur if initial training goals are set to be relatively easy. Once an initial easy goal is met, the training goal can be modestly increased.

A common reaction when an individual is first asked to attend to the biofeedback signal and to relax is for them to actually become somewhat more tense. This can be interpreted as an understandable form of performance anxiety since their behavior is readily observable through the feedback signal. This response pattern can be mitigated in part by predicting for the patient in advance that they may find the signal going in the opposite direction at first. Consider advising the patient that it is not particularly important if the signal increases or decreases during the early part of the training session—what is important is paying close attention to what they feel like or what they are thinking about when the signal increases and, similarly, to pay close attention to what they are feeling or are thinking when the signal decreases. Emphasize that developing an awareness of what may be subtle changes in muscle tension is the initial training goal. As the patient becomes more comfortable with observing their own behavior, the shift to a goal of actually decreasing sEMG activity can be made. Similarly, it can often be useful to have a patient initially “play” with the biofeedback signal, deliberately increasing muscle activity to watch the signal change and then letting the tension go.

In the late 1970s/early 1980s, the author was the lead software developer of what became one of the first commercially available computerized biofeedback systems. Some of the aforementioned principles were built into an automated goal-setting algorithm. Following an initial set-up period when signal quality could be checked, visual display ranges adjusted, and during which the patient could “play” with the signal before formal data collection began, patients were instructed to relax comfortably while a resting baseline reading of muscle activity was obtained. In cases of muscle relaxation training, at the completion of the baseline period, the initial training goal was set to be 105% of the mean baseline sEMG reading. On a computer display screen where up corresponded to increased sEMG activity and down to decreased activity, this meant that a threshold or goal line was drawn on the screen that was 5% higher than the mean level of muscle activity present during the baseline recording period. In these circumstances, even if the patient were to tense a little, they had a reasonable probability of being at or below the training goal at some point during the initial training trail, that is, of receiving some positive reinforcement. Trials were typically relatively short, often 1 minute in duration. Training goals were reassessed at the end of each trial. Performance was calculated as the percent time during a trial that the patient was at or below the training goal. If the patient met the goal less than half the time, the goal was classified as too hard and adjusted. If the patient was able to keep sEMG activity at or below the goal line between 50% and 90% of the time, the goal was kept at the same value. The training goal was only made more challenging when the patient showed relative mastery of the task by meeting goal greater than 90% of the trail. Based on the shaping principle, new goals were set to be halfway between the current goal and the patient’s mean sEMG level for the just-completed trial. In this manner, the new goal was adapted based on the patient’s performance level. When working with noncomputerized biofeedback equipment, or with systems that do not employ a similar automated training algorithm, a clinician can manually adjust training goals following the same basic principles.

In addition to the direct benefit of learning to relax overly tense muscles, biofeedback training can have important indirect effects as well. Chronic pain conditions can lead to a sense of loss of control, which may be associated with increased anxiety, secondary muscle tensing, and even depression. When a patient observes through the feedback process that he or she can voluntarily reduce muscle tension even a little, this can lead to an increased sense of control. This in turn can reduce anxiety, stress, secondary tensing, and even have a mitigating effect on depressive symptoms.

For some patients, the awareness training component of a biofeedback experience may, in some ways, be more important than the absolute level of muscle relaxation they can generate in the clinical setting. If this awareness training results in a patient becoming aware early on of when they are starting to tense muscles during the course of the day, it is much easier to attend to and think about releasing tension at that point than it is to try to relax muscles after tension has built up to the point that muscles go into spasm, trigger a headache, and so on.

While the most straightforward application of sEMG-based biofeedback training is to focus on relaxation of overly tense muscles as in Budyznski and Stoyva’s initial studies, the question of which muscles to train and for what reason should be carefully considered if maximal success is to be obtained. Selection of appropriate muscles of interest is discussed in several sections of this chapter.

Efficacy Evaluations of Biofeedback in the Treatment of TMD Disorders

Most experimental research on biofeedback-assisted treatment of TMD disorders has focused on muscle relaxation training, although a few studies may have included muscle balance reeducation in addition to relaxation work. Crider and Glaros (1999) examined outcome evaluations of treatments performed with sEMG to determine the efficacy of these studies and to estimate treatment effect size. Thirteen studies of sEMG biofeedback treatment for TMD were analyzed, consisting of 6 controlled, 4 comparative treatment, and 3 uncontrolled trials. Three types of outcome data were used: pain reports, clinical exam findings, and ratings of global improvement.

Five of the six controlled trials found sEMG biofeedback treatments to be superior to no treatment or psychological placebo controls for at least one of the three types of outcome. Data from 12 studies contributed to a meta-analysis that compared pre- to posttreatment effect sizes for EMG biofeedback treatments to effect sizes for control conditions. Mean effect sizes for both reported pain and clinical exam outcomes were substantially larger for biofeedback treatments than for control conditions. In addition, 69% of patients who received sEMG biofeedback treatments were rated as symptom-free, or significantly improved, compared with 35% of patients treated with a variety of placebo interventions. Follow-up outcomes for sEMG biofeedback treatments showed no deterioration from posttreatment levels. The authors concluded that the outcomes of these studies support the efficacy of sEMG of biofeedback treatments for TMD. A more recent review by these authors (Crider, Glaros, & Gevirtz, 2005) comes to the same conclusion.

Another meta-analysis (51 studies) was performed by Fernandez and Turk (1989). They concluded that the results supported the view that there are long-term advantages when using cognitive behavioral skills training (CBST) over other dental or pharmacological techniques in managing TMD-related pain. Flor and Birbaumer (1993) compared CBST, biofeedback, and conservative medical treatment in chronic back and TMD. Treatment with biofeedback produced the most change. At 6 and 24 months, only the biofeedback group maintained the significant decrease in pain severity.

Gardea, Gatchel, and Mishra (2001) ran a study that eliminated many shortcomings in earlier research. Chronic patients were assigned to one of three biobehavioral treatment groups—biofeedback, CBST, a combination of biofeedback and CBST, or a no-treatment comparison group. A 1-year follow-up was conducted. The major finding was that for pain scores, the biofeedback and combined groups showed the greatest improvement. The combined group showed the greatest improvement on pain-related disability. The 1-year follow-up showed that the biofeedback group was most effective in reducing TMD pain.

Dahlstrom (1989), in an exhaustive review, examined 12 studies that used sEMG. In one of the earliest (Carlsson & Gale, 1977), 11 temporomandibular joint (TMJ) patients were treated for from six to eight sessions and eight patients were reported to show significant improvement. Dohrman and Laskin (1978) treated 16 MPD patients with masseter muscle biofeedback training and 8 patients with “placebo feedback.” Mean levels of masseter activity were significantly reduced in patients in the experimental group. Only one biofeedback patient required further treatment while five of the eight controls required additional intervention. In another study (Stenn, Mothersill, & Brooke, 1979), masseter activity was measured in 11 patients with long-standing symptoms. Six were given masseter biofeedback with relaxation training for eight sessions. The remaining five were given relaxation training and masseter activity recorded with no feedback. All patients also received cognitive behavior modification. A significant reduction of masseter activity in all subjects occurred with no difference between groups. There was a significant reduction in number of symptoms and signs of MPD in all subjects. This study shows that multiple modes of therapy increase the probability of successful treatment; at the same time, if evaluated as independent therapies, biofeedback alone appeared to be the most effective.

Just as combined relaxation and biofeedback techniques may be superior to either alone, integrating biofeedback training with other conservative dental procedures may be advantageous. Turk, Hussein, and Rudy (1993) conducted two studies evaluating the effect of a biofeedback/stress management (BF/SM) treatment program and treatment with an intraoral appliance (IA), first separately and the second in combination. In the first study, the IA treatment was initially more effective in reducing pain. However, at a 6-month follow-up, the IA group showed significant relapse, particularly in depression, while the BF/SM group maintained and showed further improvement in both pain and depression indices. The use of an IA and BF/SM treatment was combined in the second study. It was reported that the combined treatment was more effective than either alone, especially in pain reduction at the 6-month follow-up. The authors concluded that the results supported the value of using both dental and psychological (behavioral) treatments to successfully treat TMD patients if treatment gains are to be maintained.

A number of conservative reviews by individuals and groups not professionally invested in the application of biofeedback techniques have concluded that the literature supports the position that relaxation techniques including biofeedback may be effective in treating TMDs (Mohl et al., 1990; Medlicott & Harris, 2006; Scrivani, Keith, & Kaban, 2008).

Summary points on clinical training:

• Introducing some form of relaxation training along with feedback from a physiological monitoring instrument is often found to be more effective than either procedure alone.
• When the goal is to learn to relax muscles, having the patient first practice gently increasing and then releasing muscle tension may be more effective than having them begin by immediately trying to decrease muscle activity.
• Initial training goals should be modest so that a patient has a high probability of obtaining positive reinforcement. Goals can be made progressively more challenging through a shaping procedure.
• The development of increased awareness of relative levels of muscle tension may be as significant for some patients as the absolute level of relaxation obtained during in-clinic training sessions.
• Similarly, in cases where there is a significant imbalance bilaterally in a set of muscles, bringing the left and right sides into ap­proximate balance and then focusing on decreasing general muscle activity may be more important than dramatically decreasing sEMG activity on a single side.

EMG as a Measure of Muscle Activity

EMG measures muscle activity by detecting changes in electrical potential that are associated with muscle action potentials. Muscle action potentials are similar to the action potentials that are generated in nerve cells, involving rapid changes in the relative concentrations of ionic charges across the cell membrane. When a motor neuron stimulates a muscle fiber, a wave of depolarization propagates bidirectionally along the length of the fiber, resulting in a cascade of molecular interactions that cause the fiber to contract. Muscle activity, such as the contraction of the masseter muscle to close the jaw, is the result of the summed activity of many muscle fibers. The overall strength of a muscle contraction is a function both of the number of fibers recruited and the frequency of firing of individual fibers.

sEMG is a noninvasive technique that places relatively large recording electrodes on the surface of the skin over a muscle or muscles of interest. Like fine-wire EMG, sEMG provides a measure of the summed electrical activity of groups of muscle fibers, but does so for a much larger area of muscle and without the discomfort and complications of inserting a lead through the skin. With proper instrumentation and recording technique, sEMG is a highly reliable measure of muscle activity across repetitions and testing days, obtaining reliability coefficients superior to that of invasive wire EMG recordings (Komi & Buskirk, 1970; Giroux & Lamontagne, 1990) and providing more information about functional activity level (Sihvonen et al., 1991). It should be noted that sEMG does have some limitations in that it cannot detect activity in deep muscle groups except when contractions from these groups are relatively strong (Wolf et al., 1991). Thus, while sEMG techniques readily record activity from muscle groups such as the masseter and anterior temporalis, low-level activity from overlaid muscles such as the pterygoid is more difficult to monitor. Fine-wire recordings remain quite useful in research for developing an understanding of the subtle features of recruitment patterns with a given muscle, such as Van Eijden, Blanksma, and Brugman’s (1993) investigation of the masseter.

Detecting the EMG Signal

The magnitude of the EMG signal that appears at the surface of the skin is extremely low when compared to the forms of electrical energy encountered in everyday settings. A fully charged standard flashlight battery has a rated electrical potential of 1.5 V. In comparison, a surface EMG reading from the forearm of a relaxed individual might be measured at 1.5 μV, one millionth the voltage of the flashlight battery. Besides this low-level electrical activity from muscle at the skin surface, external sources of electrical activity may also be present. These other sources are referred to as “noise” and include radio frequency waves (from radio and television stations, cell phones, and microwaves) and inter/>