Applied Psychophysiology and Biofeedback, Vol. 24, No. 4, 1999

Specific Muscle EMG Biofeedback for Hand Dystonia K. K. Deepak1,3 and M. Behari2

Currently available therapies have only limited success in patients having hand dystonla (writer's cramp). We employed specific muscle EMG biofeedback (audio feedback of the EMG from proximal large muscles of the limb that show abnormally high activity during writing) in 10 of 13 consecutive patients (age, 19-62 years; all males) with a duration of illness from 6 months to 8 years. In three patients, biofeedback was not applicable due to lack of abnormal EMG values. Nine patients showed dystonic posture during writing and had hypertrophy of one or more large muscles of the dominant hand. The remaining four patients showed either involvement of small muscles or muscle wasting. Ten patients were given four or more sessions of EMG audio biofeedback from the proximal large limb muscles, which showed maximum EMG activity. They also practiced writing daily with the relaxed limb for 5 to 10 min. Nine patients showed improvement from 37 to 93% in handwriting, alleviation of discomfort, and pain (assessed on a visual analogue scale). One patient did not show any improvement. Thus EMG biofeedback improved the clinical and electromyographic picture in those patients with hand dystonia who showed EMG overactivity of proximal limb muscles during writing. This specific type of EMG biofeedback appears to be a promising tool for hand dystonia and might also be applied to other types of dystonias. KEY WORDS: hand dystonia; writer's cramp; biofeedback; EMG.

INTRODUCTION Focal hand dystonias are a crippling problem and pose a challenge to neurologists. In hand dystonia, there is hyperactivity of groups of muscles inappropriate to the motor act performed, for example, writing. Further, there may be a coactivation of antagonistic muscles and hyperactivity of muscles normally not involved in writing. Although the diagnosis is simple, the treatment is difficult (Thompson, 1993; Rhoad & Stern, 1993). The commonly prescribed treatments are prolonged anticholinergic medication and repeated injections of botulinum toxin (BTX). Anticholinergic agents rarely work and require high doses that lead to side effects (Ross & Charness, 1996). Although botulinum toxin is relatively free of systemic effects, its usefulness is limited by its temporary benefit (4-6 months), induction 1Department

of Physiology, All India Institute of Medical Sciences, New Delhi 110029. Fax: +91-11-6862663. e-mail: [email protected]. 2Department of Neurology, All India Institute of Medical Sciences, New Delhi 110029, India. 3To whom correspondence should be addressed.

267 1090-0586/99/1200-0267$16.00/0© 1999 Plenum Publishing Corporation

26S

Deepak and Behari

of temporary muscle weakness, occurrence of amyotrophy (Sheean, Murray, & Marsden, 1995), and high costs. Moreover it is effective only in selected patients (Poungvarin, 1991). Recently, lidocaine injections to block muscle afferents have been used. However, the treatment is cumbersome (injections twice a week for 6 months) and results in unpleasant side effects such as pain at the injection site, nausea, and dizziness (Sawamsto, Kaji, Katayama, Kuboli, & Kumara, 1995). Various nonpharmacological interventions have also been used with little success. Patients having hand dystonia usually attempt to write with the nondominant (unaffected) hand. This use of the unaffected hand may result in similar dystonia in that hand in about a quarter of the patients (Marsden & Sheehy, 1990). Reeducation of muscles and physical reinforcement with braces do not help. Other methods such as relaxationbiofeedback, acupuncture, psychotherapy, and avoidance conditioning have shown limited and only short-term benefits (Karp & Hallet, 1994). Studies on EMG biofeedback report varying degree of success (O'Neill, Gwinn, & Adler, 1996; Lees, Kleedrofer, & Foster, 1989; Ince, Leon, & Christids, 1986; Cottraux, Juenet, & Collet, 1983). In these studies the EMG biofeedback was used either as an adjunct to relaxation technique (see review by Ince et al., 1986) or as an aid to correct dystonic posture without being specific about the muscle involved. Thus, the electrodes were either applied on the flexors of the forearm to reduce flexion at wrist [see O'Neill et al. (1996) for a single case report] or on the dorsum of the forearm to monitor the contractions of extensor muscles during writing (Cottraux et al., 1983). In one study biofeedback was unsuccessful (Lees et al., 1989). However, in this study the details of biofeedback were not reported . It appears that there is an uncontrolled outflow of abnormal involuntary efferent discharge to muscles, which do not normally participate in writing [nontargeted muscle group(s) in these patients]. Perhaps this efferent outflow gives rise to cardinal symptoms of hand dystonia, namely, dystonic posture, discomfort, fatigue and pain, labored writing, and, finally, hypertrophy of the proximal muscles. An EMG study of proximal muscles with writer's cramp confirmed this excessive activity (Hughes & McLellan, 1985). The underlying mechanisms of writer's cramp are not entirely clear. Patients with writer's cramp have abnormal central sensorimotor processing as evidenced by PET studies (Tempel & Perlmutter, 1993). In a primate model, repetitive finger movements (digital overuse) have been shown to degrade the hand representation in primary sensory cortex (Byl et al., 1997). By using a noninvasive neuroimaging technique, a use-dependent overlap or smearing of the representational zones of the digits of the dystonic hand in the somatosensory cortex has been found in musicians with focal hand dystonia (Elbert et al., 1998). Based on these reports, Candia et al. (1999) used a constraint-induced movement therapy, a type of behavioral intervention, in five patients with focal hand dystonia. A 1-month followup showed that all five patients benefited from this therapy. This probably suggests that modified motor use (under "constraint") may help in reorganizing cortical representation in patients with focal hand dystonia. A close look at the sequence of the activation of the muscles in writer's cramp reveals that the central command for volitional activity of writing spreads beyond targeted areas in the primary sensory-motor areas in these patients. This may result in cocontraction of muscles inappropriate for the task. In such a situation, we hypothesized that active inhibition of proximal muscles (affected but not directly involved in the attainment of the goal) may reduce the overflow of motor areas pertaining to them. This reduction in overflow will result in corresponding decrement in muscle activation. Since some of the symptoms are produced by overactivation of proximal muscles, such nonactivation or reduced activation

269

EMG Biofeedback for Hand Dystonia

would result in an amelioration of symptoms. Moreover, the skeletal muscles are normally under volitional control. Since we have the capacity to train specific muscle groups, we might also be able actively to "detrain" the muscle groups that are hyperactive. This possibility prompted us to investigate whether patients can be taught to relax (inhibit) selectively a muscle or muscle group that show(s) high EMG activity during writing. This could be achieved by EMG biofeedback. In the present study we attempted such muscle detraining through muscle EMG biofeedback from one of the nontargeted muscles in patients suffering from writer's cramp. METHODS Thirteen patients (numbered 1-13) with a diagnosis of writer's cramps (age 19-62 years, all male) were referred to the Department of Physiology for biofeedback intervention by the Movement Disorder Clinic at the All India Institute of Medical Sciences. The diagnosis was based on the development of discomfort and pain during writing and deterioration of handwriting in the absence of any local or general pathology. Nine patients (i.e., all except Nos. 3, 5, 6, and 10) had complaints of discomfort and pain in the limb during writing, along with loss of speed and fluency, and they adopted a dystonic posture (elbow raised from the table with a varying degree of flexion at the elbow) while writing. These patients used extra muscular effort while writing. The effort was due to visible contraction of proximal and distal muscles of the limb. Thus writing appeared to be a labored task for these patients. All patients had a normal fundus occuli and absence of a Kayser-Fleischer ring. No abnormality was detected on neurological examination. All patients except Nos. 3, 6, and 10 had hypertrophy of muscles of the dominant limb. When they wrote for 3-8 min, the affected limb became warm. The warmth could be felt by an observer on the dorsum of their hands. Patients 3, 5, and 10 had normal limb musculature and showed involvement of the small muscles of the fingers. Here "small muscles" refers to the lumbricals, extensors, and flexors of the fingers. Their involvement was assessed by awkward finger positions adopted by patients such as a fanned-out position of the fingers or a flexed position of the fingers. Table I provides clinical data on these patients. Patient 6 Table I. Clinical Characteristics and Responses to EMG Biofeedback Intervention in Patients of Hand Dystoniaa Pt.No.

Age/sex

1 2 3 4 5 6 7 8 9 10 11 12 13

30/M 36/M 19/M 23/M 26/M 42/M 55/M 31/M 62/M 26/M 21/M 44/M 45/M

a

Duration of illness

Local sign

1 yr 1 yr

RLH RLH

6 mo 7 mo

— RLH

4 yr

4 yr 2 yr 8 yr 2 yr 2 yr 1 yr 1.5 yr

2 yr

— RLW RLH RLH LLH — RLH LLH RLH

Occupation Accountant Telephone operator Student Student Executive Secretary Lawyer Cashier Accountant Business Student Engineer Cashier

Medical treatment

Muscle selected forBF

Follow-up period

THP, 6 mg THP, 6 mg THP, 6 mg THP, 6 mg THP, 6 mg BTX, 3 inj. THP, 6 mg

Triceps Triceps None

3 yr

— THP, 6 mg

— CBZ THP, 6 mg THP, 6 mg

Br Br None

Br Br Br None

Br Br Br

2 yr — l yr l yr

— 8 mo 8 mo 8 mo

— 8 mo 8 mo 6 mo

RLH, right limb hypertrophy; LLH, left limb hypertrophy; RLW, right limb wasting; THP, trihexyphenidyl; BTX, botulinum toxin; CBZ, carbamazepine; Br, brachioradialis.

270

Deepak and Behari

had received injections of BTX and showed marked wasting and grade 4 muscle power (as assessed by a standard neurological examination protocol) at the time of referral. Nine patients were receiving trihexiphenidyl (6 mg/day). Side effects in the form of dryness of mouth, dizziness, and blurring of vision were reported by patients 1 and 5. All patients gave informed consent for participation in the study. The method for selection of patients and specific muscles for biofeedback is similar to that suggested for treatment by botulinum injection (Karp & Hallett 1994). Suitability of a patient for biofeedback was judged by two criteria, viz., the presence of hypertrophy of the proximal limb musculature (confirmed by girth measurement) and associated high EMG activity (by comparison of EMG activity of involved muscle during writing with that during normal limb maneuvers). The latter point is based on the premise that proximal muscles such as the triceps should show none or minimal activity during writing, while adduction will involve moderate activation of the majority of the muscle fibers of the triceps muscle. Ten patients (Nos. 1, 2, 4, 5, 7-9, and 11-13) were found suitable for specific muscle EMG biofeedback (EMG-BF) since they showed abnormally large EMG activity in one or more proximal muscles of the dominant limb while writing compared to normal maneuvers of limb. The large EMG activity was judged on the basis of comparing EMGs during normal limb maneuvers such as flexion at the elbow, extension at the elbow, and adduction at the shoulder with EMG from these muscles recorded during the act of writing. One representative recording is shown in Fig. 1. Another three patients (Nos. 3, 6, and 10) could not be given EMG-BF. Patient 3 had involvement of small extensors, patient 6 had muscle wasting following injection of BTX, and patient 10 had intention tremors. These patients did not present a so-called dystonic posture. It was thus not possible to define a biofeedback signal for them. Hence these patients were excluded from biofeedback treatment. To select the muscle for EMG-BF, the dominant limb of the patient was physically examined at rest and during writing. In each patient the arm and forearm circumferences were measured at places where the musculature was bulky and compared with corresponding sites on the contralateral side. Understandably this portion was bulky because of hypertrophied muscles. This hypertrophied portion gave an idea of the muscle(s) that was(were) likely to show excessive contractions during writing. Finally, the muscle(s) generating excessive activity was(were) confirmed by recording surface EMG from multiple sites. The electrodes (gold-plated, disk type, 0.8-mm diameter) for EMG recording were placed and secured 4 cm (center to center) apart on the most prominent part of muscle bellies. The 4-cm distance was chosen to record activity from the whole muscle and was kept constant in all sessions and all patients. EMG was recorded from the brachioradialis, biceps, triceps, deltoid, and trapezius (superior portion) either on a Grass EEG (Model 810 C) or Neuropack 8 (Nihon Kohden, Japan) polygraph. The muscle showing maximum abnormal EMG activity in a patient during writing (compared with normal maneuvers of the same limb; Fig. 1) was selected for EMG-BF. A different muscle was selected for each patient. The muscles used for biofeedback for the patients are listed in Table I. The EMG signal was converted into an audio signal using a Grass audiomonitor (Model AM 8) and was fed back to the patient. The loudness of the sound signal was directly proportional to the amplitude of the muscle EMG. The sound was attenuated when the muscle was relaxed. The patients were trained to recognize the relationship between the loudness of the sound and the contraction of the muscle. They were instructed to reduce the intensity of the sound

EMG Biofeedback for Hand Dystonia

271

Fig. 1. Kinesiologic EMG records of the brachioradialis (Br), biceps (Bi), and triceps (Tr) muscles from the right limb of a patient suffering from right-hand dystonia (30 yrs, male). Records A-E depict EMG recorded during normal maneuvers of the limb. EMG recorded (A) at rest in a semiflexed elbow posture; (B) when the limb is held straight projecting forward; (C) when the patient is holding the limb in an adducted position against a force of 1 -kg weight; (D) when the patient is holding the limb in a forced flexion at the elbow joint against a force of 1-kg weight; (E) during flexion-extension performed at the elbow joint without any load. (F) Continuous EMG tracings while writing, accompanied by poor handwriting, pain, discomfort, and dystonic posture. Dashed lines suggest that the record continues. (G) Record obtained during a biofeedback session that represents initial rest, recording while writing without biofeedback, and a final portion (as indicated by the arrow) obtained during writing while the patient was undergoing triceps EMG biofeedback. The upward arrow indicates the beginning of triceps EMG biofeedback. The dashed line with K and > suggests further continuation of writing along with EMG biofeedback. Note that during biofeedback the contraction in three muscles subsided to a level comparable to rest.

272

Deepak and Behari

coming from the amplifier. Patients were also told that they should make an attempt to reduce the sound to the minimum possible level while they were writing. Such a reduction in the EMG of the proximal muscles ensures their minimal involvement during writing. During EMG-BF, EMGs were recorded from three or four neighboring sites (one of them was used for BF) while patients were writing. During the first visit, patients' discomfort and pain level and quality of handwriting were assessed using a visual analogue scale (0-100). The use of visual analogue scales is well accepted in the behavioral sciences for assessing subjective responses. The highest end point of the scale (100) was considered by the subject as his best preillness status. The lowest (0 point) was the worst condition perceived by patient. The first value chosen by the subject was used as the baseline level for future comparisons. Discomfort and pain were scored using a 0-100 (absence of discomfort/pain to worst condition) scale and handwriting was scored 0-100 (worst handwriting to normal handwriting). To obtain a total VAS score, the individual scores were added up and divided by 3. Before adding up, the individual scores for discomfort and pain were subtracted from 100 and their sum was added to a VAS score for handwriting [total score = (100 — discomfort score) + (100 — pain score) + (handwriting score)]. Old samples of handwriting were also collected from the patients to assess the duration of problem, the pattern of deterioration, and the degree of loss in quality of handwriting. Subsequently, patients were given four or more practice sessions (once every two weeks). In the practice sessions, the patients practiced writing while undergoing biofeedback intervention. They practiced writing with relaxed proximal muscles aided by EMG biofeedback. During the initial session the patients became aware of involuntary and abnormal contractions of the distal muscles of the hand. They were told about the possible worsening of the handwriting after the initial intervention. Instructions for home practice were clearly spelled out. The patients were told to practice writing with relaxed muscles in the proximal portion of the limb. They were also told that, with practice, they would learn to write with a relaxed limb. Patient 4 was given four sessions in a period of 1 month since he could not stay longer with us. The patients were advised to practice writing with the relaxed limb every day for 5 to 10 min, depending on the severity of the condition. Discomfort, pain, and quality of handwriting were recorded by the patients using the visual analogue scale (VAS) on a day-today basis for the first 2 months. Later, such assessments were performed on a monthly basis. The change in total VAS score served as a guide for improvement over time. Both subjective (50% improvement in VAS to pain, discomfort, and quality of handwriting) and objective assessment of handwriting were the criteria for improvement. Sustained improvement for 6 months was accepted as successful intervention. RESULTS During the initial sessions of biofeedback handwriting became worse (maximum worsening in first session, see Fig. 4), but there was no discomfort or pain. Nine of the 10 patients showed overall improvement ranging from 37.5 to 93.4% on the VAS and one (No. 5) failed to show improvement (see Table II). Seven patients showed more than 50% improvement. Thus the success rate was 90% in selected (9 of 10) and 69.2% in unselected (9 of 13) patients with writer's cramp. These patients had considerably different types of clinical presentations and responses. Therefore individual cases are described and presented in Tables I and II.

EMG Biofeedback for Hand Dyslonia

273

274

Deepak and Behari

Patient 1. This patient had had a history of poor handwriting for more than 1 year, which worsened during high-anxiety states. On examination his right hand was found to be bulky. The midcircumference measurements for the arm and forearm were larger for the right limb (1 and 0.5 cm, respectively). Tremors were observed in the outstretched hand but not during writing. The patient had been well adjusted at his job but lately had interpersonal conflicts in the office. However, a psychiatric consultation did not reveal any psychiatric problem. The patient had been given long practice on writing (straight lines, shapes, etc.) by a psychologist, but to no avail. The EMG during writing showed excessive activity in the triceps muscle (see Fig. 1G) and substantial EMG activity in the biceps muscle. The triceps muscle showed more activity during writing than in any other maneuver (see Figs. 1A-E). The patient showed control over abnormal contractions of the triceps during triceps EMGBF (see Fig. IE). Four BF sessions made him confident to write with a relaxed limb. The patient showed overall improvements in symptoms ranging from 15 to 80% on the VAS over 6 months (see Table II). In 1 month's practice the writing speed improved from 6.5 to 9.5 words per min. Pain and discomfort showed a decline of 90 and 70%, respectively. The improvement continued for 2.5 years of follow-up. Patient 2. This patient used to write by putting excessive pressure on the pen. This was assessed by observing raised markings on the back of the written material. He had undergone several sessions of physiotherapy with no improvement. He was given triceps EMG-BF. Before EMG-BF, the patient could write only 3—4 lines at a time, which increased to 9-10 lines at a time in the third session (see Fig. 2). At 6 months, he was able to write 20 lines at one time.

Fig. 2. Samples of handwriting of patient 2 obtained before and after triceps EMG biofeedback (at 2 weeks, 1 month, and 3 months). On the first day the patient could write only 2 to 2.5 lines with great difficulty. The patient developed pain and discomfort at the end. After biofeedback, in 2 weeks' time, he could write about nine lines before pain and discomfort appeared. In 3 months he could write legible handwriting with ease.

EMG Biofeedback for Hand Dystonia

275

Patient 3. He was a university student who used to develop gradual extension of the ring and middle fingers to the extent that he was unable to hold the pen. Later on, pain would develop in the ring finger. Gradually, his handwriting had become clumsy. He tried braces to keep the index and the middle fingers together, but this gave him fatigue and discomfort while writing. No other proximal groups of muscles were involved. The patient had received trihexiphenidyl without any improvement and had dryness of the mouth as side effects. Patient 4. On the first day, he could not write more than two or three lines due to pain. His right forearm was 1 cm thicker than the left. This patient showed excessive EMG activity in the brachioradialis, triceps, and biceps muscles in decreasing order. During brachioradialis biofeedback, EMG activity declined in all muscles and showed a further decline after a week (see Fig. 3). Although handwriting worsened initially during biofeedback (see Fig. 4, day 1) in the fourth week (see Fig. 4) handwriting was better than as the pretreatment level. Later he showed gradual and steady improvement until 1 year of follow-up. At the end of 1 year he could write 55-60 (about double the last sample shown in Fig. 4) lines at a time. Patient 5. In addition to other features of writer's cramp, this patient had tremors, which were present while holding an object. EMG screening during writing showed enhanced activity in the brachioradialis, triceps, and biceps muscles (in decreasing order). Handwriting during brachioradialis EMG-BF did not worsen, nor was he able to reduce the sound. Four sessions (over 3 months, compared to 2 months for the other patients) of biofeedback did not lead to any sign of improvement. The patient was not very cooperative during biofeedback sessions and was often absent on the scheduled date. Patient 6. This patient had had difficulty in writing for a period of 4 years. While writing, his index finger gradually extended, thus making the pen slip. While writing he did not use excessive force to hold the pen or to write. He had been given trihexiphenidyl (6 mg) in the past without any benefit. In the recent past he had been given injection of BTX into the brachioradialis muscle. He had wasting of the right forearm as evident by two observations; first, the circumference of the forearm was 1 cm less than on the left side; second, he showed decreased muscle strength on neurological examination. He was not found suitable as per "criteria a priori," as no proximal muscle was found to be hyperactive during writing. Patient 7. This patient had resting tremors. Excessive contractions were seen in the brachioradialis and long flexors of the right hand while writing. After brachioradialis BF he showed improvement within 2 months. Patient 8. This patient showed continuous contraction of the brachioradialis and his index finger showed increasing extension while writing. The patient used to develop pain in the elbow after writing for a couple of minutes. The discomfort continued for some time even after the writing was stopped. The deltoid also showed contraction while writing. The patient had swelling of the small joints, which was not due to any structural deformity or gout. His antistreptolysin O (ASO) titer was negative. He showed improvement after 1.5 months (brachioradialis EMG-BF, four sessions). The EMG activity of the trapezius muscle increased at min 6 of writing, which otherwise was silent. Since contractions in the trapezius muscle have been correlated with increased anxiety and sympathetic activity, this suggested increased sympathetic tone toward the end of the writing effort. Patient 9. Because of continuous deterioration in handwriting, this patient was forced to shift from the accounts department to administration. His handwriting further deteriorated over 2 years. He received trihexyphenidyl (6 mg/day), which gave him approximately 20% relief. During writing, the patient showed enormous EMG activity in the brachioradialis and

276

Deepak and Behari

Fig. 3. Kinesiologic records of EMG from the brachioradialis (Br), biceps (Bi), and triceps (Tri). This patient (No. 4) was given brachioradialis EMG biofeedback. Rtri represents the rectified triceps EMG. The upper panel represents the EMG taken on day 1 at rest (left record), while writing without biofeedback (middle record), and while writing and undergoing brachioradialis biofeedback simultaneously (right record). The lower panel depicts records obtained during similar conditions on the seventh day of the biofeedback intervention. After 1 week, the resting EMG records are comparable (see left column, day 1 and day 7). There is a reduction in amplitude of the EMG of all three muscles after 1 week of biofeedback intervention (see the middle column). The right column shows that the brachioradialis EMG biofeedback causes an immediate reduction in EMG amplitude in all muscles, with a maximum effect in the brachioradialis EMG. This patient showed improvement at 1 month and continued to show further improvement up to the fourth month of follow-up (see Fig. 4).

EMG Biofeedback for Hand Dystonia

277

Fig. 4. Samples of handwriting of a patient (No. 4) with dystonic writer's cramp obtained before, during (at days 1, 7, and 28), and 2, 3, 4, and 6 months after biofeedback intervention. Day 1 shows a sample written with his usual habit, which was accompanied by symptoms of pain and dystonic posture. The patient could not write more than 6.5 lines. The next record was obtained on the same day (day 1) when the patient was writing and simultaneously undergoing brachioradialis EMG biofeedback. (EMG records during biofeedback are depicted in Fig. 3.) Notice the deterioration in handwriting. During biofeedback the patient was asked to keep the brachioradialis relaxed. A noticeable improvement in handwriting while undergoing EMG biofeedback is seen at week 4. Now the patient could write material of similar length as on day 1. Later handwriting, in months 2, 3, 4, and 6, showed further improvement in handwriting. The last four samples (labeled post-BF intervention) were obtained during a no-biofeedback situation and represent learned control over proximal muscles.

a marked increase in triceps EMG activity. During brachioradialis EMG-BF, the activity of the brachioradialis muscle was remarkably reduced. The activity in the triceps almost disappeared. However, the trapezius EMG showed high EMG activity during biofeedback. Four biofeedback sessions with everyday practice at home improved the handwriting significantly. Patient 10. This patient used to develop rigidity and sudden pain soon after the beginning of writing. This pain used to be located in the region of the tips of the fingers. This was followed by tremors. He was not able to write more than six or seven lines. On examination, the right arm and forearm were found to be hypertrophied (1.5 and 1 cm thicker, respectively, than the left limb at corresponding points). Biofeedback could not be given because of the presence of tremors, which interfered with the amplifier sound. Patient 11. He was having discomfort and pain while writing. The pain was localized in the area of the brachioradialis. He had fine tremors in both hands. This patient was on prolonged carbamazepine therapy given to him after head injury. He showed improvement with brachioradialis EMG-BF.

278

Deepak and Behari

Patient 12. He noticed deterioration in handwriting first about 1.5 year prior to the assessment. This was followed by pain in the left hypothenar eminence of the palm. This patient was left-handed. He had no problem doing jobs such as buttoning shirts, using a screwdriver, and holding a spoon with the left hand. On examination the left limb was found 1 cm thicker than the right at the level of the brachioradialis. He was on treatment with trihexiphenidyl (6 mg/day), which gave him about 40% relief (based on VAS scale for pain and discomfort). However, he reported two side effects, namely, dryness of the mouth and increase in duration of sleep by 1 hr. With brachioradialis EMG biofeedback, he showed improvement in 3 months. Patient 13. This patient had a history of gradual deterioration in his handwriting over 2 years. He had problem in holding the pen. He was a cashier and had great difficulty in maintaining the cashbook. He had received 6 mg of trihexiphenidyl daily for 6 months. On examination his right forearm was found to be 1 cm thicker than the left at the corresponding points. He benefited from brachioradialis EMG biofeedback.

DISCUSSION In the present study specific muscle EMG biofeedback was used in patients with writer's cramp. The patients learned to relax the proximal limb muscles, which showed sustained contractions during writing. Nine of 10 patients showed both subjective and objective improvement. It was seen that patients could continue trihexyphenidyl and derive benefits from biofeedback. However, discontinuation of the drug did not interfere with the favorable outcome of EMG-BF as seen in two patients (8 and 11) who discontinued the medicine on their own. To our knowledge, this is the first attempt to use the EMG of proximal hyperactive limb muscles as a biofeedback signal. There are sporadic reports on the use of EMG biofeedback for the treatment of focal hand dystonia as mentioned earlier in this report. The approach we used is more rational, i.e., proper selection of patients, definite criteria for muscle to be used for biofeedback, avoiding the use of the general relaxation procedure (which could have colluded the results), and fairly long follow-up. Moreover, description of all patients registered for BF provides control and an insight into the likely causes of failure of BF. We gave BF treatment to a sizable number of patients, i.e., 10 of 13 patients who approached us. A web-based literature search suggests that in the last 25 years about 23 patients have been treated in six studies [four studies reviewed by Ince et al. (1986) and two more separate studies (Lees et al., 1989; O'Neill et al., 1996)]. The specific muscle biofeedback technique has certain limitations. It is difficult to record electrical activity when only small muscles of the hand are affected. Even if the EMG of small muscles is recorded by inserted wire electrodes, it may not be useful for biofeedback since contraction of the same group of muscles is essential for writing. This was exemplified by patient 3, who showed involuntary extension of the middle and ring fingers during writing. Poor compliance and inability to relax targeted muscle groups during biofeedback sessions are limiting factors for successful results. Other limitations are the presence of associated tremors while writing as seen in patient 5 and recent treatment with BTX. The present study also has an inherent limitation. It was a retrospective review of a relatively informal case series. Nevertheless, we have given adequate attention to details and the collection of possibly useful data. The patients with

EMG Biofeedback for Hand Dystonia

279

hand dystonia present a heterogeneous spectrum of involvement of muscles (Cohen & Hallett, 1988). In spite of these limitations, beneficial effects of the specific biofeedback strategy seem to be present. It appears that patients with hand dystonia involving larger muscle groups are likely to profit from EMG-BF. The presence of muscular hypertrophy and excessive EMG during writing are possible indications for potential benefit from EMG-BF. It seems that the EMG biofeedback from inappropriately contracting muscle results in nonactivation these muscles. Since some of the symptoms are produced by overactivation of proximal muscles, such nonactivation or reduced activation would result in an amelioration of symptoms. This is precisely what was observed during selective muscle EMG-BF. Such reduced activation may be brought about by attentional factors. During EMG biofeedback the patient becomes aware of overactivation and he is in a better position to manipulate muscle tension through attention. This could operate at a voluntary or an involuntary level. This finding is supported by the fact that event-related potentials of cortex are susceptible to attention and voluntary influences. Selective EMG-BF may be superior to treatment by injection of BTX in two ways: first, it influences a relatively larger number of muscle groups and reduces their overactivity; and second, unlike a BTX injection, it does not permanently decrease the performance of the muscle. In conclusion, biofeedback reduces overactivity of muscle groups responsible for dystonic posture, thereby alleviating symptoms. Since overactivity of proximal muscle is associated with the act of writing, the patients learn to undo it during the biofeedback session. This learning later, as we have observed, is manifest without the aid of EMG biofeedback. Thus EMG-BF may work in two ways: first, by making the subject aware of abnormal contractions and, second, by controlling/reducing excessive cocontraction in involved muscle groups. This could be brought about by voluntary decreased flow of motor activity (efferent control of motor mechanism) or by decreasing abnormal sensory input from muscle (muscle efferent block). One study has speculated that muscle afferents play a pivotal role in dystonia (Sawamsto et al., 1995). Since most of the symptoms are due to abnormal contractions of agonists and antagonists, such relaxation of local muscle groups leads to alleviation of discomfort and pain and correction of the dystonic posture. Improvement by biofeedback in no way denies the existence of organic lesions in the brain, as biofeedback may be working at a volitional motor learning level. There is also a possibility of the existence of local or general excessive sympathetic activation in these patients during writing. This point is favored by the tonic nature of the contraction of muscles involved in the dystonic posture and appearance of pain. Two arguments are being put forth in this connection. First, in at least two cases we noticed gradually increasing contraction of the trapezius muscle, indicating general sympathetic activation. Second, there is the possibility that local sympathetic activation might have resulted in compromised local circulation, giving rise to discomfort and pain. In this regard it is worth noting that sympathetically maintained pain is a well-known entity. However these points remains to be tested. It would be worthwhile to record certain local autonomic variables and trapezius EMG during the writing effort and BF intervention in future studies. There is also the possibility of y motor neuron dysfunction (specifically, hyperfunction), which normally enhances tone and provides background support for voluntary activities. This hyperfunction may have its origin at the spinal or brain-stem level. Such disinhibition in the y motor neuron circuitry has been proposed as a possible mechanism for myofascial pain symptoms (Donaldson, Nelson, & Shultz, 1998). The involvement of cortical sensory areas has been shown by elegant studies, both in terms of anatomical

280

Deepak and Behari

changes (Elbert et al., 1998; Byl et al., 1997) and by physiologic variables (Tempel & Perlmutter, 1993). However, what neurophysiologic processes operates during a biofeedback intervention needs to be studied. Studies using PET to image the primary sensory area and supplementary motor area during the initial period of writing and during EMG biofeedback may further shed light on the role of the central nervous system in the genesis of hand dystonias and CNS events during biofeedback. Thus, specific muscle EMG biofeedback appears to be a promising nonpharmacologic intervention procedure to manage dystonias. An attempt may be made to extend this method to other types of dystonias.

REFERENCES Byl, N. N., Merzenich, M. M., Cheung, S., Bedenbaugh, P., Nagarajan, S. S., & Jenkins, W. M. (1997). A primate model for studying focal dystonia and repetitive strain injury: Effects on primary somatosensory cortex. Physical Therapy, 77, 269-284. Candia, V., Elbert, T., Altenmuller, E., Rau, H., Schafer, T., & Taub, E. (1999). Constraint-induced movement therapy for focal hand dystonia in musicians. Lancet, 353, 42. Cohen, L. G., & Hallett, M. (1988). Hand cramps: Clinical features and electromyographic patterns in a focal dystonia. Neurology, 38, 1005-1012. Cottraux, J. H., Juenet, C., & Collet, L. (1983). The treatment of writer's cramp with multimodal behaviour therapy and biofeedback: A study of 15 cases. British Journal of Psychiatry, 142, 180-183. Donaldson, C. C. S., Nelson, D. V., & Schulz, R. (1988). Dis-inhibition in the gamma motoneuron circuitry: A neglected mechanism for understanding myofascial pain syndromes? Applied Psychophysiology and Biofeedback, 23, 43-57. Elbert, T., Candia, V., Altenmuller, E., Rau, H., Sterr, A., Rockstroh, B., Pantev, C., & Taub, E. (1998). Alteration of digital representations in somatosenEory cortex in focal hand dystonia. Neuroreport, 9, 3571-3575. Hughes, M., & McLellan, D. L. (1985). Increased co-activation of the limb muscles in writer's cramp. Journal of Neurology, Neurosurgery and Psychiatry, 48, 782-787. Ince, L. P., Leon, M. S., & Christidis, D. S. (1986). EMG biofeedback for handwriting disabilities: A critical examination of the literature. Journal of Behavioral Therapy and Experimental Psychiatry, 17(2), 95-100. Kaji, R., Kohara, N., Katayama, M., Kubori, T., Mezaki, T., Shibasaki, H., & Kimura, J. (1995). Muscle afferent block by intramuscular injection of lidocain for the treatment of writer's cramp. Muscle Nerve, 18, 234-235. Karp, B. H., & Hallett, M. (1994). Botulinum toxin treatment of focal hand dystonias. In J. Janovic & M. Hallett (Eds.), Therapy with botulinum toxin (pp. 299-306). New York: Mercel Dekker. Lees, A. J., Kleedorfer, B., & Foster, H. (1989). Treatment of writers' dystonia. Lancet, 30, 1525. Marsden, C. D., & Sheehy, M. P. (1990). Writer's cramp. Trends in the Neurosciences, 13, 148-153. O'Neill, M. E., Gwinn, K. A., & Adler, C. H. (1996). Biofeedback for writer's cramp. American Journal of Occupational Therapy, 51, 605-607. Poungvarin, N. (1991). Writer's cramp: The experience with botulinum toxin injection in 25 patients. Journal of Medical Association of Thailand, 75, 239-247. Rhoad, R. C., & Stem, P. J. (1993). Writer's cramp—a focal dystonia: Etiology, diagnosis, and treatment. Journal of Hand Surgery, 18, 541-544. Ross, M. H., & Charness, M. E. (1996). Occupational cramp. In M. A. Samuels & S. Feske (Eds.), Office practice of neurology (pp. 678-680). New York: Churchill Livingstone. Sawamsto, N., Kaji, R., Katayama, M., Kuboli, T., & Kumara, J. (1995). Muscle afferent block for the treatment of writer's cramp. Rinsho Shinkeigaku, 35, 1210-1213. Sheean, G. L., Murray, N. M., & Marsden, C. D. (1995). Pain and remote weakness in limbs injected with botulinum toxin A for writer's cramp. Lancet, 346, 154-156. Tempel, L. W., & Perlmutter, J. S. (1993). Abnormal cortical responses in patients with writer's cramp. Neurology, 43, 2252-2257. Thompson, P. D. (1993). Writer's cramp. British Journal of Hospital Medicine, 50, 91-94.