ORIGINAL STUDY
Biofeedback Treatment for Tourette Syndrome: A Preliminary Randomized Controlled Trial Yoko Nagai, PhD,* Andrea E. Cavanna, MD, PhD,w Hugo D. Critchley, DPhil, FRCPsyc,*z Jeremy J. Stern, FRCP,y Mary M. Robertson, MD, FRCPsyc,y and Eileen M. Joyce, MD, FRCPsyc8
Objective: To study the clinical effectiveness of biofeedback treatment in reducing tics in patients with Tourette syndrome. Background: Despite advances in the pharmacologic treatment of patients with Tourette syndrome, many remain troubled by their tics, which may be resistant to multiple medications at tolerable doses. Electrodermal biofeedback is a noninvasive biobehavioral intervention that can be useful in managing neuropsychiatric and neurologic conditions. Methods: We conducted a randomized controlled trial of electrodermal biofeedback training in 21 patients with Tourette syndrome. Results: After training the patients for 3 sessions a week over 4 weeks, we observed a significant reduction in tic frequency and improved indices of subjective well-being in both the active-biofeedback and sham-feedback (control) groups, but there was no difference between the groups in these measurements. Furthermore, the active-treatment group did not demonstrably learn to reduce their sympathetic electrodermal tone using biofeedback. Conclusions: Our findings indicate that this form of biofeedback training was unable to produce a clinical effect greater than placebo. The main confounding factor appeared to be the 30-minute duration of the training sessions, which made it difficult for patients to sustain a reduction in sympathetic tone when their tics themselves were generating competing phasic electrodermal arousal responses. Despite a negative finding in this study, electrodermal biofeedback training may have a role in managing tics if optimal training schedules can be identified.
Received for publication January 16, 2013; accepted April 15, 2013. From the *Brighton and Sussex Medical School, University of Sussex, Falmer Campus, Brighton, UK; wDepartment of Neuropsychiatry, Birmingham and Solihull Mental Health National Health Service Foundation Trust and University of Birmingham, Birmingham, UK; zSackler Centre for Consciousness Science, University of Sussex, Brighton, UK; ySt George’s Hospital and Medical School, London, UK; and 8Sobell Department of Motor Neuroscience, University College London Institute of Neurology, London, UK. Supported in part by the Tourette Syndrome Association. The authors declare no conflicts of interest. Reprints: Yoko Nagai, PhD, Brighton and Sussex Medical School, University of Sussex, Falmer Campus, Brighton BN1 9RR, UK (e-mail: [email protected]). Copyright r 2014 by Lippincott Williams & Wilkins
Cogn Behav Neurol
Volume 27, Number 1, March 2014
Key Words: biofeedback, electrodermal activity, sympathetic autonomic arousal, tics, Tourette syndrome (Cogn Behav Neurol 2014;27:17–24)
Reader Benefit: Electrodermal activity biofeedback may play a role in reducing tics in patients with Tourette syndrome if optimal training schedules can be determined. TS = Tourette syndrome.
T
ourette syndrome (TS) is a neurodevelopmental disorder affecting between 0.4% and 3.8% of schoolchildren worldwide, and with a calculated prevalence of 1% of the general population (Robertson, 2008). Patients with TS suffer from multiple involuntary movements (motor tics) and 1 or more phonic or vocal tics that last longer than a year. These tics can cause patients social discomfort that may affect their academic or work performance. Many patients perceive such secondary effects of TS as more psychologically problematic than the tics themselves (Robertson, 2000, 2012). Medications are typically the first-line treatment for managing tics, with neuroleptics most commonly prescribed. Because patients commonly experience side effects from these drugs (Robertson, 2000), nonpharmacologic approaches have long been sought. Behavioral interventions such as habit reversal training are helpful as alternative or adjunctive treatments (Piacentini and Chang, 2001). Habit reversal training encompasses self-monitoring, relaxation, competing response practice, and contingency management training to reinforce behavior with reward and punishment. Piacentini et al (2010) reported significantly lower tic severity in children with TS who had received the Comprehensive Behavioral Intervention for Tics than in a control group. This therapy consists of habit reversal training, relaxation training, and identification of vulnerable situations. Although habit reversal training is the primary component, all of these elements have been used to manage tics. Other established behavioral approaches include cognitive therapy, assertiveness training, and self-monitoring (Robertson, 2000). Although studies examining www.cogbehavneurol.com |
17
Nagai et al
the efficacy of behavioral treatments are mostly encouraging, these approaches are indirect and do not clearly target the pathogenetic mechanisms underlying TS. Biofeedback encompasses a set of behavioral or psychophysiological treatments that have been gaining popularity in the management of neurologic, psychological, and psychiatric conditions over the last few decades. Biofeedback is used for conditions such as anxiety disorders, epilepsy, asthma, some cardiovascular disorders, hypertension, irritable bowel syndrome, neuromuscular disorders, and migraine and tension headaches. Biofeedback characteristically enables a patient to gain voluntary control over covert physiological responses by making these responses explicit through real-time visual or auditory feedback. Patients are typically able to learn how to modify these physiological processes volitionally. Such change is thought to be achieved through an indirect combination of psychological (eg, changes in attentional focus) and physiological (eg, changes in muscle tension or in rate and depth of breathing) mechanisms, although patients do not necessarily consciously recognize why they are doing better. The earliest attempts at biofeedback treatments for TS, made in the 1980s, used electroencephalography. The target signal was enhancement of the sensory motor rhythm, an electroencephalographic frequency component of 12 to 14 Hz that emerges over the rolandic sensorimotor area when a person is still (Tansey, 1986). Sensory motor rhythm biofeedback was first explored to reduce seizure frequency in epilepsy (Sterman and Friar, 1972), and was later reported as successful in 2 patients with TS who showed a dramatic reduction in tic frequency (Tansey, 1986). The biofeedback sessions for the patients with TS lasted 30 minutes once a week; however, the overall length of the treatment seemed individually tailored. One of the patients with TS benefited after 5 weeks of treatment, and the other after 18 weeks. Still, given the lack of quantitative outcome measures in Tansey’s 2 patients, nonspecific psychological effects of treatment cannot be excluded as explanatory factors in their improvement. In a pilot study of biofeedback in TS (Nagai et al, 2009a), we tested an easy-to-implement approach targeting peripheral sympathetic arousal indexed by electrodermal activity, which reflects sympathetic nervous control of the activity of eccrine sweat glands. In an earlier neuroimaging study (Nagai et al, 2004a), we had found that reduction of electrodermal sympathetic tone was associated with greater activity within the orbitofrontal and ventromedial prefrontal cortex, regions also associated with affective responses and default (disengaged) brain states. We have also shown that electrodermal sympathetic activity is inversely related to an index of cortical arousal, slow cortical potentials, evoked experimentally using a contingent negative variation paradigm (Nagai et al, 2004c). We found brain correlates of induced cortical arousal in heightened thalamocortical activity that was related to the amplitude of slow cortical potentials (Nagai et al, 2004b). Our observations supported a proposal that
18 | www.cogbehavneurol.com
Cogn Behav Neurol
Volume 27, Number 1, March 2014
peripheral electrodermal modulation can influence thalamocortical regulation (Nagai et al, 2004b), with potential therapeutic implications for neurologic conditions, notably epilepsy (Nagai et al, 2004d; Nagai, 2011). Patients with TS have been reported to have abnormalities of the contingent negative variation: an attenuation of the early part of the wave but an enhancement of the later (dopaminergically regulated) component (van Woerkom et al, 1988; Weate et al, 1993). This pattern suggests dysregulation of thalamocortical excitability in patients with TS. This perspective is supported by the generally positive outcomes of thalamic deep brain stimulation as a treatment for TS (Ackermans et al, 2011). In our proof-of-principle TS study (Nagai et al, 2009a), which lacked an explicit control condition, we tested the short-term effects of using electrodermal biofeedback to raise and lower sympathetic activity on the expression of tics in 15 patients. We found that a reduction in sympathetic tone, elicited using electrodermal biofeedback for 5 minutes, was associated with a reduction in tic frequency (Figure 1). Before this pilot study, it had not been clear whether the greater attentional focus associated with increasing peripheral sympathetic tone would also be beneficial in TS, as it is in epilepsy (Nagai et al, 2004d). This proved not to be the case. Another uncertain factor had been the best training regimen for TS. Our epilepsy treatment study (Nagai et al, 2004d) had established an optimal protocol for electrodermal biofeedback in adults with epilepsy as 30-minute sessions, 3 times a week. On this protocol, the patients showed treatment effects within 4 weeks. Although the role of central autonomic control on tic activity has not been comprehensively investigated, patients with TS generally have the same heart rate and blood pressure as those in healthy people (Goetz et al, 1987; van Dijk et al, 1992). Drugs that act on sympathetic tone can have therapeutic benefit. Clonidine, an alpha 2 receptor agonist, is often 1 of the first-line drugs given for TS, and cholinergic (typically nicotinic) agents are also reported to reduce symptoms (Robertson, 2000). There is a single case report that vagus nerve stimulation reduced motor and phonic tic activity in a patient with TS (Diamond et al, 2006). Thus, there is evidence that both sympathetic and parasympathetic factors influence tic expression, and our pilot TS study (Nagai et al, 2009a) suggested that acute attenuation of sympathetic activity promptly reduced tic activity. In the current study, we conducted a randomized controlled trial to examine the possible translation of electrodermal biofeedback to a treatment for patients with TS. Although the study was not strictly double blind, the researcher who counted patients’ tics was blinded to the treatment condition.
METHODS Participants We recruited 21 patients (26.7 ± 10.6 years old) with TS from the National Hospital for Neurology and r
2014 Lippincott Williams & Wilkins
Cogn Behav Neurol
Volume 27, Number 1, March 2014
Biofeedback Treatment for Tourette Syndrome
Arousal Biofeedback Relaxation Biofeedback
Number of Tics
50
*
40
*
30 20 10
*
0
Relaxation Biofeedback Motor
Face
Arousal Biofeedback Body Type of Tics
Vocal
Total
FIGURE 1. Average number of tics during brief 5-minute periods of relaxation and arousal biofeedback in 15 patients with Tourette syndrome in the authors’ proof-of-principle study (Nagai et al, 2009a). Motor tics include both face and body tics. Total tics include face, body, and vocal tics. Relaxation biofeedback significantly reduced the rates of tics over arousal biofeedback in the scores for motor, face, and total tics. The asterisks denote significant differences. Reproduced from Nagai Y, Cavanna A, Critchley HD. 2009. Influence of sympathetic autonomic arousal on tics: implications for a therapeutic behavioral intervention for Tourette syndrome. J Psychosom Res. 67:599–605, with permission from Elsevier.
Neurosurgery and from the Department of Neurology at St George’s Hospital, both in London. All patients satisfied the Diagnostic and Statistical Manual of Mental Disorders, 4th edition, Text Revision (American Psychiatric Association, 2000) criteria for the diagnosis of TS. We excluded patients with a history of psychotic illness, personality disorder, or major medical or neurologic diseases, including movement disorders such as Parkinson disease, Wilson disease, Huntington disease, and dystonia. We also required patients not to have had any changes in their medication regimen for at least 1 month before the study began, and no change likely during the 4-week study. The 21 patients were randomly assigned to either the active-biofeedback (n = 12) or sham-control (n = 9) group, according to the randomization table. Six patients in the biofeedback group and 2 in the control group were taking neuroleptics (haloperidol, risperidone, pimozide, sulpiride, or aripiprazole), and 1 patient in each group was taking clonidine. Two patients in the biofeedback group and 1 in the control group were taking a selective serotonin reuptake inhibitor (fluvoxamine or fluoxetine). One patient in the biofeedback group was taking lamotrigine. The study was approved by the local Ethics Committee, and all patients provided informed written consent to the procedure.
Procedures Biofeedback Training Protocol During the 4-week treatment, the patients attended a 30-minute biofeedback session 3 times a week, for a total of 12 sessions (Table 1). During each session, the therapist instructed the patients to attend to, and change, the animation on a computer screen by relaxing both mentally and physically. The patients assigned to the active-treatment group received true biofeedback from the r
2014 Lippincott Williams & Wilkins
computer screen, and the patients assigned to the control group received sham biofeedback. Electrodermal Recording and Biofeedback To measure electrodermal activity and deliver biofeedback, we used a purpose-built biofeedback system (Inner Tuner: Professional, Ultrasis plc, London, UK) modified for experimental use. We placed 2 dry nickelplated electrodes on the palm of the patient’s left hand. The area of each electrode was approximately 1 cm2. The sampling rate of the recording was 10 Hz. Biofeedback took the form of computer-generated graphics displayed on a computer monitor in front of the patient. When patients in the active-biofeedback group relaxed and their skin conductance dropped, an animated figure in the foreground of the computer screen moved rightward through a successive sequence of animations. The sequence started with a picture of fish, which turned into a mermaid, then a walking lady, and then a star. When patients’ skin conductance increased, the display moved leftward, returning to an earlier part of the animation sequence. Thus, the active-biofeedback condition “rewarded” patients for reducing their electrodermal sympathetic activity by showing them more of the animation sequence. The patients in the sham-control group viewed the same computer-generated graphics as those used in the
TABLE 1. Design of This Study Before Treatment Treatment Assessment 1. Video monitoring of tics 2. Questionnaires
Electrodermal biofeedback treatment: 30-minute sessions 3 times/week for 4 weeks
After Treatment Assessment 1. Video monitoring of tics 2. Questionnaires
www.cogbehavneurol.com |
19
Nagai et al
active-biofeedback condition, and also tried to reduce their electrodermal activity level using the evolving sequence of animations. However, in contrast to the activebiofeedback group, here the video animation changed direction unrelated to the ups and downs in the patients’ own electrodermal activity. Thus, the sham-control patients could not learn direct biofeedback control. To avoid frustrating the patients, we altered the progression of their animation changes across sessions to match the sequences that the patients in the active-biofeedback group were generating by their changes in electrodermal activity. Because the sham-control patients believed that they were actively performing the task, they did not report negative perceptions of having failed to learn to relax. Video Recording and Psychological Assessment To assess the effects of treatment on our patients’ tic severity, we started the first and last biofeedback sessions by counting tics and then taking a psychological profile. We recorded each patient’s tics with a digital video camera (Sony, DVD, Sony Europe). The patients sat on a comfortable chair and relaxed for 10 minutes while the camera recorded their whole-body movements. Author A.E.C., a neurologist with special expertise in TS and blinded to the treatment condition, viewed the video recording and counted the tics. After the video recording, we gave the Yale Global Tic Severity Scale (Leckman et al, 1989) to quantify the patients’ self-reported symptoms of TS. To assess comorbid symptoms of obsessive-compulsive disorder, depression, and anxiety, we gave the Yale-Brown Obsessive Compulsive Scale (Goodman et al, 1989), Beck Depression Inventory (Beck et al, 1961), and State-Trait Anxiety Inventory (Spielberger, 1983). We gave the Gilles de la Tourette Syndrome Quality of Life Scale (Cavanna et al, 2008) to assess self-reported changes in patients’ healthrelated quality of life. We gave the Social Readjustment Rating Scale (Holmes and Rahe, 1967), commonly called the Holmes and Rahe Stress Scale, to monitor the stress of external life events that patients experienced during the study.
Data Analysis We calculated the change in each patient’s electrodermal activity during each biofeedback session and then averaged those changes over the 12 sessions. We used paired t tests to investigate significant electrodermal changes between the beginning and end of each session in both study groups. To measure the progress in performance of biofeedback, we calculated the average electrodermal activity changes over the first 3 and last 3 sessions. This permitted us to determine if patients improved in the degree to which they could change their skin resistance level over the month of active- or sham-biofeedback treatment. We used independent t tests to examine the differences in electrodermal activity between the active-biofeedback and sham-control groups.
20 | www.cogbehavneurol.com
Cogn Behav Neurol
Volume 27, Number 1, March 2014
We calculated the number of tics before and after treatment. The main outcome measure was the change in number of tics from the baseline period. We categorized the types of tics as motor (body or facial) and vocal. Total tics represented the sum of the motor and vocal tics. We used repeated measures analyses of variance to analyze the mean number of tics before and after treatment in each group. We corrected degrees of freedom using Greenhouse-Geisser estimates. We further explored all significant main effects and interactions using post hoc t tests. We performed paired t tests to investigate changes in the psychological questionnaires before and after the intervention in both study groups. We analyzed differences between these conditions using independent t tests. We also performed Pearson correlation analyses of the psychological and physiological data. We set significance at P < 0.05.
RESULTS Statistical analysis revealed that there were no group differences in the patients’ age, initial psychological state, or tonic electrodermal activity level. However, the number of tics (mean ± standard deviation) at baseline at the first session was significantly higher in the active-biofeedback group (143.17 ± 97.55) than in the control group (43.00 ± 33.52). Nine patients (6 in the biofeedback group and 3 controls) dropped out in the middle of the treatment, citing lack of motivation. To avoid the effects of dropout, we used intention-to-treat data analysis.
Changes in Electrodermal Activity We compared the tonic level of electrodermal activity at the beginning and end of each 30-minute biofeedback session. All of the patients in the activebiofeedback group significantly changed their tonic electrodermal activity level (t = 3.43, P = 0.006) (Figure 2). However, despite the instruction to relax their physiological state, the patients in the active-biofeedback group actually increased their sympathetic arousal, as indexed by their skin response. Their mean skin resistance at the beginning of biofeedback sessions was 1035.24 ± 339.44 kO, and at the end it had fallen to 735.15 ± 422.29 kO. The control group showed no significant difference in skin response between the beginning and the end of each session. Their mean skin resistance at the beginning of biofeedback sessions was 921.47 ± 571.09 kO, and, at the end, 922.60 ± 933.72 kO. The difference in electrodermal activity change between the active-biofeedback and sham-control groups did not reach significance. Over the course of training, the patients in both groups were unable to improve their biofeedback skills (t = 1.97, P = 0.064).
Changes in Tic Activity Figure 3 shows the number of all types of tics before and after the intervention in both study groups. Despite both groups’ inability to change their electrodermal r
2014 Lippincott Williams & Wilkins
Cogn Behav Neurol
3500 3000
Volume 27, Number 1, March 2014
Correlations Among Electrodermal Activity, Tics, and Questionnaires
Biofeedback Series21 Control
Kohm
2500 2000 1500 1000 500 0 Beginning
End
FIGURE 2. Change in electrodermal activity (galvanic skin response) at the beginning and end of each 30-minute session in all 21 patients in this study, averaged over the 12 biofeedback sessions. Despite being instructed to increase their skin resistance level, most patients in both groups reduced their skin resistance. (In the biofeedback group, the patients with the second- and third-lowest beginning levels had essentially overlapping traces. For the patients who dropped out before the end of the study, we averaged the results of their completed sessions.)
activity, we found a significant main effect of Intervention Time in reducing the number of tics (F1,19 = 4.68, P = 0.04). However, there was no significant interaction with group, suggesting no difference in the effects of biofeedback and sham control. The mean total number of tics in the biofeedback group was 143.17 ± 97.55 before and 110.25 ± 77.69 after intervention; in the control group, the mean was 43.00 ± 33.52 before and 21.22 ± 19.65 after. The number of motor tics in the combined groups was significantly lower (t = 2.21, P = 0.039), especially body tics (t = 2.20, P = 0.039). Facial and vocal tics were unchanged.
Evaluation of the Questionnaires Independent sample t tests revealed no difference in psychological and clinical state between the 2 groups at the beginning of the study; neither was there any significant difference between the groups in the changes in pre- and post-treatment questionnaire scores. Both groups had a tendency for Beck Depression Inventory scores to improve (t = 1.95, P = 0.067) (Figure 4). Both groups showed a significant reduction in Yale-Brown Obsessive Compulsive Scale scores (t = 2.427, P = 0.034) and a significantly improved Tourette Syndrome Quality of Life Scale score (t = 2.45, P = 0.024). There was no significant difference in the Holmes and Rahe Stress Scale, suggesting that the patients’ external circumstances had not changed during the study period. r
2014 Lippincott Williams & Wilkins
Biofeedback Treatment for Tourette Syndrome
For the most part, we found no correlation between changes in patients’ tics and psychological questionnaires, suggesting that the overall reduction of tics could not be accounted for by psychological change. One exception was the observed correlation between a reduction of vocal tics and improvement of obsessive-compulsive symptoms (P = 0.004, Pearson correlation coefficient [r] = 0.60). Changes in vocal tics were also related to performance on the biofeedback task over the first 3 sessions. Overall, treatment was unsuccessful. Patients who increased their sympathetic activity despite their intention to relax had the greatest reduction in frequency of vocal tics (P = 0.016, r = 0.57). These patients had also had more vocal tics at the beginning of treatment (P = 0.023, r = 0.49).
DISCUSSION Understanding of the central neural mechanisms of TS has improved through advances in neuroimaging studies and the development of interventions such as deep brain stimulation. However, the relationships among attentional control, emotional stress, autonomic arousal, and the expression of tics have yet to be characterized comprehensively. In our proof-of-principle study of TS (Nagai et al, 2009a), we showed that effective biofeedback modulation of electrodermal activity to decrease sympathetic tone could diminish tic activity, while biofeedback to increase sympathetic arousal could not. We hypothesized that biofeedback might reduce tics mechanistically through psychophysiological changes generated by the sympathetic modulation. Electrodermal biofeedback alters both activity within the medial orbitofrontal cortex and amplitude of slow cortical potentials, indexing cortical excitability (Nagai et al, 2004a, 2009b). TS has been linked in part to abnormal function of the orbitofrontal cortex (George et al, 1992; Worbe et al, 2010), and most patients with TS have low preparatory motor potentials (O’Connor et al, 2001; van Woerkom et al, 1988). Thus, by affecting cortical excitability, electrodermal biofeedback may influence neural networks associated with the generation of motor tics. Our pilot study (Nagai et al, 2009a) led to the current study aimed at determining whether training patients with TS to decrease their sympathetic activity volitionally using electrodermal biofeedback could reduce their tics. Although our patients succeeded in modulating their sympathetic activity briefly using biofeedback, we were unable to find evidence that our methods enabled the patients to learn to control their sympathetic arousal or to benefit with a reduction in tics. This outcome might reflect the relatively high dropout rate in our active-biofeedback group. Our successful outcome study using this method in patients with epilepsy did not have dropouts (Nagai et al, 2004d). Thus, effective biofeedback performance and accompanying clinical improvement appear to be crucial factors in www.cogbehavneurol.com |
21
Cogn Behav Neurol
Nagai et al
Both Bothconditions condition
Control
Volume 27, Number 1, March 2014
Biofeedback
160 *
Number of Tics
140
*
120 100 80
*
60 40 20 0 Before After
Before After
Motor
Before After
Face
Before After
Before After
Body Vocal
Total
Type of Tics
FIGURE 3. Averaged number of tics at the beginning and end of the 12 galvanic skin response biofeedback treatment sessions. The asterisks denote significant differences.
ensuring that people will complete biofeedback treatment. We suspect that frustration with poor performance might have been even higher in our active-biofeedback group than in our controls because the display truly mirrored the patients’ physiological state. One critical difference between our pilot TS work (Nagai et al, 2009a) and this study was the duration of biofeedback sessions. In the pilot study, patients guided by electrodermal biofeedback significantly raised and lowered their sympathetic tone over periods of 5 minutes. However, for the current study we adapted the protocol that we had used successfully in biofeedback training for epilepsy (Nagai et al, 2004d), requiring 30 minutes of continuous biofeedback. This 30-minute training caused both of our TS study groups, despite being instructed to try to relax, to increase their sympathetic arousal. We speculate that patients with TS have a particularly difficult time sustaining attention for periods as
long as 30 minutes, although they can attend for 5 minutes (Nagai et al, 2009a). We also speculate that the unavoidable occurrence of tics during our training sessions evoked increases in sympathetic arousal, thus hindering performance for both physiological and psychological reasons. Whatever the reasons, our patients with TS had an overall increase in sympathetic arousal; over the 30-minute sessions, they fluctuated between periods of successful relaxation and phases of rapid arousal. Given the lack of previous studies or standardized biofeedback protocols for patients with TS, we did not foresee our patients’ difficulty. Electrodermal activity is typically a sensitive autonomic measure of movement; our treatment protocol might have needed to be modified for people with tics. Nevertheless, tasks requiring sustained attention (including driving) typically suppress the expression of tics, and our pilot study (Nagai et al, 2009a) had suggested that biofeedback might have similar effects.
Yale-Brown Obsessive Compulsive Scale
Beck Depression Inventory
Tourette Syndrome Quality of Life Scale
15 14
18
13
Score
Score
11 10 9 8
* Score
17
12
16 15 14
7
13
6 Before
After
Before
61 60 59 58 57 56 55 54 53 52 51 50
After
*
Before
After
FIGURE 4. Changes in combined-group psychological questionnaire scores before and after electrodermal biofeedback treatment. Beck Depression Inventory scores tended to improve (t = 1.95, P = 0.067). Yale-Brown Obsessive Compulsive Scale scores (t = 2.427, P = 0.034) and Tourette Syndrome Quality of Life Scale scores (t = 2.45, P = 0.024) improved significantly.
22 | www.cogbehavneurol.com
r
2014 Lippincott Williams & Wilkins
Cogn Behav Neurol
Volume 27, Number 1, March 2014
Despite our current patients’ difficulty in suppressing their tics with biofeedback, both of our study groups had an overall reduction in tic frequency and improved questionnaire indices of their psychological state. Most clinical interventions produce placebo effects, with the size typically reaching as high as 35% (Finniss et al, 2010). Our current study had a considerably higher placebo effect of 57%. A previous randomized controlled trial of deep brain stimulation in patients with TS also showed an unusually high placebo effect, which actually grew from 37% during the intervention to 49% at 1-year follow-up (Ackermans et al, 2011). The tic activity of patients with TS reflects high levels of suggestibility (Leckman et al, 2006), as shown by the higher rate of adverse events reported in the placebo group than in the treatment group during a double-blind randomized controlled trial of topiramate for TS (Jankovic et al, 2010). As for our current study, we suspect that both the experience of biofeedback and the frequent treatment sessions may have increased the placebo effects in our control patients, who would have been especially vulnerable to suggestion. Across patients in both of our study groups, we did not find a significant correlation between the reduction in tics and the improved mood and quality-of-life scores, suggesting that these effects are independent or dissociable. Still, the reduction in tics and the improved mood are consistent with a placebo effect of the intervention. Also consistent with a placebo effect is the regular contact with the therapist who ran the biofeedback sessions. The high placebo effect in this study needs to be contrasted with the lack of observed placebo effects with sham treatment in our study of patients with epilepsy, using a similar approach of electrodermal biofeedback (Nagai et al, 2004d). In that study, all patients were able to follow the biofeedback procedure, learning the desired direction of electrodermal alteration without problems, and we found significantly better seizure reduction in the active-biofeedback than in the sham-control group. Moreover, the reduction in seizures correlated linearly with performance on the biofeedback task: The better the biofeedback task performance, the better the clinical outcome. In the current study, although our TS groups may have had some periods of tic attenuation during effective relaxation during the 30-minute biofeedback task, we found no effect beyond placebo. Interestingly, we did observe a reduction in vocal tics, related to an improved Yale-Brown Obsessive Compulsive Scale score. The biological interpretation of this finding is not clear, but a common substrate may exist within the orbitofrontal cortical region, where activity has been associated with expression or control of tics in TS (George et al, 1992), rumination in obsessive-compulsive disorder and depression (eg, Rauch et al, 1994), and control of autonomic activity (van Dijk et al, 1992). In conditions such as depression, a predominance of sympathetic tone is further associated with rumination (Ottaviani et al, 2009). r
2014 Lippincott Williams & Wilkins
Biofeedback Treatment for Tourette Syndrome
Because only a few patients in our study had vocal tics, our finding deserves further investigation in more patients. It is notable that our patients’ reduction in vocal tics was associated with a higher skin conductance level during the first 3 sessions. However, we had already observed this relationship between sympathetic tone and expression of vocal tics in our pilot study (Nagai et al, 2009a). In conclusion, our randomized controlled trial did not demonstrate a specific biofeedback effect in reducing tic frequency in patients with TS. After a month of intervention, however, patients in both the treatment and control groups had fewer tics and a better psychological state. These nonspecific benefits of a clinical intervention in TS would, of course, be difficult to investigate further, but they stand in contrast to a lack of such benefits in patients with other disorders. Our negative findings were unexpected given the robust effects of our study of short-term (5-minute) biofeedback relaxation in patients with TS (Nagai et al, 2009a). The small number of patients in our current study also poses an issue in generalizing the results. There is still value in investigating the clinical effects of an electrodermal biofeedback intervention using a different treatment design. Ideally, we would provide biofeedback training in much shorter blocks, perhaps several 5-minute periods during each therapy session, to help patients learn and benefit from biofeedback. This remains a goal. ACKNOWLEDGMENTS The authors thank Jack Hawksley and Anna Thake for their assistance during data collection. REFERENCES Ackermans L, Duits A, van der Linden C, et al. 2011. Double-blind clinical trial of thalamic stimulation in patients with Tourette syndrome. Brain. 134:832–844. American Psychiatric Association. 2000. Diagnostic and Statistical Manual of Mental Disorders. 4th edition, Text Revision. Washington, DC: American Psychiatric Association. Beck AT, Ward CH, Mendelson M, et al. 1961. An inventory for measuring depression. Arch Gen Psychiatry. 4:561–571. Cavanna AE, Schrag A, Morley D, et al. 2008. The Gilles de la Tourette syndrome-quality of life scale (GTS-QOL): development and validation. Neurology. 71:1410–1416. Diamond A, Kenney C, Jankovic J. 2006. Effect of vagal nerve stimulation in a case of Tourette’s syndrome and complex partial epilepsy. Mov Disord. 21:1273–1275. Finniss DG, Kaptchuk TJ, Miller F, et al. 2010. Biological, clinical, and ethical advances of placebo effects. Lancet. 375:686–695. George MS, Trimble MR, Costa DC, et al. 1992. Elevated frontal cerebral blood flow in Gilles de la Tourette syndrome: a 99TcmHMPAO SPECT study. Psychiatry Res. 45:143–151. Goetz CG, Shannon KM, Carroll VS, et al. 1987. The autonomic nervous system in Gilles de la Tourette’s syndrome. Mov Disord. 2:99–102. Goodman WK, Price LH, Rasmussen SA, et al. 1989. The Yale-Brown Obsessive Compulsive Scale. I. Development, use, and reliability. Arch Gen Psychiatry. 46:1006–1011. Holmes TH, Rahe RH. 1967. The social readjustment rating scale. J Psychosom Res. 11:213–218. Jankovic J, Jimenez-Shahed J, Brown LW. 2010. A randomised, doubleblind, placebo-controlled study of topiramate in the treatment of Tourette syndrome. J Neurol Neurosurg Psychiatry. 81:70–73.
www.cogbehavneurol.com |
23
Nagai et al
Leckman JF, Block MH, King RA, et al. 2006. Phenomenology of tics and natural history of tic disorders. Adv Neurol. 99:1–16. Leckman JF, Riddle MA, Hardin MT, et al. 1989. The Yale Global Tic Severity Scale: initial testing of a clinician-rated scale of tic severity. J Am Acad Child Adolesc Psychiatry. 28:566–573. Nagai Y. 2011. Biofeedback and epilepsy. Curr Neurol Neurosci Rep. 11:433–450. Nagai Y, Cavanna A, Critchley HD. 2009a. Influence of sympathetic autonomic arousal on tics: implications for a therapeutic behavioral intervention for Tourette syndrome. J Psychosom Res. 67:599–605. Nagai Y, Critchley HD, Featherstone E, et al. 2004a. Activity in ventromedial prefrontal cortex covaries with sympathetic skin conductance level: a physiological account of a “default mode” of brain function. Neuroimage. 22:243–251. Nagai Y, Critchley HD, Featherstone E, et al. 2004b. Brain activity relating to the Contingent Negative Variation (CNV): fMRI investigation. Neuroimage. 21:1232–1241. Nagai Y, Critchley HD, Rothwell JC, et al. 2009b. Changes in cortical potential associated with modulation of peripheral sympathetic activity in patients with epilepsy. Psychosom Med. 71:84–92. Nagai Y, Goldstein LH, Critchley HD, et al. 2004c. Influence of sympathetic autonomic arousal on contingent negative variation: implications for a therapeutic behavioural intervention in epilepsy. Epilepsy Res. 58:185–193. Nagai Y, Goldstein LH, Fenwick PB, et al. 2004d. Clinical efficacy of biofeedback treatment on reducing seizures in adult epilepsy: a preliminary randomized controlled study. Epilepsy Behav. 5: 216–223. O’Connor K, Lavoie ME, Robert M. 2001. Preparation and motor potentials in chronic tic and Tourette syndromes. Brain Cogn. 46:224–226. Ottaviani C, Shapiro D, Davydov DM, et al. 2009. The autonomic phenotype of rumination. Int J Psychophysiol. 72:267–275. Piacentini J, Chang S. 2001. Behavioral treatments for Tourette syndrome and tic disorders: state of the art. Adv Neurol. 85: 319–331.
24 | www.cogbehavneurol.com
Cogn Behav Neurol
Volume 27, Number 1, March 2014
Piacentini J, Woods DW, Scahill L, et al. 2010. Behavior therapy for children with Tourette disorder: a randomized controlled trial. JAMA. 303:1929–1937. Rauch SL, Jenike MA, Alpert NM, et al. 1994. Regional cerebral blood flow measured during symptom provocation in obsessive-compulsive disorder using oxygen 15-labeled carbon dioxide and positron emission tomography. Arch Gen Psychiatry. 51:62–70. Robertson MM. 2000. Tourette Syndrome, associated conditions and the complexities of treatment. Brain. 123:425–462. Robertson MM. 2008. The prevalence and epidemiology of Gilles de la Tourette syndrome: Part 1: The epidemiological and prevalence studies. J Psychosom Res. 65:461–472. Robertson MM. 2012. The psychosocial aspects of Tourette Syndrome: a family guide and perspective. In: Walkup JT, Mink JW, KStP McNaught, eds. A Family’s Guide to Tourette Syndrome. Bayside, NY: Tourette Syndrome Association; 171–192. Spielberger CD. 1983. Manual for the State-Trait Anxiety Inventory. Palo Alto, CA: Consulting Psychologists Press. Sterman MB, Friar L. 1972. Suppression of seizures in an epileptic following sensorimotor EEG feedback training. Electroencephalogr Clin Neurophysiol. 33:89–95. Tansey MA. 1986. A simple and a complex tic (Gilles de la Tourette’s syndrome): their response to EEG sensorimotor rhythm biofeedback training. Int J Psychophysiol. 4:91–97. van Dijk JG, Koenderink M, Kramer CG, et al. 1992. Non-invasive assessment of autonomic nervous function in Gilles de la Tourette syndrome. Clin Neurol Neurosurg. 94:157–159. van Woerkom TC, Fortgens C, van de Wetering BJ, et al. 1988. Contingent negative variation in adults with Gilles de la Tourette syndrome. J Neurol Neurosurg Psychiatry. 51:630–634. Weate SJ, Newell SA, Bogner JE, et al. 1993. Contingent negative variation in Gilles de la Tourette syndrome. Clin Electroencephalogr. 24:188–191. Worbe Y, Gerardin E, Hartmann A, et al. 2010. Distinct structural changes underpin clinical phenotypes in patients with Gilles de la Tourette syndrome. Brain. 133:3649–3660.
r
2014 Lippincott Williams & Wilkins
Biofeedback Treatment for Tourette Syndrome: A Preliminary Randomized Controlled Trial Yoko Nagai, PhD,* Andrea E. Cavanna, MD, PhD,w Hugo D. Critchley, DPhil, FRCPsyc,*z Jeremy J. Stern, FRCP,y Mary M. Robertson, MD, FRCPsyc,y and Eileen M. Joyce, MD, FRCPsyc8
Objective: To study the clinical effectiveness of biofeedback treatment in reducing tics in patients with Tourette syndrome. Background: Despite advances in the pharmacologic treatment of patients with Tourette syndrome, many remain troubled by their tics, which may be resistant to multiple medications at tolerable doses. Electrodermal biofeedback is a noninvasive biobehavioral intervention that can be useful in managing neuropsychiatric and neurologic conditions. Methods: We conducted a randomized controlled trial of electrodermal biofeedback training in 21 patients with Tourette syndrome. Results: After training the patients for 3 sessions a week over 4 weeks, we observed a significant reduction in tic frequency and improved indices of subjective well-being in both the active-biofeedback and sham-feedback (control) groups, but there was no difference between the groups in these measurements. Furthermore, the active-treatment group did not demonstrably learn to reduce their sympathetic electrodermal tone using biofeedback. Conclusions: Our findings indicate that this form of biofeedback training was unable to produce a clinical effect greater than placebo. The main confounding factor appeared to be the 30-minute duration of the training sessions, which made it difficult for patients to sustain a reduction in sympathetic tone when their tics themselves were generating competing phasic electrodermal arousal responses. Despite a negative finding in this study, electrodermal biofeedback training may have a role in managing tics if optimal training schedules can be identified.
Received for publication January 16, 2013; accepted April 15, 2013. From the *Brighton and Sussex Medical School, University of Sussex, Falmer Campus, Brighton, UK; wDepartment of Neuropsychiatry, Birmingham and Solihull Mental Health National Health Service Foundation Trust and University of Birmingham, Birmingham, UK; zSackler Centre for Consciousness Science, University of Sussex, Brighton, UK; ySt George’s Hospital and Medical School, London, UK; and 8Sobell Department of Motor Neuroscience, University College London Institute of Neurology, London, UK. Supported in part by the Tourette Syndrome Association. The authors declare no conflicts of interest. Reprints: Yoko Nagai, PhD, Brighton and Sussex Medical School, University of Sussex, Falmer Campus, Brighton BN1 9RR, UK (e-mail: [email protected]). Copyright r 2014 by Lippincott Williams & Wilkins
Cogn Behav Neurol
Volume 27, Number 1, March 2014
Key Words: biofeedback, electrodermal activity, sympathetic autonomic arousal, tics, Tourette syndrome (Cogn Behav Neurol 2014;27:17–24)
Reader Benefit: Electrodermal activity biofeedback may play a role in reducing tics in patients with Tourette syndrome if optimal training schedules can be determined. TS = Tourette syndrome.
T
ourette syndrome (TS) is a neurodevelopmental disorder affecting between 0.4% and 3.8% of schoolchildren worldwide, and with a calculated prevalence of 1% of the general population (Robertson, 2008). Patients with TS suffer from multiple involuntary movements (motor tics) and 1 or more phonic or vocal tics that last longer than a year. These tics can cause patients social discomfort that may affect their academic or work performance. Many patients perceive such secondary effects of TS as more psychologically problematic than the tics themselves (Robertson, 2000, 2012). Medications are typically the first-line treatment for managing tics, with neuroleptics most commonly prescribed. Because patients commonly experience side effects from these drugs (Robertson, 2000), nonpharmacologic approaches have long been sought. Behavioral interventions such as habit reversal training are helpful as alternative or adjunctive treatments (Piacentini and Chang, 2001). Habit reversal training encompasses self-monitoring, relaxation, competing response practice, and contingency management training to reinforce behavior with reward and punishment. Piacentini et al (2010) reported significantly lower tic severity in children with TS who had received the Comprehensive Behavioral Intervention for Tics than in a control group. This therapy consists of habit reversal training, relaxation training, and identification of vulnerable situations. Although habit reversal training is the primary component, all of these elements have been used to manage tics. Other established behavioral approaches include cognitive therapy, assertiveness training, and self-monitoring (Robertson, 2000). Although studies examining www.cogbehavneurol.com |
17
Nagai et al
the efficacy of behavioral treatments are mostly encouraging, these approaches are indirect and do not clearly target the pathogenetic mechanisms underlying TS. Biofeedback encompasses a set of behavioral or psychophysiological treatments that have been gaining popularity in the management of neurologic, psychological, and psychiatric conditions over the last few decades. Biofeedback is used for conditions such as anxiety disorders, epilepsy, asthma, some cardiovascular disorders, hypertension, irritable bowel syndrome, neuromuscular disorders, and migraine and tension headaches. Biofeedback characteristically enables a patient to gain voluntary control over covert physiological responses by making these responses explicit through real-time visual or auditory feedback. Patients are typically able to learn how to modify these physiological processes volitionally. Such change is thought to be achieved through an indirect combination of psychological (eg, changes in attentional focus) and physiological (eg, changes in muscle tension or in rate and depth of breathing) mechanisms, although patients do not necessarily consciously recognize why they are doing better. The earliest attempts at biofeedback treatments for TS, made in the 1980s, used electroencephalography. The target signal was enhancement of the sensory motor rhythm, an electroencephalographic frequency component of 12 to 14 Hz that emerges over the rolandic sensorimotor area when a person is still (Tansey, 1986). Sensory motor rhythm biofeedback was first explored to reduce seizure frequency in epilepsy (Sterman and Friar, 1972), and was later reported as successful in 2 patients with TS who showed a dramatic reduction in tic frequency (Tansey, 1986). The biofeedback sessions for the patients with TS lasted 30 minutes once a week; however, the overall length of the treatment seemed individually tailored. One of the patients with TS benefited after 5 weeks of treatment, and the other after 18 weeks. Still, given the lack of quantitative outcome measures in Tansey’s 2 patients, nonspecific psychological effects of treatment cannot be excluded as explanatory factors in their improvement. In a pilot study of biofeedback in TS (Nagai et al, 2009a), we tested an easy-to-implement approach targeting peripheral sympathetic arousal indexed by electrodermal activity, which reflects sympathetic nervous control of the activity of eccrine sweat glands. In an earlier neuroimaging study (Nagai et al, 2004a), we had found that reduction of electrodermal sympathetic tone was associated with greater activity within the orbitofrontal and ventromedial prefrontal cortex, regions also associated with affective responses and default (disengaged) brain states. We have also shown that electrodermal sympathetic activity is inversely related to an index of cortical arousal, slow cortical potentials, evoked experimentally using a contingent negative variation paradigm (Nagai et al, 2004c). We found brain correlates of induced cortical arousal in heightened thalamocortical activity that was related to the amplitude of slow cortical potentials (Nagai et al, 2004b). Our observations supported a proposal that
18 | www.cogbehavneurol.com
Cogn Behav Neurol
Volume 27, Number 1, March 2014
peripheral electrodermal modulation can influence thalamocortical regulation (Nagai et al, 2004b), with potential therapeutic implications for neurologic conditions, notably epilepsy (Nagai et al, 2004d; Nagai, 2011). Patients with TS have been reported to have abnormalities of the contingent negative variation: an attenuation of the early part of the wave but an enhancement of the later (dopaminergically regulated) component (van Woerkom et al, 1988; Weate et al, 1993). This pattern suggests dysregulation of thalamocortical excitability in patients with TS. This perspective is supported by the generally positive outcomes of thalamic deep brain stimulation as a treatment for TS (Ackermans et al, 2011). In our proof-of-principle TS study (Nagai et al, 2009a), which lacked an explicit control condition, we tested the short-term effects of using electrodermal biofeedback to raise and lower sympathetic activity on the expression of tics in 15 patients. We found that a reduction in sympathetic tone, elicited using electrodermal biofeedback for 5 minutes, was associated with a reduction in tic frequency (Figure 1). Before this pilot study, it had not been clear whether the greater attentional focus associated with increasing peripheral sympathetic tone would also be beneficial in TS, as it is in epilepsy (Nagai et al, 2004d). This proved not to be the case. Another uncertain factor had been the best training regimen for TS. Our epilepsy treatment study (Nagai et al, 2004d) had established an optimal protocol for electrodermal biofeedback in adults with epilepsy as 30-minute sessions, 3 times a week. On this protocol, the patients showed treatment effects within 4 weeks. Although the role of central autonomic control on tic activity has not been comprehensively investigated, patients with TS generally have the same heart rate and blood pressure as those in healthy people (Goetz et al, 1987; van Dijk et al, 1992). Drugs that act on sympathetic tone can have therapeutic benefit. Clonidine, an alpha 2 receptor agonist, is often 1 of the first-line drugs given for TS, and cholinergic (typically nicotinic) agents are also reported to reduce symptoms (Robertson, 2000). There is a single case report that vagus nerve stimulation reduced motor and phonic tic activity in a patient with TS (Diamond et al, 2006). Thus, there is evidence that both sympathetic and parasympathetic factors influence tic expression, and our pilot TS study (Nagai et al, 2009a) suggested that acute attenuation of sympathetic activity promptly reduced tic activity. In the current study, we conducted a randomized controlled trial to examine the possible translation of electrodermal biofeedback to a treatment for patients with TS. Although the study was not strictly double blind, the researcher who counted patients’ tics was blinded to the treatment condition.
METHODS Participants We recruited 21 patients (26.7 ± 10.6 years old) with TS from the National Hospital for Neurology and r
2014 Lippincott Williams & Wilkins
Cogn Behav Neurol
Volume 27, Number 1, March 2014
Biofeedback Treatment for Tourette Syndrome
Arousal Biofeedback Relaxation Biofeedback
Number of Tics
50
*
40
*
30 20 10
*
0
Relaxation Biofeedback Motor
Face
Arousal Biofeedback Body Type of Tics
Vocal
Total
FIGURE 1. Average number of tics during brief 5-minute periods of relaxation and arousal biofeedback in 15 patients with Tourette syndrome in the authors’ proof-of-principle study (Nagai et al, 2009a). Motor tics include both face and body tics. Total tics include face, body, and vocal tics. Relaxation biofeedback significantly reduced the rates of tics over arousal biofeedback in the scores for motor, face, and total tics. The asterisks denote significant differences. Reproduced from Nagai Y, Cavanna A, Critchley HD. 2009. Influence of sympathetic autonomic arousal on tics: implications for a therapeutic behavioral intervention for Tourette syndrome. J Psychosom Res. 67:599–605, with permission from Elsevier.
Neurosurgery and from the Department of Neurology at St George’s Hospital, both in London. All patients satisfied the Diagnostic and Statistical Manual of Mental Disorders, 4th edition, Text Revision (American Psychiatric Association, 2000) criteria for the diagnosis of TS. We excluded patients with a history of psychotic illness, personality disorder, or major medical or neurologic diseases, including movement disorders such as Parkinson disease, Wilson disease, Huntington disease, and dystonia. We also required patients not to have had any changes in their medication regimen for at least 1 month before the study began, and no change likely during the 4-week study. The 21 patients were randomly assigned to either the active-biofeedback (n = 12) or sham-control (n = 9) group, according to the randomization table. Six patients in the biofeedback group and 2 in the control group were taking neuroleptics (haloperidol, risperidone, pimozide, sulpiride, or aripiprazole), and 1 patient in each group was taking clonidine. Two patients in the biofeedback group and 1 in the control group were taking a selective serotonin reuptake inhibitor (fluvoxamine or fluoxetine). One patient in the biofeedback group was taking lamotrigine. The study was approved by the local Ethics Committee, and all patients provided informed written consent to the procedure.
Procedures Biofeedback Training Protocol During the 4-week treatment, the patients attended a 30-minute biofeedback session 3 times a week, for a total of 12 sessions (Table 1). During each session, the therapist instructed the patients to attend to, and change, the animation on a computer screen by relaxing both mentally and physically. The patients assigned to the active-treatment group received true biofeedback from the r
2014 Lippincott Williams & Wilkins
computer screen, and the patients assigned to the control group received sham biofeedback. Electrodermal Recording and Biofeedback To measure electrodermal activity and deliver biofeedback, we used a purpose-built biofeedback system (Inner Tuner: Professional, Ultrasis plc, London, UK) modified for experimental use. We placed 2 dry nickelplated electrodes on the palm of the patient’s left hand. The area of each electrode was approximately 1 cm2. The sampling rate of the recording was 10 Hz. Biofeedback took the form of computer-generated graphics displayed on a computer monitor in front of the patient. When patients in the active-biofeedback group relaxed and their skin conductance dropped, an animated figure in the foreground of the computer screen moved rightward through a successive sequence of animations. The sequence started with a picture of fish, which turned into a mermaid, then a walking lady, and then a star. When patients’ skin conductance increased, the display moved leftward, returning to an earlier part of the animation sequence. Thus, the active-biofeedback condition “rewarded” patients for reducing their electrodermal sympathetic activity by showing them more of the animation sequence. The patients in the sham-control group viewed the same computer-generated graphics as those used in the
TABLE 1. Design of This Study Before Treatment Treatment Assessment 1. Video monitoring of tics 2. Questionnaires
Electrodermal biofeedback treatment: 30-minute sessions 3 times/week for 4 weeks
After Treatment Assessment 1. Video monitoring of tics 2. Questionnaires
www.cogbehavneurol.com |
19
Nagai et al
active-biofeedback condition, and also tried to reduce their electrodermal activity level using the evolving sequence of animations. However, in contrast to the activebiofeedback group, here the video animation changed direction unrelated to the ups and downs in the patients’ own electrodermal activity. Thus, the sham-control patients could not learn direct biofeedback control. To avoid frustrating the patients, we altered the progression of their animation changes across sessions to match the sequences that the patients in the active-biofeedback group were generating by their changes in electrodermal activity. Because the sham-control patients believed that they were actively performing the task, they did not report negative perceptions of having failed to learn to relax. Video Recording and Psychological Assessment To assess the effects of treatment on our patients’ tic severity, we started the first and last biofeedback sessions by counting tics and then taking a psychological profile. We recorded each patient’s tics with a digital video camera (Sony, DVD, Sony Europe). The patients sat on a comfortable chair and relaxed for 10 minutes while the camera recorded their whole-body movements. Author A.E.C., a neurologist with special expertise in TS and blinded to the treatment condition, viewed the video recording and counted the tics. After the video recording, we gave the Yale Global Tic Severity Scale (Leckman et al, 1989) to quantify the patients’ self-reported symptoms of TS. To assess comorbid symptoms of obsessive-compulsive disorder, depression, and anxiety, we gave the Yale-Brown Obsessive Compulsive Scale (Goodman et al, 1989), Beck Depression Inventory (Beck et al, 1961), and State-Trait Anxiety Inventory (Spielberger, 1983). We gave the Gilles de la Tourette Syndrome Quality of Life Scale (Cavanna et al, 2008) to assess self-reported changes in patients’ healthrelated quality of life. We gave the Social Readjustment Rating Scale (Holmes and Rahe, 1967), commonly called the Holmes and Rahe Stress Scale, to monitor the stress of external life events that patients experienced during the study.
Data Analysis We calculated the change in each patient’s electrodermal activity during each biofeedback session and then averaged those changes over the 12 sessions. We used paired t tests to investigate significant electrodermal changes between the beginning and end of each session in both study groups. To measure the progress in performance of biofeedback, we calculated the average electrodermal activity changes over the first 3 and last 3 sessions. This permitted us to determine if patients improved in the degree to which they could change their skin resistance level over the month of active- or sham-biofeedback treatment. We used independent t tests to examine the differences in electrodermal activity between the active-biofeedback and sham-control groups.
20 | www.cogbehavneurol.com
Cogn Behav Neurol
Volume 27, Number 1, March 2014
We calculated the number of tics before and after treatment. The main outcome measure was the change in number of tics from the baseline period. We categorized the types of tics as motor (body or facial) and vocal. Total tics represented the sum of the motor and vocal tics. We used repeated measures analyses of variance to analyze the mean number of tics before and after treatment in each group. We corrected degrees of freedom using Greenhouse-Geisser estimates. We further explored all significant main effects and interactions using post hoc t tests. We performed paired t tests to investigate changes in the psychological questionnaires before and after the intervention in both study groups. We analyzed differences between these conditions using independent t tests. We also performed Pearson correlation analyses of the psychological and physiological data. We set significance at P < 0.05.
RESULTS Statistical analysis revealed that there were no group differences in the patients’ age, initial psychological state, or tonic electrodermal activity level. However, the number of tics (mean ± standard deviation) at baseline at the first session was significantly higher in the active-biofeedback group (143.17 ± 97.55) than in the control group (43.00 ± 33.52). Nine patients (6 in the biofeedback group and 3 controls) dropped out in the middle of the treatment, citing lack of motivation. To avoid the effects of dropout, we used intention-to-treat data analysis.
Changes in Electrodermal Activity We compared the tonic level of electrodermal activity at the beginning and end of each 30-minute biofeedback session. All of the patients in the activebiofeedback group significantly changed their tonic electrodermal activity level (t = 3.43, P = 0.006) (Figure 2). However, despite the instruction to relax their physiological state, the patients in the active-biofeedback group actually increased their sympathetic arousal, as indexed by their skin response. Their mean skin resistance at the beginning of biofeedback sessions was 1035.24 ± 339.44 kO, and at the end it had fallen to 735.15 ± 422.29 kO. The control group showed no significant difference in skin response between the beginning and the end of each session. Their mean skin resistance at the beginning of biofeedback sessions was 921.47 ± 571.09 kO, and, at the end, 922.60 ± 933.72 kO. The difference in electrodermal activity change between the active-biofeedback and sham-control groups did not reach significance. Over the course of training, the patients in both groups were unable to improve their biofeedback skills (t = 1.97, P = 0.064).
Changes in Tic Activity Figure 3 shows the number of all types of tics before and after the intervention in both study groups. Despite both groups’ inability to change their electrodermal r
2014 Lippincott Williams & Wilkins
Cogn Behav Neurol
3500 3000
Volume 27, Number 1, March 2014
Correlations Among Electrodermal Activity, Tics, and Questionnaires
Biofeedback Series21 Control
Kohm
2500 2000 1500 1000 500 0 Beginning
End
FIGURE 2. Change in electrodermal activity (galvanic skin response) at the beginning and end of each 30-minute session in all 21 patients in this study, averaged over the 12 biofeedback sessions. Despite being instructed to increase their skin resistance level, most patients in both groups reduced their skin resistance. (In the biofeedback group, the patients with the second- and third-lowest beginning levels had essentially overlapping traces. For the patients who dropped out before the end of the study, we averaged the results of their completed sessions.)
activity, we found a significant main effect of Intervention Time in reducing the number of tics (F1,19 = 4.68, P = 0.04). However, there was no significant interaction with group, suggesting no difference in the effects of biofeedback and sham control. The mean total number of tics in the biofeedback group was 143.17 ± 97.55 before and 110.25 ± 77.69 after intervention; in the control group, the mean was 43.00 ± 33.52 before and 21.22 ± 19.65 after. The number of motor tics in the combined groups was significantly lower (t = 2.21, P = 0.039), especially body tics (t = 2.20, P = 0.039). Facial and vocal tics were unchanged.
Evaluation of the Questionnaires Independent sample t tests revealed no difference in psychological and clinical state between the 2 groups at the beginning of the study; neither was there any significant difference between the groups in the changes in pre- and post-treatment questionnaire scores. Both groups had a tendency for Beck Depression Inventory scores to improve (t = 1.95, P = 0.067) (Figure 4). Both groups showed a significant reduction in Yale-Brown Obsessive Compulsive Scale scores (t = 2.427, P = 0.034) and a significantly improved Tourette Syndrome Quality of Life Scale score (t = 2.45, P = 0.024). There was no significant difference in the Holmes and Rahe Stress Scale, suggesting that the patients’ external circumstances had not changed during the study period. r
2014 Lippincott Williams & Wilkins
Biofeedback Treatment for Tourette Syndrome
For the most part, we found no correlation between changes in patients’ tics and psychological questionnaires, suggesting that the overall reduction of tics could not be accounted for by psychological change. One exception was the observed correlation between a reduction of vocal tics and improvement of obsessive-compulsive symptoms (P = 0.004, Pearson correlation coefficient [r] = 0.60). Changes in vocal tics were also related to performance on the biofeedback task over the first 3 sessions. Overall, treatment was unsuccessful. Patients who increased their sympathetic activity despite their intention to relax had the greatest reduction in frequency of vocal tics (P = 0.016, r = 0.57). These patients had also had more vocal tics at the beginning of treatment (P = 0.023, r = 0.49).
DISCUSSION Understanding of the central neural mechanisms of TS has improved through advances in neuroimaging studies and the development of interventions such as deep brain stimulation. However, the relationships among attentional control, emotional stress, autonomic arousal, and the expression of tics have yet to be characterized comprehensively. In our proof-of-principle study of TS (Nagai et al, 2009a), we showed that effective biofeedback modulation of electrodermal activity to decrease sympathetic tone could diminish tic activity, while biofeedback to increase sympathetic arousal could not. We hypothesized that biofeedback might reduce tics mechanistically through psychophysiological changes generated by the sympathetic modulation. Electrodermal biofeedback alters both activity within the medial orbitofrontal cortex and amplitude of slow cortical potentials, indexing cortical excitability (Nagai et al, 2004a, 2009b). TS has been linked in part to abnormal function of the orbitofrontal cortex (George et al, 1992; Worbe et al, 2010), and most patients with TS have low preparatory motor potentials (O’Connor et al, 2001; van Woerkom et al, 1988). Thus, by affecting cortical excitability, electrodermal biofeedback may influence neural networks associated with the generation of motor tics. Our pilot study (Nagai et al, 2009a) led to the current study aimed at determining whether training patients with TS to decrease their sympathetic activity volitionally using electrodermal biofeedback could reduce their tics. Although our patients succeeded in modulating their sympathetic activity briefly using biofeedback, we were unable to find evidence that our methods enabled the patients to learn to control their sympathetic arousal or to benefit with a reduction in tics. This outcome might reflect the relatively high dropout rate in our active-biofeedback group. Our successful outcome study using this method in patients with epilepsy did not have dropouts (Nagai et al, 2004d). Thus, effective biofeedback performance and accompanying clinical improvement appear to be crucial factors in www.cogbehavneurol.com |
21
Cogn Behav Neurol
Nagai et al
Both Bothconditions condition
Control
Volume 27, Number 1, March 2014
Biofeedback
160 *
Number of Tics
140
*
120 100 80
*
60 40 20 0 Before After
Before After
Motor
Before After
Face
Before After
Before After
Body Vocal
Total
Type of Tics
FIGURE 3. Averaged number of tics at the beginning and end of the 12 galvanic skin response biofeedback treatment sessions. The asterisks denote significant differences.
ensuring that people will complete biofeedback treatment. We suspect that frustration with poor performance might have been even higher in our active-biofeedback group than in our controls because the display truly mirrored the patients’ physiological state. One critical difference between our pilot TS work (Nagai et al, 2009a) and this study was the duration of biofeedback sessions. In the pilot study, patients guided by electrodermal biofeedback significantly raised and lowered their sympathetic tone over periods of 5 minutes. However, for the current study we adapted the protocol that we had used successfully in biofeedback training for epilepsy (Nagai et al, 2004d), requiring 30 minutes of continuous biofeedback. This 30-minute training caused both of our TS study groups, despite being instructed to try to relax, to increase their sympathetic arousal. We speculate that patients with TS have a particularly difficult time sustaining attention for periods as
long as 30 minutes, although they can attend for 5 minutes (Nagai et al, 2009a). We also speculate that the unavoidable occurrence of tics during our training sessions evoked increases in sympathetic arousal, thus hindering performance for both physiological and psychological reasons. Whatever the reasons, our patients with TS had an overall increase in sympathetic arousal; over the 30-minute sessions, they fluctuated between periods of successful relaxation and phases of rapid arousal. Given the lack of previous studies or standardized biofeedback protocols for patients with TS, we did not foresee our patients’ difficulty. Electrodermal activity is typically a sensitive autonomic measure of movement; our treatment protocol might have needed to be modified for people with tics. Nevertheless, tasks requiring sustained attention (including driving) typically suppress the expression of tics, and our pilot study (Nagai et al, 2009a) had suggested that biofeedback might have similar effects.
Yale-Brown Obsessive Compulsive Scale
Beck Depression Inventory
Tourette Syndrome Quality of Life Scale
15 14
18
13
Score
Score
11 10 9 8
* Score
17
12
16 15 14
7
13
6 Before
After
Before
61 60 59 58 57 56 55 54 53 52 51 50
After
*
Before
After
FIGURE 4. Changes in combined-group psychological questionnaire scores before and after electrodermal biofeedback treatment. Beck Depression Inventory scores tended to improve (t = 1.95, P = 0.067). Yale-Brown Obsessive Compulsive Scale scores (t = 2.427, P = 0.034) and Tourette Syndrome Quality of Life Scale scores (t = 2.45, P = 0.024) improved significantly.
22 | www.cogbehavneurol.com
r
2014 Lippincott Williams & Wilkins
Cogn Behav Neurol
Volume 27, Number 1, March 2014
Despite our current patients’ difficulty in suppressing their tics with biofeedback, both of our study groups had an overall reduction in tic frequency and improved questionnaire indices of their psychological state. Most clinical interventions produce placebo effects, with the size typically reaching as high as 35% (Finniss et al, 2010). Our current study had a considerably higher placebo effect of 57%. A previous randomized controlled trial of deep brain stimulation in patients with TS also showed an unusually high placebo effect, which actually grew from 37% during the intervention to 49% at 1-year follow-up (Ackermans et al, 2011). The tic activity of patients with TS reflects high levels of suggestibility (Leckman et al, 2006), as shown by the higher rate of adverse events reported in the placebo group than in the treatment group during a double-blind randomized controlled trial of topiramate for TS (Jankovic et al, 2010). As for our current study, we suspect that both the experience of biofeedback and the frequent treatment sessions may have increased the placebo effects in our control patients, who would have been especially vulnerable to suggestion. Across patients in both of our study groups, we did not find a significant correlation between the reduction in tics and the improved mood and quality-of-life scores, suggesting that these effects are independent or dissociable. Still, the reduction in tics and the improved mood are consistent with a placebo effect of the intervention. Also consistent with a placebo effect is the regular contact with the therapist who ran the biofeedback sessions. The high placebo effect in this study needs to be contrasted with the lack of observed placebo effects with sham treatment in our study of patients with epilepsy, using a similar approach of electrodermal biofeedback (Nagai et al, 2004d). In that study, all patients were able to follow the biofeedback procedure, learning the desired direction of electrodermal alteration without problems, and we found significantly better seizure reduction in the active-biofeedback than in the sham-control group. Moreover, the reduction in seizures correlated linearly with performance on the biofeedback task: The better the biofeedback task performance, the better the clinical outcome. In the current study, although our TS groups may have had some periods of tic attenuation during effective relaxation during the 30-minute biofeedback task, we found no effect beyond placebo. Interestingly, we did observe a reduction in vocal tics, related to an improved Yale-Brown Obsessive Compulsive Scale score. The biological interpretation of this finding is not clear, but a common substrate may exist within the orbitofrontal cortical region, where activity has been associated with expression or control of tics in TS (George et al, 1992), rumination in obsessive-compulsive disorder and depression (eg, Rauch et al, 1994), and control of autonomic activity (van Dijk et al, 1992). In conditions such as depression, a predominance of sympathetic tone is further associated with rumination (Ottaviani et al, 2009). r
2014 Lippincott Williams & Wilkins
Biofeedback Treatment for Tourette Syndrome
Because only a few patients in our study had vocal tics, our finding deserves further investigation in more patients. It is notable that our patients’ reduction in vocal tics was associated with a higher skin conductance level during the first 3 sessions. However, we had already observed this relationship between sympathetic tone and expression of vocal tics in our pilot study (Nagai et al, 2009a). In conclusion, our randomized controlled trial did not demonstrate a specific biofeedback effect in reducing tic frequency in patients with TS. After a month of intervention, however, patients in both the treatment and control groups had fewer tics and a better psychological state. These nonspecific benefits of a clinical intervention in TS would, of course, be difficult to investigate further, but they stand in contrast to a lack of such benefits in patients with other disorders. Our negative findings were unexpected given the robust effects of our study of short-term (5-minute) biofeedback relaxation in patients with TS (Nagai et al, 2009a). The small number of patients in our current study also poses an issue in generalizing the results. There is still value in investigating the clinical effects of an electrodermal biofeedback intervention using a different treatment design. Ideally, we would provide biofeedback training in much shorter blocks, perhaps several 5-minute periods during each therapy session, to help patients learn and benefit from biofeedback. This remains a goal. ACKNOWLEDGMENTS The authors thank Jack Hawksley and Anna Thake for their assistance during data collection. REFERENCES Ackermans L, Duits A, van der Linden C, et al. 2011. Double-blind clinical trial of thalamic stimulation in patients with Tourette syndrome. Brain. 134:832–844. American Psychiatric Association. 2000. Diagnostic and Statistical Manual of Mental Disorders. 4th edition, Text Revision. Washington, DC: American Psychiatric Association. Beck AT, Ward CH, Mendelson M, et al. 1961. An inventory for measuring depression. Arch Gen Psychiatry. 4:561–571. Cavanna AE, Schrag A, Morley D, et al. 2008. The Gilles de la Tourette syndrome-quality of life scale (GTS-QOL): development and validation. Neurology. 71:1410–1416. Diamond A, Kenney C, Jankovic J. 2006. Effect of vagal nerve stimulation in a case of Tourette’s syndrome and complex partial epilepsy. Mov Disord. 21:1273–1275. Finniss DG, Kaptchuk TJ, Miller F, et al. 2010. Biological, clinical, and ethical advances of placebo effects. Lancet. 375:686–695. George MS, Trimble MR, Costa DC, et al. 1992. Elevated frontal cerebral blood flow in Gilles de la Tourette syndrome: a 99TcmHMPAO SPECT study. Psychiatry Res. 45:143–151. Goetz CG, Shannon KM, Carroll VS, et al. 1987. The autonomic nervous system in Gilles de la Tourette’s syndrome. Mov Disord. 2:99–102. Goodman WK, Price LH, Rasmussen SA, et al. 1989. The Yale-Brown Obsessive Compulsive Scale. I. Development, use, and reliability. Arch Gen Psychiatry. 46:1006–1011. Holmes TH, Rahe RH. 1967. The social readjustment rating scale. J Psychosom Res. 11:213–218. Jankovic J, Jimenez-Shahed J, Brown LW. 2010. A randomised, doubleblind, placebo-controlled study of topiramate in the treatment of Tourette syndrome. J Neurol Neurosurg Psychiatry. 81:70–73.
www.cogbehavneurol.com |
23
Nagai et al
Leckman JF, Block MH, King RA, et al. 2006. Phenomenology of tics and natural history of tic disorders. Adv Neurol. 99:1–16. Leckman JF, Riddle MA, Hardin MT, et al. 1989. The Yale Global Tic Severity Scale: initial testing of a clinician-rated scale of tic severity. J Am Acad Child Adolesc Psychiatry. 28:566–573. Nagai Y. 2011. Biofeedback and epilepsy. Curr Neurol Neurosci Rep. 11:433–450. Nagai Y, Cavanna A, Critchley HD. 2009a. Influence of sympathetic autonomic arousal on tics: implications for a therapeutic behavioral intervention for Tourette syndrome. J Psychosom Res. 67:599–605. Nagai Y, Critchley HD, Featherstone E, et al. 2004a. Activity in ventromedial prefrontal cortex covaries with sympathetic skin conductance level: a physiological account of a “default mode” of brain function. Neuroimage. 22:243–251. Nagai Y, Critchley HD, Featherstone E, et al. 2004b. Brain activity relating to the Contingent Negative Variation (CNV): fMRI investigation. Neuroimage. 21:1232–1241. Nagai Y, Critchley HD, Rothwell JC, et al. 2009b. Changes in cortical potential associated with modulation of peripheral sympathetic activity in patients with epilepsy. Psychosom Med. 71:84–92. Nagai Y, Goldstein LH, Critchley HD, et al. 2004c. Influence of sympathetic autonomic arousal on contingent negative variation: implications for a therapeutic behavioural intervention in epilepsy. Epilepsy Res. 58:185–193. Nagai Y, Goldstein LH, Fenwick PB, et al. 2004d. Clinical efficacy of biofeedback treatment on reducing seizures in adult epilepsy: a preliminary randomized controlled study. Epilepsy Behav. 5: 216–223. O’Connor K, Lavoie ME, Robert M. 2001. Preparation and motor potentials in chronic tic and Tourette syndromes. Brain Cogn. 46:224–226. Ottaviani C, Shapiro D, Davydov DM, et al. 2009. The autonomic phenotype of rumination. Int J Psychophysiol. 72:267–275. Piacentini J, Chang S. 2001. Behavioral treatments for Tourette syndrome and tic disorders: state of the art. Adv Neurol. 85: 319–331.
24 | www.cogbehavneurol.com
Cogn Behav Neurol
Volume 27, Number 1, March 2014
Piacentini J, Woods DW, Scahill L, et al. 2010. Behavior therapy for children with Tourette disorder: a randomized controlled trial. JAMA. 303:1929–1937. Rauch SL, Jenike MA, Alpert NM, et al. 1994. Regional cerebral blood flow measured during symptom provocation in obsessive-compulsive disorder using oxygen 15-labeled carbon dioxide and positron emission tomography. Arch Gen Psychiatry. 51:62–70. Robertson MM. 2000. Tourette Syndrome, associated conditions and the complexities of treatment. Brain. 123:425–462. Robertson MM. 2008. The prevalence and epidemiology of Gilles de la Tourette syndrome: Part 1: The epidemiological and prevalence studies. J Psychosom Res. 65:461–472. Robertson MM. 2012. The psychosocial aspects of Tourette Syndrome: a family guide and perspective. In: Walkup JT, Mink JW, KStP McNaught, eds. A Family’s Guide to Tourette Syndrome. Bayside, NY: Tourette Syndrome Association; 171–192. Spielberger CD. 1983. Manual for the State-Trait Anxiety Inventory. Palo Alto, CA: Consulting Psychologists Press. Sterman MB, Friar L. 1972. Suppression of seizures in an epileptic following sensorimotor EEG feedback training. Electroencephalogr Clin Neurophysiol. 33:89–95. Tansey MA. 1986. A simple and a complex tic (Gilles de la Tourette’s syndrome): their response to EEG sensorimotor rhythm biofeedback training. Int J Psychophysiol. 4:91–97. van Dijk JG, Koenderink M, Kramer CG, et al. 1992. Non-invasive assessment of autonomic nervous function in Gilles de la Tourette syndrome. Clin Neurol Neurosurg. 94:157–159. van Woerkom TC, Fortgens C, van de Wetering BJ, et al. 1988. Contingent negative variation in adults with Gilles de la Tourette syndrome. J Neurol Neurosurg Psychiatry. 51:630–634. Weate SJ, Newell SA, Bogner JE, et al. 1993. Contingent negative variation in Gilles de la Tourette syndrome. Clin Electroencephalogr. 24:188–191. Worbe Y, Gerardin E, Hartmann A, et al. 2010. Distinct structural changes underpin clinical phenotypes in patients with Gilles de la Tourette syndrome. Brain. 133:3649–3660.
r
2014 Lippincott Williams & Wilkins
