Biofeedback and Self-Regulation, Vol. 3, No. 1, 1978
Biofeedback Modification of Frontal EMG in Normal Subjects Yoshiteru Ohno, Yoshiharu Tanaka, Tsutomu Takeya, and Hideki Matsubara Department of Psychosomatic Medicine Kyushu University, Fukuoka City, Japan Norikazu Kuriya Inatsuki Hospital, Fukuoka Sadao Komemushi Department of PharmaceuticalEngineering Osaka University We carried out a controlled study on the voluntary control o f the frontalis muscle by biofeedback procedures employing 20 normal subjects. Subjects were randomly divided into two groups of 10: (1) the biofeedback group and (2) the control group. Each o f the two groups received five training sessions o f about 40 minutes' duration each on different days. The results obtained are as follows: (1) In the biofeedback group, mean EMG levels decreused progressively and markedly from 2.16~ Vp-p rain in the first session to 1.54 ta Vp-p rain in the last session. On the contrary, the control group did not show constant decreases in EMG levels over sessions. (2) The changes in the heart rate did not correlate with the changes in EMG activity. (3) The changes in the respiratory rate correlated with the changes in EMG activity.
Since the success of Kimmel and Hill (1960) in operantly conditioning the GSR, many reports on operant conditioning of autonomic responses have been published (Fromer, 1963; Shapiro, Crider, & Turskey, 1964; Miller & Carmona, 1967). Not only autonomic responses but also modification of the EEG, EMG, evoked potentials, and other parameters have been studied using operant conditioning procedures, generally called biofeedback.
61 0363-3586/78/0300-0061505.00/0
© 1978 Plenum Publishing C o r p o r a t i o n
62
Ohno et al.
Modification of the EMG by biofeedback procedures in normal subjects has been reported by many researchers (Budzynski & Stoyva, 1969; Coursey, 1975; Reinking & Kohl, 1975; Haynes, Moseley, & McGowan, 1975; Kinsman, O'Brian, Robinson, & Staudenmayer, 1975), but little is known about the relationship between the EMG changes and the changes in the other physiological responses. This study was undertaken to evaluate the biofeedback modification of frontal EMG in normal subjects and the relationship between the EMG changes and the changes in the other physiological responses (the heart rate and the respiratory rate).
METHODS
Subjects The study was performed on 20 healthy subjects (6 males and 14 females), 19 to 33 years of age (average, 21.1 years). The subjects were paid one thousand yen (approximately $3.00) for each session.
Apparatus The apparatus used was a Model B-1 Bio-Electric Information Feedback System unit designed by Bio-Feedback Systems, Inc. (U.S.A.), which integrates EMG activity and shows precise quantification of the muscle tension levels on a Nixie digit readout. It can also inform the subjects of the EMG changes by changes in light and sound. For visual feedback, a red light indicates a high level, amber a moderate level, and green a low activity level. Auditory feedback is given by a continuous series of clicks and a pure variable-frequency audio tone. The rate of these clicks is identical to the counting rate of the Nixie readout, whereas the audio tone becomes louder and higher pitched with an increase in EMG activity. The gain of the tone can be adjusted to permit gradual shaping of deep muscle relaxation. In this study, however, only auditory feedback was employed, and shaping was not done by a fixed feedback gain setting. The EMG of the frontalis muscle was recorded bipolarly by surface electrodes, 1.2 cm in diameter, placed 2.5 cm above each eyebrow, and spaced 9 cm apart on the subject's forehead. One reference electrode was placed in the center of the forehead. A high EMG level produced a high click rate. As the EMG level declined, the click frequency decreased. The subjects were seated with the eyes closed in an easy chair in a dimly lit, electrically shielded room, the temper-
EMG Biofeedback in Normal Subjects
63
ature of which was maintained at 20-22°C. The ECG and respiration were recorded simultaneously using a Polygraph (Sanei Sokki Co.). Procedure Subjects were randomly divided into two groups of 10: (1) the biofeedback group, which was told to relax their forehead by the help of the sound proportional to the integrated EMG level, and (2) the control group, which was told to relax their forehead without any other information. No audio signals were presented to the control group. All subjects were instructed to refrain from moving and falling asleep, and to keep breathing as regularly as possible. The biofeedback group was told to relax the forehead, i.e., to make the click rate slow down. The control group was just told to relax the forehead. Each of the two groups received five training sessions of about 40 minutes' duration each on different days over a period of 2 weeks. One session consisted of a pretraining period of 3 minutes, a training period of 27 minutes (seven experimental trials of 3 minutes each and six intertrial rests of 1 minute each), and a posttraining period of 3 minutes. During a trial of 3 minutes, integration of the EMG for 64 seconds was done twice, and the mean value of the two integral values was used for the following analysis. The EMG level in the pretraining period of each session was used as a baseline value. Each subject was subjected to psychological tests (Cornell Medical Index and Manifest Anxiety Scale) before the experiment and was interviewed after the experiment to determine his subjective feelings during the sessions.
RESULTS The mean levels of frontal EMG activity in the biofeedback and control groups are summarized in Table I. In the first session, the biofeedback group showed a mean 2.16 t~Vp-p min as the baseline value. This decreased progressively with the number of trials and reached the minimum level of 1.45/~Vp-p min in the fourth trial. The control group, however, showed a mean 2.15/~Vp-p min as the baseline value, and the levels did not change progressively in a particular direction with subsequent trials. In the fifth session, the biofeedback group showed a mean baseline value of 2.04/aVp-p min, which decreased with trials, reaching a minimum level of 1.44/~Vp-p min in the seventh trial. The control group showed a mean of 1.89/~rp_p min as the baseline value and did not show a decreased EMG level in most trials. Similar patterns were observed
64
Ohno et al.
Table I. Mean Levels of Frontal EMG Activitya Trial Session
Baseline
1
2
3
4
5
6
7
After
Mean level of 7 trials
Biofeedback group 1 2 3 4 5
2.16 2.00 2.09 2.14 2.33 1 . 9 8 1.97 2.00 2.04 1 . 8 7
1.86 1.77 1.55 1.71 1.58
1.68 1.64 1.52 1.58 1.51
1.45 1.60 1.55 1.35 1.49
1.51 1.48 1.50 1.54 1.46
1.45 1.45 1.49 1.36 1.46
1.48 1.52 1.49 1.36 1.44
2.18 2.25 2.07 1.65 1.88
1.63 1.66 1.58 1.56 1.54
Control group 1 2 3 4 5
2.15 2.07 2.12 1.76 1.89
2.45 2.08 2.04 1.58 2.02
2.47 2.14 1.96 1.58 2.05
1.84 2.19 2.07 1.83 2.03
1.90 2.07 1.72 1.71 2.00
2.10 2.51 2.43 1.93 1.81
1.95 2.29 2.55 1.83 2.00
2.63 2.26 2.08 2.61 1.81
2.14 2.20 2.13 1.76 2.05
2.24 2.10 2.16 1.87 2.47
aUnit:/~Vp-p minutes. for t h e s eco n d , t h i r d , a n d f o u r t h sessions (Table I). F o r simplicity, the m e a n E M G levels o f all seven trials f o r each session are s h o w n in Fig. 1. T h e baseline v a l u e o f t h e first session was t a k e n as the baseline v al u e o f t h e c u r v e . In the b i o f e e d b a c k g r o u p , m e a n E M G levels d ecr eased p r o gressively a n d m a r k e d l y f r o m 2 . 1 6 / a V p - p m i n in t h e first session t o 1.54 /aVp-p m i n in the last session. j~V
: o
2.5"
= bioteedback group o control group
2.0'
L
1.5I i
~,_ i line
~
~ SESSION
Fig. 1. The mean EMG levels of all seven trials for each session.
EMG Biofeedback in Normal Subjects
65
An analysis of variance was applied to confirm the changes of the EMG activity, under the assumption that the experiment was performed on nested design (Table II). There was a significant effect for trials, group X trial, session X trial, individual X session X trial, and for group X session X trial. Namely, the biofeedback group and the control group showed a significant EMG differences with the number of trials in each session. Group X session X trial was highly significant, meaning that the biofeedback group could get a better training effect over sessions once they became good at relaxing their forehead EMG. The differences between the baseline value and the EMG levels in the seventh trial were tested in each session applying t test. The differences were statistically significant except for those in the second session in the biofeedback group (Is, t = 2.96, p < .05; 2s, t = 2.03, n.s.; 3s, t = 3.68, p < .01; 4s, t = 2.79, p < .05; 5s, t = 3.22, p < .05). The control group did not show any significant differences. Furthermore, the differences between groups in the baseline and the seventh trials in each sessions were not significant using t test. The heart rate (obtained from the ECG) and the respiratory rate were subjected to the same analysis of variance. Although the heart rate and the respiratory rate decreased slightly during the session, there were no significant differences between the groups, the sessions, the group X session, the group X trial, etc. The changes in heart rate did not correlate with the changes in EMG activity (correlation coefficient = .0899, n.s.), but the changes in respira-
Table II. Analysis o f Variance of Nested Design E x p e r i m e n t on EMG S.V. Individuals Groups E1 Sessions Individual X session G r o u p X session E2 Trials Individual X trial G r o u p × trial E3 Session × trial Individual X session X trial G r o u p × session × trial E4
SS
df
MS
633.3931 33.9895 601.4036 8.0864 174.7906 1.0847 173.7059 21.7349 98.2962 13.0629 85.2336 9.3861
19 1 18 4 76 4 72 8 152 8 144 32
33.3365 31.9895 33.4113 2.0216 2.2999 .2712 2.4126 2.7169 .6467 1.6329 .5919 .2933
184.5745
608
98.4200 89.1545
32 576
F
p
.9978 .9574
n.s. n.s.
.8378 .9533 .1124
n.s. n.s. n.s.
4.5901 1.0926 2.7588
p < .01 n.s. p < .01
1.9606
p < .01
.3036
2.0294
p < .01
3.0756 .1496
20.5588
p < .01
66
Ohno et al.
Table III. Results of PsychologicalTests Region by CMIa MAS (M +-SD)
PF group Control group Test
12.4 -+6.1 14.9 -+4.5 t = -1.045 n-s-
I
II
III
IV
3 6 1 0 5 2 3 0 p (F) = .190n.s(2 x 3)
aA discriminative chart as a convenient test in screening neurotics by CMI was drawn by Fukamachi (1959). The chart is divided into four regions. Region I: diagnosed to be normal. Region II: provisionally diagnosed to be normal. Region III: provisionally diagnosed to be neurotic. Region IV: diagnosed to be neurotic.
tory rate correlated with the changes in E M G activity (correlation coefficient = .2299, p < .05). Most o f the subjects were in the normal range according to scores f r o m the Cornell Medical Index revised by Fukamachi (1959) and Manifest Anxiety Scale (Table III). There were no differences of scores of these psychological tests between the biofeedback and control groups. Finally, verbal reports of the subjective experiences o f subjects during the experiment are summarized as follows: The subjects in the biofeedback group said they could estimate the degree of their muscular tension and relax themselves well, and that it was not very effective to force the relaxation. However, some subjects said that to know the results quickly made them tense in case they could not achieve the relaxation. Most subjects in the control group reported they were not sure of their muscle relaxation.
DISCUSSION There have been m a n y reports on the beneficial effects o f E M G biofeedback. This is used alone or in combination with existing relaxation methods to treat patients with chronic headache (Wickramasekera, 1972; Budzynski, Stoyva, Adler, & Mullaney, 1973), anxiety, insomnia (Raskin, Johnson, & Rondestvedt, 1973), cervical injury (Jacob & Felton, 1969), and writer's cramp (Reavley, 1975), etc. Budzynski et al. (1973) reported that E M G feedback was useful in training patients with tension headache to relax their frontalis muscle and obtain relief f r o m headache. Wickramasekera (1972) also reported that 5 female patients with chronic headache reduced the frequency and intensity of headaches with contingent E M G
EMG Biofeedback in Normal Subjects
67
feedback. Raskin et al. (1973) treated 10 chronically anxious patients with frontalis E M G feedback training. One patient improved markedly, 3 moderately, and 6 showed no improvement. Furthermore, there are many reports of E M G modification by biofeedback procedures employing normal subjects (Budzynski & Stoyva, 1969; Coursey, 1975; Reinking & Kohl, 1975; Haynes et al., 1975; Kinsman et al., 1975). We carried out a controlled study on the voluntary control of the frontalis muscle by biofeedback procedures employing 20 normal subjects. On this occasion, we made each experimental trial fairly short (3 min) in order to avoid the influence of muscle relaxation induced by drowsiness and thus get complete muscle relaxation in a short time. The subjects were divided into two groups, the biofeedback group and the control group. After five training sessions o f 40 minutes each, it was shown that the biofeedback group relaxed the forehead muscle more quickly and consistently than the control group. The results are similar to those o f Budzynski & Stoyva (1969) and other researchers. Our results are of interest not only as confirmatory evidence obtained in a different laboratory but because we used subjects o f a different race (Japanese) and the period o f each trial was shorter than theirs. The E M G levels in the experiment were generally low relative to anything reported thus far. This may be due to the fact the subjects were young and healthy normals. Furthermore, little is known about the relationship between the E M G changes and the changes in the other physiological responses. In our study, the heart rate did not correlate with changes in the E M G , although the respiratory rate correlated with changes in the EMG, showing specificity o f learning partly. According to verbal reports it is suggested that subjects o f the biofeedback group might acquire a mental control o f muscle relaxation as in autogenic training. REFERENCES
Budzynski, "lZ.H., & Stoyva, J. M. An instrument for producing deep relaxation by means of analog information feedback. Journal of Applied Behavior Analysis, 1969,2, 231-237. Budzynski, T. H., Stoyva, J. M., Adler, C. S., & Mullaney, D. J. EMG biofeedback and tension headache: A controlled outcome study. Psychosomatic Medicine, 1973, 35, 484-496. Coursey, R. D. Electromyograph feedback as a relaxation technique. Journal of Counseling and Clinical Psychology, 1975, 43, 825-834. Fromer, R. Conditioned vasomotor responsesin the rabbit. Journal of Comparative and Physiological Psychology, 1963, 56, 1050-1055. Fukamachi, K. The study on the Cornell Medical Index (II). Fukuoka Acta Medica, 1959, 50, 3001-3009. Haynes, S. N., Moseley,D., & McGowan, W. T. Relaxation training and biofeedback in the reduction of frontalis muscle tension. Psychophysiology, 1975, 12, 547-552.
68
Ohno et al.
Jacob, A., & Felton, G. S. Visual feedback of myeloelectric output to facilitate muscle relaxation in normal persons and patients with neck injuries. Archives of Physical Medicine and Rehabilitation, 1969, 50, 34-39. Kimmel, H. D., & Hill, F. A. Operant conditioning of the GSR. Psychological Report, 1960, 7, 555-562. Kinsman, R. A., O'Brian, K., Robinson, S., & Staudenmayer, H. Continuous biofeedback and discrete posttrial verbal feedback in frontalis muscle relaxation. Psychophysiology, 1975, 12, 30-35. Miller, N. E., & Carmona, A. Modification of a visceral response, salivation in thirsty dogs by instrumental training with water reward. Journal of Comparative and Physiological Psychology, 1967, 63, 1-6. Raskin, M., Johnson, G., & Rondestvedt, J. W. Chronic anxiety treated by feedback-induced muscle relaxation. Archives of General Psychiatry, 1973, 28, 263-267. Reavley, W. The use of feedback in the treatment of writer's cramp. Journal of Behavior Therapy and Experimental Psychiatry, 1975, 6, 335-338. Reinking, R. H., & Kohl, M. L. Effects of various forms of relaxation training on physiological and self-report measures of relaxation. Journal of Counseling and Clinical Psychology, 1975, 43, 595-600. Shapiro, D., Crider, A. B., & Turskey, B. Differentiation of an autonomic response through operant reinforcement. Psychosomatic Science, 1964, 1, 147-148. Wickramasekera, I. Electromyographic feedback training and tension headache: Preliminary observations. American Journal of Clinical Hypnosis, 1972, 15, 83-85. (Revision received July 18, 1977)
Biofeedback Modification of Frontal EMG in Normal Subjects Yoshiteru Ohno, Yoshiharu Tanaka, Tsutomu Takeya, and Hideki Matsubara Department of Psychosomatic Medicine Kyushu University, Fukuoka City, Japan Norikazu Kuriya Inatsuki Hospital, Fukuoka Sadao Komemushi Department of PharmaceuticalEngineering Osaka University We carried out a controlled study on the voluntary control o f the frontalis muscle by biofeedback procedures employing 20 normal subjects. Subjects were randomly divided into two groups of 10: (1) the biofeedback group and (2) the control group. Each o f the two groups received five training sessions o f about 40 minutes' duration each on different days. The results obtained are as follows: (1) In the biofeedback group, mean EMG levels decreused progressively and markedly from 2.16~ Vp-p rain in the first session to 1.54 ta Vp-p rain in the last session. On the contrary, the control group did not show constant decreases in EMG levels over sessions. (2) The changes in the heart rate did not correlate with the changes in EMG activity. (3) The changes in the respiratory rate correlated with the changes in EMG activity.
Since the success of Kimmel and Hill (1960) in operantly conditioning the GSR, many reports on operant conditioning of autonomic responses have been published (Fromer, 1963; Shapiro, Crider, & Turskey, 1964; Miller & Carmona, 1967). Not only autonomic responses but also modification of the EEG, EMG, evoked potentials, and other parameters have been studied using operant conditioning procedures, generally called biofeedback.
61 0363-3586/78/0300-0061505.00/0
© 1978 Plenum Publishing C o r p o r a t i o n
62
Ohno et al.
Modification of the EMG by biofeedback procedures in normal subjects has been reported by many researchers (Budzynski & Stoyva, 1969; Coursey, 1975; Reinking & Kohl, 1975; Haynes, Moseley, & McGowan, 1975; Kinsman, O'Brian, Robinson, & Staudenmayer, 1975), but little is known about the relationship between the EMG changes and the changes in the other physiological responses. This study was undertaken to evaluate the biofeedback modification of frontal EMG in normal subjects and the relationship between the EMG changes and the changes in the other physiological responses (the heart rate and the respiratory rate).
METHODS
Subjects The study was performed on 20 healthy subjects (6 males and 14 females), 19 to 33 years of age (average, 21.1 years). The subjects were paid one thousand yen (approximately $3.00) for each session.
Apparatus The apparatus used was a Model B-1 Bio-Electric Information Feedback System unit designed by Bio-Feedback Systems, Inc. (U.S.A.), which integrates EMG activity and shows precise quantification of the muscle tension levels on a Nixie digit readout. It can also inform the subjects of the EMG changes by changes in light and sound. For visual feedback, a red light indicates a high level, amber a moderate level, and green a low activity level. Auditory feedback is given by a continuous series of clicks and a pure variable-frequency audio tone. The rate of these clicks is identical to the counting rate of the Nixie readout, whereas the audio tone becomes louder and higher pitched with an increase in EMG activity. The gain of the tone can be adjusted to permit gradual shaping of deep muscle relaxation. In this study, however, only auditory feedback was employed, and shaping was not done by a fixed feedback gain setting. The EMG of the frontalis muscle was recorded bipolarly by surface electrodes, 1.2 cm in diameter, placed 2.5 cm above each eyebrow, and spaced 9 cm apart on the subject's forehead. One reference electrode was placed in the center of the forehead. A high EMG level produced a high click rate. As the EMG level declined, the click frequency decreased. The subjects were seated with the eyes closed in an easy chair in a dimly lit, electrically shielded room, the temper-
EMG Biofeedback in Normal Subjects
63
ature of which was maintained at 20-22°C. The ECG and respiration were recorded simultaneously using a Polygraph (Sanei Sokki Co.). Procedure Subjects were randomly divided into two groups of 10: (1) the biofeedback group, which was told to relax their forehead by the help of the sound proportional to the integrated EMG level, and (2) the control group, which was told to relax their forehead without any other information. No audio signals were presented to the control group. All subjects were instructed to refrain from moving and falling asleep, and to keep breathing as regularly as possible. The biofeedback group was told to relax the forehead, i.e., to make the click rate slow down. The control group was just told to relax the forehead. Each of the two groups received five training sessions of about 40 minutes' duration each on different days over a period of 2 weeks. One session consisted of a pretraining period of 3 minutes, a training period of 27 minutes (seven experimental trials of 3 minutes each and six intertrial rests of 1 minute each), and a posttraining period of 3 minutes. During a trial of 3 minutes, integration of the EMG for 64 seconds was done twice, and the mean value of the two integral values was used for the following analysis. The EMG level in the pretraining period of each session was used as a baseline value. Each subject was subjected to psychological tests (Cornell Medical Index and Manifest Anxiety Scale) before the experiment and was interviewed after the experiment to determine his subjective feelings during the sessions.
RESULTS The mean levels of frontal EMG activity in the biofeedback and control groups are summarized in Table I. In the first session, the biofeedback group showed a mean 2.16 t~Vp-p min as the baseline value. This decreased progressively with the number of trials and reached the minimum level of 1.45/~Vp-p min in the fourth trial. The control group, however, showed a mean 2.15/~Vp-p min as the baseline value, and the levels did not change progressively in a particular direction with subsequent trials. In the fifth session, the biofeedback group showed a mean baseline value of 2.04/aVp-p min, which decreased with trials, reaching a minimum level of 1.44/~Vp-p min in the seventh trial. The control group showed a mean of 1.89/~rp_p min as the baseline value and did not show a decreased EMG level in most trials. Similar patterns were observed
64
Ohno et al.
Table I. Mean Levels of Frontal EMG Activitya Trial Session
Baseline
1
2
3
4
5
6
7
After
Mean level of 7 trials
Biofeedback group 1 2 3 4 5
2.16 2.00 2.09 2.14 2.33 1 . 9 8 1.97 2.00 2.04 1 . 8 7
1.86 1.77 1.55 1.71 1.58
1.68 1.64 1.52 1.58 1.51
1.45 1.60 1.55 1.35 1.49
1.51 1.48 1.50 1.54 1.46
1.45 1.45 1.49 1.36 1.46
1.48 1.52 1.49 1.36 1.44
2.18 2.25 2.07 1.65 1.88
1.63 1.66 1.58 1.56 1.54
Control group 1 2 3 4 5
2.15 2.07 2.12 1.76 1.89
2.45 2.08 2.04 1.58 2.02
2.47 2.14 1.96 1.58 2.05
1.84 2.19 2.07 1.83 2.03
1.90 2.07 1.72 1.71 2.00
2.10 2.51 2.43 1.93 1.81
1.95 2.29 2.55 1.83 2.00
2.63 2.26 2.08 2.61 1.81
2.14 2.20 2.13 1.76 2.05
2.24 2.10 2.16 1.87 2.47
aUnit:/~Vp-p minutes. for t h e s eco n d , t h i r d , a n d f o u r t h sessions (Table I). F o r simplicity, the m e a n E M G levels o f all seven trials f o r each session are s h o w n in Fig. 1. T h e baseline v a l u e o f t h e first session was t a k e n as the baseline v al u e o f t h e c u r v e . In the b i o f e e d b a c k g r o u p , m e a n E M G levels d ecr eased p r o gressively a n d m a r k e d l y f r o m 2 . 1 6 / a V p - p m i n in t h e first session t o 1.54 /aVp-p m i n in the last session. j~V
: o
2.5"
= bioteedback group o control group
2.0'
L
1.5I i
~,_ i line
~
~ SESSION
Fig. 1. The mean EMG levels of all seven trials for each session.
EMG Biofeedback in Normal Subjects
65
An analysis of variance was applied to confirm the changes of the EMG activity, under the assumption that the experiment was performed on nested design (Table II). There was a significant effect for trials, group X trial, session X trial, individual X session X trial, and for group X session X trial. Namely, the biofeedback group and the control group showed a significant EMG differences with the number of trials in each session. Group X session X trial was highly significant, meaning that the biofeedback group could get a better training effect over sessions once they became good at relaxing their forehead EMG. The differences between the baseline value and the EMG levels in the seventh trial were tested in each session applying t test. The differences were statistically significant except for those in the second session in the biofeedback group (Is, t = 2.96, p < .05; 2s, t = 2.03, n.s.; 3s, t = 3.68, p < .01; 4s, t = 2.79, p < .05; 5s, t = 3.22, p < .05). The control group did not show any significant differences. Furthermore, the differences between groups in the baseline and the seventh trials in each sessions were not significant using t test. The heart rate (obtained from the ECG) and the respiratory rate were subjected to the same analysis of variance. Although the heart rate and the respiratory rate decreased slightly during the session, there were no significant differences between the groups, the sessions, the group X session, the group X trial, etc. The changes in heart rate did not correlate with the changes in EMG activity (correlation coefficient = .0899, n.s.), but the changes in respira-
Table II. Analysis o f Variance of Nested Design E x p e r i m e n t on EMG S.V. Individuals Groups E1 Sessions Individual X session G r o u p X session E2 Trials Individual X trial G r o u p × trial E3 Session × trial Individual X session X trial G r o u p × session × trial E4
SS
df
MS
633.3931 33.9895 601.4036 8.0864 174.7906 1.0847 173.7059 21.7349 98.2962 13.0629 85.2336 9.3861
19 1 18 4 76 4 72 8 152 8 144 32
33.3365 31.9895 33.4113 2.0216 2.2999 .2712 2.4126 2.7169 .6467 1.6329 .5919 .2933
184.5745
608
98.4200 89.1545
32 576
F
p
.9978 .9574
n.s. n.s.
.8378 .9533 .1124
n.s. n.s. n.s.
4.5901 1.0926 2.7588
p < .01 n.s. p < .01
1.9606
p < .01
.3036
2.0294
p < .01
3.0756 .1496
20.5588
p < .01
66
Ohno et al.
Table III. Results of PsychologicalTests Region by CMIa MAS (M +-SD)
PF group Control group Test
12.4 -+6.1 14.9 -+4.5 t = -1.045 n-s-
I
II
III
IV
3 6 1 0 5 2 3 0 p (F) = .190n.s(2 x 3)
aA discriminative chart as a convenient test in screening neurotics by CMI was drawn by Fukamachi (1959). The chart is divided into four regions. Region I: diagnosed to be normal. Region II: provisionally diagnosed to be normal. Region III: provisionally diagnosed to be neurotic. Region IV: diagnosed to be neurotic.
tory rate correlated with the changes in E M G activity (correlation coefficient = .2299, p < .05). Most o f the subjects were in the normal range according to scores f r o m the Cornell Medical Index revised by Fukamachi (1959) and Manifest Anxiety Scale (Table III). There were no differences of scores of these psychological tests between the biofeedback and control groups. Finally, verbal reports of the subjective experiences o f subjects during the experiment are summarized as follows: The subjects in the biofeedback group said they could estimate the degree of their muscular tension and relax themselves well, and that it was not very effective to force the relaxation. However, some subjects said that to know the results quickly made them tense in case they could not achieve the relaxation. Most subjects in the control group reported they were not sure of their muscle relaxation.
DISCUSSION There have been m a n y reports on the beneficial effects o f E M G biofeedback. This is used alone or in combination with existing relaxation methods to treat patients with chronic headache (Wickramasekera, 1972; Budzynski, Stoyva, Adler, & Mullaney, 1973), anxiety, insomnia (Raskin, Johnson, & Rondestvedt, 1973), cervical injury (Jacob & Felton, 1969), and writer's cramp (Reavley, 1975), etc. Budzynski et al. (1973) reported that E M G feedback was useful in training patients with tension headache to relax their frontalis muscle and obtain relief f r o m headache. Wickramasekera (1972) also reported that 5 female patients with chronic headache reduced the frequency and intensity of headaches with contingent E M G
EMG Biofeedback in Normal Subjects
67
feedback. Raskin et al. (1973) treated 10 chronically anxious patients with frontalis E M G feedback training. One patient improved markedly, 3 moderately, and 6 showed no improvement. Furthermore, there are many reports of E M G modification by biofeedback procedures employing normal subjects (Budzynski & Stoyva, 1969; Coursey, 1975; Reinking & Kohl, 1975; Haynes et al., 1975; Kinsman et al., 1975). We carried out a controlled study on the voluntary control of the frontalis muscle by biofeedback procedures employing 20 normal subjects. On this occasion, we made each experimental trial fairly short (3 min) in order to avoid the influence of muscle relaxation induced by drowsiness and thus get complete muscle relaxation in a short time. The subjects were divided into two groups, the biofeedback group and the control group. After five training sessions o f 40 minutes each, it was shown that the biofeedback group relaxed the forehead muscle more quickly and consistently than the control group. The results are similar to those o f Budzynski & Stoyva (1969) and other researchers. Our results are of interest not only as confirmatory evidence obtained in a different laboratory but because we used subjects o f a different race (Japanese) and the period o f each trial was shorter than theirs. The E M G levels in the experiment were generally low relative to anything reported thus far. This may be due to the fact the subjects were young and healthy normals. Furthermore, little is known about the relationship between the E M G changes and the changes in the other physiological responses. In our study, the heart rate did not correlate with changes in the E M G , although the respiratory rate correlated with changes in the EMG, showing specificity o f learning partly. According to verbal reports it is suggested that subjects o f the biofeedback group might acquire a mental control o f muscle relaxation as in autogenic training. REFERENCES
Budzynski, "lZ.H., & Stoyva, J. M. An instrument for producing deep relaxation by means of analog information feedback. Journal of Applied Behavior Analysis, 1969,2, 231-237. Budzynski, T. H., Stoyva, J. M., Adler, C. S., & Mullaney, D. J. EMG biofeedback and tension headache: A controlled outcome study. Psychosomatic Medicine, 1973, 35, 484-496. Coursey, R. D. Electromyograph feedback as a relaxation technique. Journal of Counseling and Clinical Psychology, 1975, 43, 825-834. Fromer, R. Conditioned vasomotor responsesin the rabbit. Journal of Comparative and Physiological Psychology, 1963, 56, 1050-1055. Fukamachi, K. The study on the Cornell Medical Index (II). Fukuoka Acta Medica, 1959, 50, 3001-3009. Haynes, S. N., Moseley,D., & McGowan, W. T. Relaxation training and biofeedback in the reduction of frontalis muscle tension. Psychophysiology, 1975, 12, 547-552.
68
Ohno et al.
Jacob, A., & Felton, G. S. Visual feedback of myeloelectric output to facilitate muscle relaxation in normal persons and patients with neck injuries. Archives of Physical Medicine and Rehabilitation, 1969, 50, 34-39. Kimmel, H. D., & Hill, F. A. Operant conditioning of the GSR. Psychological Report, 1960, 7, 555-562. Kinsman, R. A., O'Brian, K., Robinson, S., & Staudenmayer, H. Continuous biofeedback and discrete posttrial verbal feedback in frontalis muscle relaxation. Psychophysiology, 1975, 12, 30-35. Miller, N. E., & Carmona, A. Modification of a visceral response, salivation in thirsty dogs by instrumental training with water reward. Journal of Comparative and Physiological Psychology, 1967, 63, 1-6. Raskin, M., Johnson, G., & Rondestvedt, J. W. Chronic anxiety treated by feedback-induced muscle relaxation. Archives of General Psychiatry, 1973, 28, 263-267. Reavley, W. The use of feedback in the treatment of writer's cramp. Journal of Behavior Therapy and Experimental Psychiatry, 1975, 6, 335-338. Reinking, R. H., & Kohl, M. L. Effects of various forms of relaxation training on physiological and self-report measures of relaxation. Journal of Counseling and Clinical Psychology, 1975, 43, 595-600. Shapiro, D., Crider, A. B., & Turskey, B. Differentiation of an autonomic response through operant reinforcement. Psychosomatic Science, 1964, 1, 147-148. Wickramasekera, I. Electromyographic feedback training and tension headache: Preliminary observations. American Journal of Clinical Hypnosis, 1972, 15, 83-85. (Revision received July 18, 1977)
