Angular Biofeedback Device for Sitting Balance of Stroke Patients
Background and Purpose Impaired sitting balance is an important and time-consuming complication for stroke patients. We examined the effect of the use of an angular biofeedback device in addition to physical therapy in training stroke patients with impaired sitting balance compared with outcome in patients receiving conventional physical therapy only.
Methods The biofeedback group consisted of 24 patients who received angular biofeedback training in addition to conventional physical therapy. The number of biofeedback signals and the lengths of time a patient could sit balanced throughout a period of 5 minutes before the training program, after 10 days of treatment, and at discharge were recorded and compared with those of the control group of 13 patients who received conventional physical therapy only.
Results It was found that 75% of the biofeedback group gained sitting balance after 10 days of treatment in comparison with 15.4% of the control group (P<.001). At discharge, 91.6% of the biofeedback group and 84.6% of the control group gained sitting balance (P=.510), and 45.8% of the biofeedback group and 46.2% of the control group managed independent ambulation (P=.985). The mean rehabilitation periods among the ambulatory patients of the biofeedback and control groups were 9.45±0.71 and 13.83±1.70 weeks, respectively (P=.049). The mean training time in which the biofeedback group gained sitting balance was significantly shorter than that of the control group (P<.001).
Conclusions Angular biofeedback intervention, by providing earlier postural trunk control, is a useful adjunct to conventional physical therapy in the rehabilitation of stroke patients with impaired sitting balance.
Biofeedback is a technique of increasing importance in the field of rehabilitation. Biofeedback training is considered an adjunct to other treatment procedures used in the rehabilitation of stroke patients, and its applications have grown in number as clinician demand has increased.
Impaired sitting balance is not a rare complication in the treatment of stroke patients, usually occurring shortly after the onset of stroke and especially in cases with bilateral involvement. A severe stroke will cause an absence of righting and equilibrium reactions; however, after a mild stroke, these reactions are present but decreased in quality and timing and/or delayed. Good sitting balance is a prerequisite for functional transfers, standing balance, and ambulation of stroke patients. Visual, proprioceptive, vestibular, and auditory input are important to help a patient regain good sitting balance.1
The present study was undertaken to determine whether sitting balance can be improved and the hospitalization period for rehabilitation shortened by use of an angular biofeedback device for stroke patients with impaired sitting balance.
Subjects and Methods
We developed, implemented, and used a simple angular biofeedback device to improve the postural trunk control of hemiplegic patients. This system consists of the following parts (Fig 1⇓): an adjustable, multidimensional, and round on-off mercury switch; an electronic digital counter and its screen; an audio amplifier and a loudspeaker; and a warning lamp. The mercury switch is a small metal cylinder containing mercury of 99% purity (Fig 2⇓). Parts A and B are made of conductive material, isolated from each other and connected to the electronic counter. The cover and surrounding areas of the mercury switch are also made of conductive material and are isolated from part A. When the mercury switch is placed in a vertical position, the mercury inside the switch does not come into contact with part A. Since there is no contact between parts A and B, the switch is off. When the position of the mercury switch is changed from vertical to horizontal in any direction, and when the angle of the slope is sufficient, the mercury comes into contact with the metal of part A, causing a short circuit between parts A and B. When this occurs, the switch is on, and it sends a signal to the electronic counter. The degree of sensitivity of the mercury switch can be adjusted manually by tightening or loosening the screw-type cover. The tighter the cover is screwed on, the more sensitive the switch will be to the slightest tilting of the patient.
The mercury switch is placed on the midline of the upper back portion of the body at vertebra T-7 to T-8 level and is attached firmly with straps. The other parts of the system are placed in front of the patient in such a way that the visual and auditory signals can be easily noticed. When the patient tilts from the erect position in any direction, shifting of the mercury causes a short circuit, and the loudspeaker and the warning lamp give feedback to the patient. The electronic counter increases its previous record number with each tilting of the patient, and the number of “dispostures” is calculated over a certain period of time.
Thirty-seven patients with impaired sitting balance were selected among hemiplegic patients who attended Ankara Rehabilitation Center. All these patients had undergone a detailed physical and neurological examination. The biofeedback group consisted of 24 patients who each received angular biofeedback training (30 minutes per day) plus conventional physical therapy (PT; 2.5 hours per day). The control group consisted of 13 patients who each received conventional PT only (3 hours per day). Conventional PT included control and coordination exercises, range-of-motion and selective stretching exercises, muscle-strengthening exercises, and training in activities of daily living. Neurophysiological therapies, other biofeedback applications, and functional electrical stimulation were not used. During the training program, the cover of the mercury switch was adjusted according to the severity of impairment of each patient, and the patient was supported if necessary. As the sitting balance of the patient improved, the cover of the mercury switch was readjusted to a more advanced level, which produces biofeedback signals with sways of smaller degrees. The number of biofeedback signals and the lengths of time a patient could sit balanced throughout a period of 5 minutes before the training program, after 10 days of treatment, and at discharge were recorded for each group. To ensure standardization during these recordings, the cover of the mercury switch was fixed at a level that allowed the patient a maximum of 5° of tilting without producing biofeedback signals.
Demographic results were descriptive and expressed as percent or as mean±standard error. Comparison between the groups of the demographic results, number of biofeedback signals, lengths of time the patients sat balanced, and lengths of the rehabilitation period were performed by using χ2 and unpaired t tests. Comparisons of the records before treatment, after 10 days of treatment, and at discharge within each group were performed using one-way ANOVA for paired data.
Table 1⇓ shows mean age, duration of disease, distribution of gender, side of involvement, and type of stroke of the biofeedback and control groups. No significant differences were found between the two groups regarding these data. Proprioceptive loss was present in 6 (25%) of the biofeedback and 5 (38.5%) of the control patients (P=.306). Neither of the patient groups showed markedly increased spasticity, which prevents the passive movement of the joints of the affected limbs through the full range of motion.
Table 2⇓ displays changes in mean number of the biofeedback signals (“dispostures” of patient) for each group. The initial recordings showed no significant difference between the biofeedback and control groups (P=.453). After 10 days of treatment and at discharge, the mean number of signals decreased significantly in both the biofeedback (P<.001) and control (P<.001) groups. However, the mean number of signals recorded after 10 days of treatment of the biofeedback group was significantly lower than that of the control group (P=.005). Recordings of the biofeedback and control groups at discharge showed a similar pattern (P=.907).
Before the training program, in 6 (25%) of the biofeedback and 3 (23.1%) of the control patients who were not able to sit in an erect position at all, the biofeedback signals were continuous. The mean number of biofeedback signals of patients in the biofeedback group dropped to 14 after 10 days of treatment and to 4 at discharge. One of the 3 patients in the control group still produced continuous signals after 10 days of conventional PT, and the other 2 gave 20 and 36 signals. At discharge, these recordings had decreased to 76, 4, and 0, respectively.
Table 3⇓ shows changes in mean length of time for which the patients could sit without producing biofeedback signals throughout a period of 5 minutes. After 10 days of treatment and at discharge, the mean length of time the patients could sit balanced increased significantly in both the biofeedback (P<.001) and control (P<.001) groups. No statistically significant differences were found between the biofeedback and control groups with respect to the initial (P=.959) and discharge (P=.231) recordings. However, recordings of the biofeedback group were significantly higher than those of the control group after 10 days of treatment (P<.001).
Eighteen (75%) of the biofeedback patients compared with 2 (15.4%) of the control patients gained sitting balance after 10 days of treatment (P<.001). At discharge, 22 (91.6%) of the biofeedback group and 11 (84.6%) of the control group were able to sit balanced (P=.510). Postural trunk control of the remaining 4 patients (2 biofeedback and 2 control patients) also improved, but they still needed slight assistance while sitting.
Eleven (45.8%) of the biofeedback patients and 6 (46.2%) of the control patients managed independent ambulation at discharge (P=.985). Whereas the mean length of time in which the ambulatory biofeedback patients gained sitting balance was 8.18±0.87 treatment days, the ambulatory control patients gained sitting balance in a mean of 24±4.49 treatment days (P<.001). After sitting balance was gained, the lengths of the rehabilitation period for the biofeedback and control groups were 7.92±0.82 and 9.17±1.26 weeks, respectively (P=.403). The mean lengths of the rehabilitation program among the ambulatory patients of the biofeedback and control groups were 9.45±0.71 and 13.83±1.70 weeks, respectively. The rehabilitation period of the ambulatory group was significantly shorter than that of the ambulatory control group (P=.049).
In total, 13 (54.2%) of the biofeedback and 7 (53.8%) of the control patients were nonambulatory at discharge. Among the nonambulatory patients, 2 of the biofeedback group (15.4%) and 2 of the control group (28.6%) did not gain sitting balance at discharge (P=.907). The nonambulatory biofeedback and control patients gained sitting balance in a mean of 10.73±0.82 and 27.80±1.66 treatment days, respectively (P<.001). Four (16.7%) of the biofeedback and 1 (7.7%) of the control patients whose rehabilitation programs could not be completed were transferred to the cardiology department because of problems such as unstable angina pectoris, atrial fibrillation, and severe hypertension or hypotension. Nine (37.5%) of the biofeedback and 6 (46.2%) of the control patients were not candidates for ambulation training, and hemiplegic wheelchairs were assigned to these patients. The ambulatory inability of these patients (biofeedback/control) resulted from knee flexion contracture (1/0), reactive depression and lack of motivation (2/1), lack of postural trunk control (1/1), and weak hip muscles (5/4).
The most widely accepted use of biofeedback is for movement disorders. Numerous research efforts using electromyographic (EMG) biofeedback for stroke patients have emerged with very encouraging results, indicating that these applications can enhance function of hemiplegic patients.2 3
EMG biofeedback has been used in attempts to control shoulder subluxation, as well as to improve hand function in the upper extremity4 5 and ambulatory function in the lower extremity.6 7 Positional biofeedback systems have also been successfully used in knee-joint position training of hemiplegic stroke patients.8 9 10
Postural control of the head, neck, and trunk is essential for normal functions of the upper and lower extremities. The hemiplegic patient must have control of the head and trunk to manage shifting and bearing weight on each side to free an extremity for function. The establishment of head, neck, and trunk control allows for dissociation of the shoulder and pelvic girdles from the trunk and dissociation of the extremities from the girdles.
Bjork and Wetzel11 used a positional biofeedback device to improve sitting balance of a bilateral stroke patient and reported that after 2 weeks of training the patient was able to sit independently for 20 minutes. The device they used consisted of a glass mercury switch that was connected to a portable EMG unit. They attached the mercury switch to the shoulder of the patient, and as the patient tilted a noxious auditory signal was produced by the device. However, only the medial-lateral trunk tilt could be monitored by this device, and other motions such as shoulder elevations produced undesired feedback signals. The device we have developed consists of an adjustable, multidimensional, and round mercury switch that is placed on the midline of the upper back portion of the body. Thus, trunk tilting in any direction (including medial, lateral, anterior, and posterior sway and/or any combination) can be monitored, and no other motions such as shoulder elevation will produce undesired feedback signals.
In our study, the biofeedback group was compatible with the control group regarding age, duration of stroke, sex, side of involvement, type of stroke, and presence of proprioceptive loss. The number of biofeedback signals and the lengths of time the patients could sit balanced throughout 5 minutes were recorded at the end of the 10th treatment day because, in our previous study, a positional biofeedback device was able to allow control of genu recurvatum in a very short period.10 The mean number of biofeedback signals decreased, the mean period for which the patients could sit balanced increased significantly in both the biofeedback and control groups, and the number of patients who gained sitting balance at discharge were not significantly different in either of the groups. However, the mean number of the biofeedback signals was significantly less, and the mean period for which patients could sit balanced was significantly longer in the biofeedback group after 10 days of treatment. Moreover, 75% of the biofeedback patients compared with only 15.4% of the control patients gained sitting balance after 10 days. These findings show that although conventional physical therapy is a beneficial therapeutic approach, combining conventional physical therapy with biofeedback training decreases the time needed to achieve good sitting balance. After sitting balance was gained, the rehabilitation period of the nonambulatory patients was interrupted by medical problems, and some of these patients were transferred to other departments. Thus, we could not evaluate and compare the rehabilitation period of these patients. The rehabilitation periods of the ambulatory biofeedback and control patients were properly compared, since their rehabilitation program continued without any interruption. Among the ambulatory patients, the total rehabilitation period of the biofeedback group was significantly shorter. Although not statistically significant, their rehabilitation period after sitting balance was gained was also shorter. The number of patients who managed independent ambulation did not differ between groups. Therefore, in this study biofeedback training did not show any effect on ambulatory status of the patients, and whether it shortened the ambulatory training remains unclear. Its major effect was to reduce the time needed to achieve sitting balance. The ambulatory and nonambulatory biofeedback patients gained sitting balance in a mean of 8 and 11 treatment days, respectively; the ambulatory and nonambulatory control patients gained sitting balance in a mean of 24 and 28 treatment days, respectively. These findings suggest that angular biofeedback training of stroke patients with impaired sitting balance can shorten the hospitalization period of these patients by providing an earlier postural trunk control, which is essential for ambulatory training. Its effect on ambulatory status needs further investigation.
Application of the angular biofeedback device to stroke patients with impaired sitting balance may favorably alter the rehabilitation outcome of these patients. Taken in total perspective, the findings from this study suggest that angular biofeedback training is beneficial in improving postural trunk control among stroke patients.
- Received September 26, 1995.
- Revision received March 25, 1996.
- Accepted March 28, 1996.
- Copyright © 1996 by American Heart Association
Ryerson SD. Hemiplegia resulting from vascular insult or disease. In: Umphred OA, ed. Neurological Rehabilitation. St Louis, Mo: CV Mosby Co; 1990:619-660.
Wolf SL. Use of biofeedback in the treatment of stroke patients. Stroke.. 1990;21:11-22.
Wolf SL. Electromyographic biofeedback applications to stroke patients. Phys Ther.. 1983;63:1449-1455.
Wolf SL, Binder-Macleod SA. Electromyographic biofeedback applications to the hemiplegic patient: changes in upper extremity. Phys Ther.. 1983;63:1393-1402.
Wolf SL, Binder-Macleod SA. Electromyographic biofeedback applications to the hemiplegic patient: changes in lower extremity neuromuscular and functional status. Phys Ther.. 1983;63:1404-1413.
Intiso D, Santilli V, Grasso MG, Rossi R, Caruso I. Rehabilitation of walking with electromyographic biofeedback in foot-drop after stroke. Stroke.. 1994;25:1189-1192.
Dursun E, Ceceli E, Çakçì A, Tüzünalp Ö, Öztürk O, Telatar Z. Effectiveness of a new joint position biofeedback training system development and its application on hemiplegic patients. In: Proceedings of the International Biomedical Engineering Days, August 18-20, 1992; Istanbul, Turkey. Pp 246-248.
Bjork L, Wetzel A. A positional biofeedback device for sitting balance. Phys Ther.. 1983;63:1460-1462.