Restoration of gait during two to three weeks of therapy with multichannel electrical stimulation.

Restoration of Gait During Two to Three Weeks of Therapy with Multichannel Electrical Stimulation Soon after the introduction of the first single-channel electrical stimulators for correction of foot drop, attempts were made to stimulate, in addition to the ankle dorsiflexors, other muscle groups of the hemiplegic leg, which also play an important role in gait. Vodovnik et al suggested in 1965 a six-channel stimulator for stimulation of six muscle groups in the affected leg. [1] Thereafter numerous multichannel devices were developed, principles for closed-loop control were studied, stimulation sequences were tested and analyzed, and gait improvement and therapeutic effects of electrical stimulation were investigated. [2-6] Clinical assessment and kinesiological studies showed that the application of cutaneous electrical stimulation to the six main antagonistic muscle groups of the lower limb could modify pathological gait. [7-10] Multichannel functional electrical stimulation leads to a higher correction level of gait anomalies, to a faster recovery rate, and to better endurance than the standard approaches in motor-disabled patients. [7-10]

Despite good correction of gait, multichannel surface FES is not yet widely used as an orthotic aid because of its complexity, the size of the stimulator, and the numerous electrodes, the positioning of which poses a great problem for the hemiplegic patient. Considerable therapeutic effects, such as the improvement of gait during the swing and the stance phases, a better posture, and higher endurance were also observed in the course of therapy. [7,8] The effects, however, tended to disappear within six months after therapy when no significant difference was found between stimulated and control groups, both under conventional therapy. [7,8] These are probably the main reasons why no reports have been published in the last several years on the use of multichannel surface FES in the rehabilitation of hemiplegic patients.

None of the patients treated in previous studies with multichannel FES suffered from severe motor disability. All of the patients could walk, and multichannel FES was used to correct their pathological gait pattern. Because the short-term results of this therapeutic modality were promising, we decided to apply multichannel FES to a group of severely impaired patients after cerebrovascular insult or craniocerebral trauma. They were disabled to the degree of being incapable of walking or even attaining standing from the wheelchair without the help of a therapist. We expected multichannel FES to help initiate the gait pattern, reestablish antigravity support, and lead to partial or complete ambulation of the patient.

The purposes of this article are 1) to present the approach using intensive initial multichannel FES therapy in severely impaired patients and 2) to evaluate its effects with a stride analyzer incorporated into the stimulator. Measurements with the ground reaction measuring system were also performed at the beginning and at the end of the therapy period to determine changes in the patients' gait pattern.

Method

Instrumentation

The stimulator used in this study contained six independent, galvanically separated channels with intermittent stimulation pulses, each optionally triggered by a left or a right heel-switch. Stimulation during one gait cycle (stimulation sequence) was timed for each channel by using 16 switches, 8 for the stance phase and 8 for the swing phase (Fig. 1). When the switch was on, the stimulation occurred during the selected time interval of the stride phase. The sequences were automatically adjusted to the walking rate of the patient, permitting a stride time of up to seven seconds. The stimulation was triggered by two heel-switches.

The stride analyzer enabled us to record the following gait measurements during stimulation without any additional equipment: number of steps, mean stride time, temporal symmetry, and mean heel-on times with their standard deviations for both legs. The temporal symmetry of gait was calculated as the ratio of the heel-on time of the left leg to the heel-on time of the right leg. All gait measurements were derived and calculated from the signals from both heel-switches. These data are scattered with some random error around the correct value of the gait variable because of unreliable and inaccurate heel contact and push-off of the foot, especially on the affected extremity. When averaged, however, the data provide very reliable values of the gait variables, with accuracy in a range of [+ or -] 2%. A more detailed description of the stimulator was published elsewhere. [11]

At the beginning and at the end of the therapy period, the patients' gait was recorded on videotape for visual documentation and measured with the ground reaction measuring system for purposes of graphic representation of the ground reaction force and the spatial distribution of its point of action. [12,13] The measuring system includes five sizes of leather shoes with eight or nine force transducers inlayed in each sole. [14] The shoes were connected by cable and through amplifiers to a HP-2100 computer (*1) where the data were recorded and later processed. The accuracy of the system is within a range of [+ or -] 3%.

We used 5- X 10-cm silicon conductive rubber electrodes with karaya pads for stimulation of larger muscles and 5- X 5-cm electrodes for smaller muscles. We used 2.5-cm gauze button electrodes for peroneal nerve stimulation.

Subjects

Subjects were selected on the basis of neurological, internist, physiatric, and psychological examinations. The level and nature of the lesion were considered together with compensation of the cardiovascular system, passive range of motion in lower extremity joints, proprioception, response of individual muscle groups to electrical stimulation, intactness of the skin, and peripheral circulation.

In addition to general tests, more specific tests included a manual muscle test, a motor function test, a clinical kinesiological analysis of gait, and a test related to daily activities. In addition to the patient's informed consent, an adequate psychosocial condition, communicativeness, and motivation were also required. Patients with contraindications (eg, demand pacemaker, pains in the joints, dizziness, changes on skin, rejection of stimulation) were excluded from the study. All tests and examinations were performed to determine whether the patient could be included in the gait therapy with multichannel FES without risk. These data also provided the necessary information about the patient's condition and abilities and pointed to the deficiencies on which therapy should focus.

Twenty hemiplegic patients with severe motor deficiencies (5 female, 15 male), aged 14 to 74 years (X = 48, s = 16), participated in the study. Sixteen subjects were hemiplegic secondary to stroke, and 4 were hemiplegic secondary to head injury. Seven subjects were right hemiplegic, and 13 were left hemiplegic. At the beginning of the FES therapy program, most of the subjects could walk only leaning heavily on the therapist. Four subjects were able to walk with minor assistance and support from the therapist. None of the subjects could walk independently. All subjects were instructed in the use of a crutch. No mechanical bracing was used during the testing sessions. Patients in the early to later phases of rehabilitation after the onset of injury were candidates for multichannel FES. The treatment program started from 1.5 to 72 months post-onset (X = 9.5, s = 16.5). The duration of the therapy period varied according to the progress of the subjects, and lasted from 7 to 21 days (X = 14, s = 3.5).

Treatment

Trains of 0- to 120-V stimulation pulses of 30-Hz frequency and 0.2-msec pulse duration were applied in most cases to the common peroneal nerve for ankle dorsiflexion, to the soleus muscle for plantar flexion, to the quadriceps femoris muscle for knee extension, to the hamstring muscles for knee flexion, to the gluteus maximus muscle for hip extension, and to the triceps brachii muscle for reciprocal arm swing during the swing phase of the ipsilateral leg. [15] In subjects with bilateral impairment, the common peroneal nerve, the quadriceps femoris muscle, or the hamstring muscles of the contralateral leg were stimulated instead of the triceps brachii or soleus muscles. Muscles on the contralateral side were selected for stimulation according to the kinesiological deficiencies of the subject. In one case, the rhomboideus minor muscle was stimulated to correct the kyphotic posture of the subject. Exact stimulation sites were determined by cyclic stimulation. With the subject in the seated or in the prone position, the electrodes were shifted along each muscle selected for stimulation until an optimal functional response was obtained. When the sites were determined, they were marked on the skin with nonconductive permanent ink. Amplitude was raised until the functional response was satisfactory, or up to the pain threshold if the contraction was not satisfactory. In some cases, where muscles were stimulated on the contralateral side, amplitudes were set below the contraction level for providing sensory information as to when the stimulated muscle should be activated.

A stimulation sequence was determined for each subject, starting with an initial pattern and modifying it during the first couple of stimulation sessions until an optimal gait was achieved. The subjects walked on a 90-m walkway. They started with the support of a therapist covering a shorter distance, repeating it after a rest period. The initial distance depended on the subject's ability to avoid overexertion, or it was determined by the subject's physician. In the course of treatment, the distance was gradually increased. The subjects, however, were instructed not to ambulate more than 600 m per session because the purpose of the treatment was to correct the gait pattern, not to increase the distance.

Each subject participated in one therapy session per day, five times a week. No therapy sessions were conducted on Saturdays and Sundays so that the subjects could rest or go home for a weekend. The duration of therapy varied according to the subject's progress in gait. Previous studies demonstrated a high correlation between the gait variables and the clinical gait improvement of the patient. [16-19] Thus, the gait variables measured by the stimulator, which were available immediately after the session, permitted a quantitative follow-up and analysis of the subject's gait, and consequently, simultaneous planning and modification of the course of therapy. The therapy period was to be as short as possible, but long enough to enable the subject to start walking independently or reach the necessary degree of ambulation to be able to follow the standard gait therapy. The therapist's (NG) visual assessment of the subject's gait and the pattern of change in the measured gait variables were relevant in determining whether the subject's gait was improving. As long as the subject's gait improved, the therapy continued. If the subject's gait did not continue to improve for a few days, the therapy was stopped. Throughout their treatment with multichannel FES, the subjects continued with the prescribed rehabilitation program and chemotherapy, except that the exercises were less intensive to avoid excessive fatigue.

Measurements

The gait variables were measured each time the subject walked with FES. The heel-switches provided the necessary data for the stride analyzer. No additional equipment was needed. The distance that the subject walked was also measured so that the mean stride length and the mean velocity could be calculated. The mean stride length was calculated as the ratio of the distance to number of steps, and the mean velocity was calculated as the ratio of mean stride length to mean stride time. These variables, as well as those measured by the stride analyzer, were available immediately after the session and entered on a special form.

For additional quantification of the subject's gait, the ground reaction measuring system was used. The measurements were performed in the kinesiological laboratory. The gait of every subject was recorded at the beginning of the therapy period before multichannel FES was applied. At the end of the therapy period, the subject's gait was measured again without and with stimulation. First, the gait without FES was measured. The subject did not receive any FES in this measurement session to preclude any carry-over effect. After this measurement session, multichannel FES was applied and the subject walked for a few minutes. His gait was then remeasured with FES. Gait measurements were performed while the subject walked on a 20-m walkway and repeated three times with rests between repetitions.

All data were gathered and stored by the computer and later averaged over the steps. A couple of first and last steps were not included in these statistics. These data were then used to plot the mean ground reaction force and trajectory of its point of application, with their standard deviations, over the stance phase.

Results

Figure 2 shows the mean and standard deviation of stride time, stride length, and velocity of gait at the beginning and at the end of the therapy period for our group of 20 subjects. All gait variables were measured over a few hundred steps, which helped to ensure that the data accurately represented performance. Mean stride time decreased an average of 20% (s = 12%) by the end of the therapy period in all except one subject. In that particular subject, the longer mean stride time was due to considerably longer steps with the application of FES, which was also an improvement. The mean gait velocity increased an average of 61.6% (s = 47.5%), and the mean stride length an average of 46.3% (s = 32.5%). The data collected before and after the therapy period were compared by a t test, which showed a highly significant difference (p [is less than] .005) in all three variables. These data were obtained during the application of FES. Measurements performed with the ground reaction force measuring shoes in the gait laboratory also showed changes in gait without FES after the completion of therapy. The mean stride time was considerably shorter in all subjects, gait was more regular, and the load on the impaired leg was greater and lasted longer. The load on the crutch, if used, was decreased, and the trajectory of the point of action was closer to normal as compared with that at the beginning of the therapy period. Three-point gait (crutch-impaired leg-healthy leg) spontaneously changed into two-point gait (crutch and impaired leg-healthy leg) in all subjects. Endurance increased rapidly in all subjects and was evidenced by the progression in the number of steps they were able to perform in each session throughout the therapy period. A considerable improvement of posture was also observed throughout the therapy period. Some degree of gait improvement was observed in all subjects.

Representative Case

This section is a detailed description of the treatment response of one subject who is representative of the whole group. Subject "MD" (aged 45 years) developed right hemiplegia after sustaining a head injury in a car accident. He took 136 steps with the help of his physical therapist during the first FES session. He achieved about 500 steps after five days of therapy. He maintained this number of steps in subsequent sessions and concluded his 12-day therapy period with a left crutch and some minor assistance from the therapist. As shown in Figure 3, his mean stride time decreased from 3.36 sec to 2.09 sec; his mean heel-on times for both feet decreased by about 40%; and his relative dispersion, calculated as the ratio of the standard deviation to the mean, was constantly about 0.2 (20%) for both legs. As shown in Figure 4, the symmetry of gait in the same subject varied between 0.71 and 1.46, the mean velocity increased from 0.12 to 0.41 m/sec, and the mean stride length increased from 0.41 to 0.86 m.

Figures 5 to 7 show the results of the subject's ground reaction force measurements. The upper diagram shows the mean vertical force component with its standard deviation and body weight of the subject versus the mean stride time. The lower two diagrams show the spatial distribution of the force point of action on the shoe soles with standard deviations in the X and Y directions represented by rectangles at preselected time intervals.

A comparison of the forces during the gait of the subject without FES at the beginning and at the end of the therapy period (Figs. 5 and 6, respectively) revealed considerable improvement. Mean stride time decreased from 5.24 sec (s = 1.34) to 3.58 sec (s = 0.42). Mean velocity and stride length increased from 0.08 m/sec and 0.43 m, respectively, at the beginning of the therapy period to 0.15 m/sec and 0.57 m, respectively, at the end of the therapy period. The stance time for the subject's impaired leg increased from 45% to 65% of the stride time, and the load on this leg increased by 25% because the subject walked without the support of his therapist. The load on the impaired leg at the end of therapy period was still less than the load on the unimpaired leg, perhaps partly because the subject was using a crutch; however, the stance times for both legs were almost equal. The slope of the ground reaction force showed that the subject was slowly shifting body weight from the unimpaired to the impaired leg, possibly because he lacked confidence in its ability to provide positive support. The force distribution diagrams showed that the subject continued to concentrate his body weight on the middle lateral edge of his right foot during the whole stance phase, with a slight shift toward the toes at the end of the therapy period. That is, the subject was stepping with a flatfoot because he was not able to perform a dorsiflexion of the foot during the swing phase. The trajectory of the force point of action in the left leg changed more than in the right leg. At the beginning of the therapy period, it started 80 mm from the heel and ended 70 mm from the distal toes, which meant that the loading and push-off were performed with a flatfoot. At the end of the therapy period, we observed a considerable improvement in the loading and less improvement in the push-off.

Gait with FES in the same subject showed greater improvement at the end of the therapy period (Fig. 7). The mean stride time was 2.3 sec (s = 0.09). The stance time of the impaired leg was still 20% below the stride time, but the force pattern was closer to normal. The subject shifted his body weight onto his impaired leg as fast as onto his umimpaired leg because he felt more secure. The force amplitude remained lower in the impaired leg than in the unimpaired leg because he was using a crutch. Spatial distributions of the point of action showed greater improvement, especially on the affected side. With heel contact on both feet and the distribution from the center of the soles toward the toes, the force trajectories approached a normal pattern, whereas the standard deviations in the impaired left leg remained higher than in a healthy person.

Discussion

Previous studies focused on patients with less severe motor disabilities who had developed a gait capacity prior to the beginning of the therapy period. [7-9,20] They showed how a pathological gait pattern could be corrected through therapy with multichannel FES. The purposes of our approach were to initiate a new, correct gait pattern and to include ambulation in patients who could not walk at all or who could walk only with major help of a therapist. The data provided by the stride analyzer and the ground reaction measuring system, as well as observations of the subjects' gait, suggest that multichannel FES may be an appropriate treatment. In the rehabilitation of patients with stroke or head trauma, therapists are often faced with the problem of the patient's difficulty in comprehending or following verbal instructions. A likely advantage of the therapeutic approach reported in this article is that the information about correct gait is mostly conveyed through the patient's sensory system and only partly in the form of verbal instructions. The sensation induced by FES provides the information as to when a given muscle should be contracted during the gait cycle.

In discussing the effects of FES, the terms "long-term" and "short-term" are generally used. [7,8] In this particular study, neither of the terms is suitable because the gait of all subjects improved significantly, resulting in a partly or completely independent gait. Such independence in gait enables such patients to participate in conventional gait rehabilitation programs, which has not been possible before. The duration of these effects depends on the motivation of the patient and on the environment. If the patient continues to walk, the effects will be permanent; if not, the effects will probably diminish, and the patient will likely require the use of a wheelchair.

Physical therapists also are often faced with the problem of motivation of the patient. Cooperation is often difficult to induce, especially in exercises that are repetitious and that are considered boring by the patient. These types of exercises form the basis of classical gait training. In this study, the subjects mostly skipped over these exercises and started immediately with gait training. In almost all subjects, we observed an increase of motivation for gait rehabilitation after the first few days of therapy. They demonstrated a rapid improvement in their gait and a large increase in their endurance in the first few days of therapy, and each day they doubled the number of steps that they were able to perform. They tended to walk more and more to achieve independent gait instead of returning to the "safety" and convenience of a wheelchair, which in our experience is the usual tendency of such patients.

Regardless of the objective effectiveness of this method, physical therapists should understand its limitations and bear in mind that the reported results were achieved in combination with the available conventional rehabilitation methods. Use of our method should be considered only within the framework of a complex rehabilitation program for hemiplegic patients.

Conclusions

The noninvasive surface approach restricts multichannel FES to the clinical environment and requires skilled personnel, increasing the cost of treatment and reducing the number of patients who can be treated. Use of the multichannel FES approach during the first couple of weeks of therapy can foster independent gait in patients with severe motor disability, enabling them to continue with less demanding treatment and permitting the use of single- or dual-channel electrical stimulation for correction of gait deformities. [20]

Most importantly, patients who had not been able to walk, had not dared to walk for lack of self-confidence, or for various reasons had not been included in the gait training program were able to walk again and live a more independent life after a two- to three-week therapy period. With such recovery, the reduction of rehabilitation costs might be expected. A detailed comparative study is needed to determine the real value of this therapeutic approach as compared with conventional methods of gait rehabilitation.

(*1) Model HP-2100, Hewlett-Packard Co, 1501 Page Mill Rd. Palo Alto, CA 94304.

References

[1] Vodovnik L, Dimitrijevic MR, Prevec T, et al: Electronic walking aids for patients with peroneal palsy. In: Proceedings of the European Symposium on Medical Electronics. Brighton, England, September 28-October 1, 1965, vol 5, pp 58-61

[2] Stanic U, Trnkoczy A: Closed-loop positioning of hemiplegic patient's joint by means of functional electrical stimulation. IEEE Trans Biomed Eng 21:365-370, 1974

[3] Kralj A, Trnkoczy A, Acimovic R: Improvement of locomotion in hemiplegic patients with multichannel electrical stimulation. In: Human Locomotor Engineering, London, England, The Institution of Mechanical Engineers, 1974, pp 45-50

[4] Strojnik P, Kralj A, Ursic I: Programmed six-channel electrical stimulator for complex stimulation of leg muscles during walking. IEEE Trans Biomed Eng 26:112-116, 1979

[5] Stanic U, Trnkoczy A, Kljajic M, et al: Optimal stimulating sequences to normalize hemiplegics' gait. In: Proceedings of the Third Conference on Bioengineering. Budapest, Hungary, September 16-21, 1974, pp 252-256

[6] Trnkoczy A, Stanic U, Malezic M: Present state and prospects in the design of multichannel FES stimulators for gait correction in paretic patients. TIT Journal of Life Sciences 8: 17-27, 1978

[7] Malezic M, Krajnik J, Stanic U, et al: Long-term effects of multichannel stimulation on pathological gait. In: Proceedings of the Seventh International Symposium on External Control of Human Extremities. Dubrovnik, Yugoslavia, September 7-12, 1981, pp 409-419

[8] Malezic M, Stanic U, Kljajic M, et al: Multichannel electrical stimulation of gait in motor-disabled patients. Orthopedics 7:1187-1195, 1984

[9] Bogataj U, Kljajic M, Stanic U, et al: Gait pattern behaviour of hemiplegic patients under the influence of a six-channel microprocessor stimulator in a real environment. In: Proceedings of the Second International Conference on Rehabilitation Engineering. Ottawa, Canada, June 17-22, 1984, pp 529-530

[10] Kelih B, Malezic M, Gros N, et al: Evaluation of gait during therapy with six-channel electrical stimulation. In: Proceedings of the Fourteenth International Conference on Medical and Biological Engineering and the Seventh International Conference on Medical Physics. Espoo, Finland, August 11-16, 1989 vol 9, no 23, pp 424-425

[11] Pirnat P, Trnkoczy A: Further technical improvements of multichannel FES using microprocessor control. In: Proceedings of Seventh International Symposium on External Control of Human Extremeties. Dubrovnik, Yugoslavia, September 7-12, 1981, pp 317-326

[12] Lamoreaux LW: Kinematic measurements in the study of human walking. Bulletin of Prosthetic Research 10:3-84, 1971

[14] Kljajic M, Krajnik J: The use of ground reaction measuring shoes in gait evaluation. Clin Phys Physiol Meas 8:133-142, 1987

[14] Kljajic M, Krajnik J, Trnkoczy A: Determination of ground reaction and its distribution on the foot by measuring shoes. In: Digest of the International Conference on Medical Biological Engineering. Jerusalem, Israel, 1979, vol 40, no 1

[15] Rebersek S, Gros N, Valencic V: Functional electrical stimulation of arm swing during walking of hemiplegic patients. In: Proceedings of the World Congress on Physics and Biomedical Engineering. Hamburg, Germany, September 5-11, 1982, vol 12, no 35

[16] Brandstater M, deBruin H, Gowland C, et al: Hemiplegic gait: Analysis of temporal variables. Arch Phys Med Rehabil 64:583-587, 1983

[17] Holden MK, Gill KM, Magliozzi MR, et al: Clinical gait assessment in the neurologically impaired: Reliability and meaningfulness. Phys Ther 64:35-40, 1984

[18] Holden MK, Gill KM, Magliozzi MR: Gait assessment for neurologically impaired patients: Standards for outcome assessment. Phys Ther 66:1530-1539, 1986

[19] Mizrahi J, Susan Z, Heller L, et al: Variation on time-distance parameters of the stride as related to clinical gait improvements in hemiplegics. Scand J Rehabil Med 14:133-140, 1982

[20] Malezic M, Krajnik J, Stanic U, et al: Optimization of number of channels for electrical stimulation of pathological gait. In: Proceedings of the World Congress on Physics and Biomedical Engineering. Hamburg, Germany, September 5-11, 1982, vol 12, no 32

U Bogataj, MSc, is Research Engineer, Department of Biocybernetics, Automation, and Robotics, Jozef Stefan Institute, Jamova 39, 61111 Ljubljana, Yugoslavia. Address correspondence to Mr Bogataj.

N Gros, RPT, is Research Physiotherapist and Instructor, Rehabilitation Institute, Linhartova 52, 61111 Ljubljana, Yugoslavia.

M Malezic, BSc, is Research Engineer, Department of Biocybernetics, Automation, and Robotics, Jozef Stefan Institute.

B Kelih, BSc, is Research Engineer, Department of Biocybernetics, Automation, and Robotics, Jozef Stefan Institute.

M Kljajic, PhD, is Research Engineer, Professor, and Head of Laboratory of Biocybernetics, Department of Biocybernetics, Automation, and Robotics, Jozef Stefan Institute.

R Acimovic, MD, is Co-Director, Rehabilitation Institute.

This work was supported by the VL Smith Foundation for Restorative Neurology, Houston, TX; by Research Grant 23-P-59231/F of the National Institute of Handicapped Research, US Department of Education, Washington, DC; and by Research Grant C2-0123/106 of the Research Community of Solvenia, Ljubljana, Yugoslavia.

This article was submitted April 13, 1987; was with the authors for revision for 62 weeks; and was accepted December 1, 1988.