Secondary Logo

Journal Logo

Feasibility of Adapted Aerobic Cycle Ergometry Tasks to Encourage Paretic Limb Use After Stroke: A Case Series

Sibley, Kathryn M. MSc; Tang, Ada PT, MSc; Brooks, Dina PT, PhD; Brown, David A. PT, PhD; McIlroy, William E. PhD

Journal of Neurologic Physical Therapy: June 2008 - Volume 32 - Issue 2 - p 80-87
doi: 10.1097/NPT.0b013e318176b466
Case Report
Free
SDC

Cardiorespiratory fitness, along with sensorimotor recovery, is important for optimal function after stroke. Development of exercises that simultaneously address aerobic training and increase paretic limb involvement may improve outcomes and maximize productivity of therapy sessions. This case series assessed the feasibility of and characterized the cardiorespiratory and sensorimotor demands of adapted aerobic cycle ergometer activities hypothesized to increase paretic limb use. Mechanically loaded and electromyographic (EMG) feedback pedaling were compared to traditional pedaling in three poststroke case studies and a healthy control group. Submaximal oxygen uptake (Vo2), heart rate, perceived rate of exertion (RPE), and EMG of four leg muscles were assessed. Mechanically loaded ergometry increased RPE and altered muscle activity in healthy participants, while participants with stroke did not consistently increase paretic limb activation. EMG feedback pedaling increased target limb activity in healthy participants and decreased nonparetic activity in stroke participants. This paper highlights the challenges involved in adapting training tasks for individuals who are not able to walk at training intensities. Further work is necessary to refine adapted tasks for optimal effectiveness, and consideration of additional methods that permit differential interlimb loading may have additional value.

Institute of Medical Science (K.M.S., A.T., D.B., W.E.M.), Department of Physical Therapy (K.M.S., A.T., D.B.), University of Toronto, Toronto, Ontario; Canada; Department of Kinesiology (W.E.M.), University of Waterloo, Waterloo, Ontario, Canada; Toronto Rehabilitation Institute (K.M.S., A.T., D.B., W.E.M.), Toronto, Ontario, Canada; Department of Physical Therapy and Human Movement Sciences (D.A.B.), Northwestern University, Chicago, Illinois.

Address correspondence to: William E. McIlroy, E-mail: wmcilroy@uwaterloo.ca

Back to Top | Article Outline

INTRODUCTION

Cardiorespiratory fitness is increasingly recognized as an important focus to maximize functional recovery after stroke.1 However, the nature of sensorimotor impairments associated with stroke presents challenges for selecting appropriate aerobic training modalities. Although a subset of stroke survivors may be capable of walking at sufficient speeds to achieve meaningful aerobic gains, paresis and postural dyscontrol make walking exercise difficult, if not impossible, for many stroke survivors. Aerobic walking exercise offers both cardiorespiratory and sensorimotor benefits; therefore, alternate exercise modalities that offer comparable aerobic challenge and neuromuscular demand need to be identified for individuals with a range of abilities post-stroke.

Cycle ergometry is a commonly used form of aerobic training, and the ability to exercise while seated is advantageous for clinical populations, particularly in the early stages after injury when overground or treadmill walking may be especially difficult. In contrast to the standard upright position, semirecumbent (SR) ergometry offers additional support for individuals with compromised trunk stability. SR ergometry has been used as the training or testing modality for several protocols involving stroke patients.2,3

Although standard ergometry is effective for aerobic training in healthy individuals, it features characteristics that may limit potential usefulness among stroke survivors, particularly those with dyscontrol of the paretic lower limb. For example, when pedaling after stroke, the nonparetic limb is favored for force production, with some phase-inappropriate activity and reduced mechanical work production from the paretic limb.4 Therefore, although individuals with stroke can achieve cardiorespiratory benefit from cycle ergometry,5 their control strategies are compensatory and potentially maladaptive. Potential cardiorespiratory benefits, requiring extended periods of rhythmic lower limb movements, may be countered by reinforcement of asymmetric sensorimotor control that could limit recovery of functional walking.

Specific tasks designed to engage the paretic limb to improve recovery of motor control combined with aerobic activity may result in more efficient use of training sessions. Forced or constrained use of the limb to optimize poststroke motor recovery has been successfully applied to upper limb retraining after stroke.6 The application of forced-use principles is more difficult in the lower limbs due to the inherent coupling of the limbs in walking (and other locomotor movements). However, encouraging use of the paretic limb during locomotor aerobic training offers the potential to improve motor recovery as well. This is due in part to the large number of movement repetitions that occur during training relative to the number of movement repetitions typical of everyday activities.7 The challenge is, first, to establish whether adapted encouraged-use aerobic training models are feasible and, second, to determine the effectiveness of such an intervention. Doing so would provide the opportunity to enhance both aerobic capacity and sensorimotor control by using adapted training programs for individuals in the early stages of stroke recovery.

The purpose of the present work was to establish a potential exercise modality using SR ergometry, which in future studies could be applied to optimize sensorimotor and cardiorespiratory recovery after stroke. Three case studies examined the feasibility of performing two encouraged-use pedaling tasks in persons post-stroke whose low gait speed and mild to severe asymmetry8 were sufficient to exclude them from an aerobic walking program, but who would benefit from increased control of the paretic limb. The results from these case studies were contrasted with cardiorespiratory and sensorimotor adaptations from a control group of persons without neurological impairment. The present study provided important preliminary steps for (1) examining the potential utility of these exercise interventions and (2) providing insight into the short-term physiologic changes associated with encouraged use of the paretic limb. Tasks were based on SR ergometry with variations to encourage use of a specific limb: (1) mechanically manipulating the challenge of the task to increase paretic limb involvement and (2) electromyographic (EMG) feedback (FB) to provide the participant with ongoing knowledge of task performance. We hypothesized that the encouraged-use paradigms would increase appropriate muscular contributions from the paretic limb and that the increased challenge associated with engaging the paretic limb would be reflected by increased cardiorespiratory demand and perceived effort.

Back to Top | Article Outline

METHODS

The study procedures were followed in accordance with institutional guidelines and were approved by the local university and hospital research ethics committees. Informed written consent was obtained from all study participants.

Back to Top | Article Outline

Participants

Three individuals with stroke were recruited from a larger trial examining the effects of exercise on cardiorespiratory fitness after stroke (Tables 1 and 2). Participants were referred by a neurologist and underwent medical screening for appropriateness to exercise by a physiatrist. Participants were all at least two years post-stroke and living in the community. These individuals were recruited for the case series as they each presented with significant residual impairment and walking dysfunction.

TABLE 1

TABLE 1

TABLE 2

TABLE 2

In order to compare and characterize the expected performance variations produced by the pedaling adaptations, a convenience sample of 10 healthy individuals (five men, five women) were also recruited (Table 2). Participants verified by a self-report questionnaire that they were safe to exercise and were free of any musculoskeletal, neurological, or cardiovascular conditions that would affect pedaling performance.

Back to Top | Article Outline

Protocol

All pedaling tasks were performed on a commercially available SR ergometer. Back angle was fixed at 11 degrees from vertical, and the top-dead-center position (TDC) occurred at 45 degrees posterior to the vertical. During an initial testing session, a maximal exercise test was conducted to establish a standardized exercise intensity for evaluating the encouraged-use pedaling tasks. Due to travel considerations, participants with stroke performed the exercise test and task variations on the same day, with a minimum of two hours rest between sessions. Healthy participants performed the two testing sessions on separate days.

Participants pedaled at a predetermined cadence (50 rpm for patients with stroke, 70 rpm for healthy individuals) as work rate was increased every two minutes until the participant indicated that he or she could not pedal any longer or the investigator terminated the test according to the American College of Sports Medicine guidelines.11 Details of this protocol are published elsewhere.3 During the exercise test, gas exchange was recorded and analyzed, and blood pressure, heart rate (HR), electrocardiogram, and rate of perceived exertion (RPE) were monitored (modified Borg 0-10 Scale12). Peak oxygen uptake (Vo2peak) and power output (POpeak) were recorded.

Two encouraged-use pedaling tasks, mechanically limb-loaded pedaling and EMG FB–monitored pedaling were compared with standard SR ergometry (the control task). Pedaling intensity was set at 50% of POpeak from the initial session for all conditions, and participants pedaled at 50 rpm. Work rate and cadence were displayed continuously on the ergometer to verify appropriate pedaling intensity, and cadence was maintained within five rpm of the target. Participants pedaled for four minutes in each task condition, and five minutes of rest were provided between each task. In the control task, participants pedaled with their preferred strategy, without any FB or knowledge of performance. The control task was performed both at the beginning and end of the testing session to evaluate for signs of fatigue. Comparisons of submaximal HR, Vo2, RPE, and EMG profiles did not demonstrate any systematic significant differences between repetitions, and the two trials of the control task were averaged together. The two adapted task variations were presented in a counterbalanced order among participants, and participants were randomly assigned to task order.

The first pedaling variation, limb-loaded ergometry (LLE), is task-oriented pedaling that requires coordinated weight transfer between limbs.13 A spring-loaded seat pan requires constant tonic extensor activation during pedaling to control seat position and avoid sliding too far forward. Springs can be loaded in 4.5-kg increments to a maximum of 45 kg. Previous work has evaluated short periods of LLE (20 revolutions) in stroke patients13; however, neither coordination nor cardiorespiratory effort has been assessed for this paradigm. Based on manufacturer-recommended guidelines for minimal loading for therapeutic benefits (the only guidelines available for setting load intensity), springs were loaded to 30% of body weight. Second, we evaluated pedaling with EMG FB. The purpose of this task was to evaluate whether participants could use visual EMG FB to modify pedaling control strategies, specifically to selectively increase the involvement of a target muscle without disrupting the overall movement. Healthy controls were instructed to pedal primarily with their self-selected dominant limb, and participants with stroke were instructed to focus on the paretic limb while they received a visual display of the magnitude of vastus lateralis (VL) muscle activity produced by each leg. EMG FB was generated by calculating the total integrated activity when the quadriceps muscles are highly active during the power phase of the cycle14 (from approximately 10% of the cycle [36 degrees] to 50% of the cycle [180 degrees]) in the previous 30 seconds. For each limb, total integrated VL muscle activity was displayed pictorially to the participant in a bar graph format and was updated every 30 seconds. Participants were encouraged to increase the activity in the target limb, which was confirmed by the visual display.

Back to Top | Article Outline

Data Collection

Surface EMG recordings were measured at 512 Hz using surface electrodes positioned two centimeters apart (center to center) along the belly of four leg muscles bilaterally: VL, biceps femoris long head (BF), tibialis anterior (TA), and medial gastrocnemius (MG).15 Vo2 and HR quantified cardiorespiratory demand, assessed continuously and averaged at 30-second intervals. Participants rated their overall whole-body and leg-specific RPE.12

Back to Top | Article Outline

Data Processing

Offline EMG processing was performed with Labview 7.0 software (National Instruments, Austin, TX). EMG signals were band pass filtered (20–150 Hz), full-wave rectified, and smoothed (low-pass 20 Hz second-order filter). Each revolution was referenced to the TDC position representing zero degrees and averaged to produce a mean activation profile. Onset of muscle activity was identified as the point in the cycle at which activity was greater than three standard deviations above the mean background activity within the cycle and that lasted longer than 20 milliseconds.16 Offset of muscle activity was defined as the point in the cycle at which activity fell within three standard deviations of the mean background activity of the revolution for longer than 20 milliseconds. Total integrated EMG (iEMG) was calculated as the area under the curve between the points of onset and offset. For each measure, in each adapted task, the percentage of change relative to the control task was calculated. This percentage of change score was calculated for both limbs in the FB task and averaged across legs for LLE. Vo2 and HR data were calculated over the last thirty seconds of each trial. Overall and leg RPE were assessed at the end of each trial.

Back to Top | Article Outline

Data Analysis

Descriptive analysis of feasibility of performance was performed for participants with stroke. Paired t tests evaluated cardiorespiratory parameters, perceived exertion measures, and EMG changes (by muscle) comparing the LLE and FB tasks with the control task in the healthy participants. Reported values are mean ± SD. Statistical significance was set at P < 0.01.

Back to Top | Article Outline

RESULTS

Healthy participants were able to perform the variations in the task conditions at the required loads and durations (Table 3). Mean EMG profiles for all tasks are shown in Figure 1. LLE was characterized by a nonsignificant increase in Vo2 as well as significant increases in HR, overall RPE, and leg RPE compared to the control task (P < 0.01). Both VL and MG muscles showed an earlier onset during LLE relative to the control task (mean VL shift: 32.3 ± 20.7 degrees, P = 0.0008; mean MG shift: 61.5 ± 48.9 degrees, P = 0.003), and VL also exhibited 46.4 ± 36.7% more activity than traditional SR ergometry (P = 0.003). The EMG FB task did not induce any significant changes in Vo2, HR, or overall RPE relative to the control task, although Leg RPE was significantly higher (P = 0.01). Total iEMG was 32.8 ± 15.2% higher in the targeted VL (P = 0.0002), while simultaneously reduced 27.2 ± 19.9% in the nontargeted VL (P = 0.003). TA of the targeted limb also showed a 59.5 ± 48.6% increase in total activity (P = 0.01).

TABLE 3

TABLE 3

FIGURE 1.

FIGURE 1.

Participants with stroke were moderately impaired (Chedoke-McMaster Stroke Assessment [CMSA] leg score = 3-6/7, foot score = 2-6/7), had low aerobic capacity (8.1–11.1 mL/kg · min), slow preferred walking speeds (37–64 cm/sec) and mild to severe asymmetry during gait.8 EMG patterns varied across individuals, although cycle-cycle variability within a task and an individual was comparable to the variability observed across the healthy participants. Common to the participants with stroke was the observation of asymmetric activation favoring the nonparetic limb as well as phase-inappropriate activity and/or tonic activation and lack of modulation (Fig. 2). Individual case profiles are detailed below, followed by a summary of all results.

FIGURE 2.

FIGURE 2.

Back to Top | Article Outline

Patient 1 (P1)

P1 was a 51-year-old woman, two years post-stroke. A CMSA leg score of 6/7 indicated near-normal movement patterns where only faster or more complex movements were difficult, but the foot score was only 3/7, indicating marked spasticity and movement primarily within the synergy patterns. Accordingly, her gait speed was only 37 cm/sec, and she demonstrated severe temporal asymmetry favoring the nonparetic limb. EMG profiles during the control task clearly demonstrate minimal contributions from the paretic VL. P1 was not able to perform LLE at the prescribed intensity of four springs, and accordingly they were reduced to three springs and then to two springs in order to complete four minutes of pedaling. Although pedaling was discontinuous, Vo2 was increased 29% compared to the standard SR task. TA activity was increased in both limbs, and there were no observable increases in paretic VL activity. The EMG FB task was completed with a high RPE (RPE = 7). Despite a lack of increase in paretic VL activity, nonparetic VL was reduced and pedaling was more efficient (Vo2 reduced 28% compared to control task).

Back to Top | Article Outline

Patient 2 (P2)

P2 (male, five years post-stroke, 71 years old) had the lowest Vo2peak (8.1 mL/kg/min) but the highest level of recovery overall (CMSA leg = 5/7, foot = 6/7) and was also the fastest walker among the three patients. This individual demonstrated a high level of tonic activation during the control task, particularly in the paretic BF and MG, as well as phase-inappropriate activity in the paretic TA and nonparetic BF. The prescribed spring load was six springs, but P2 was only able to perform two minutes of LLE at one third of the prescribed load. Vo2 demonstrated a 17% decrease relative to the control task. EMG patterns during LLE showed little variation from the control task. In contrast, in the FB task paretic VL activation was slightly increased, along with corresponding decreases in nonparetic VL activity. This was associated with a 9% increase in Vo2 compared to the control task.

Back to Top | Article Outline

Patient 3 (P3)

P3 (54-year-old man, four years post-stroke) was the most impaired of the patients (CMSA leg = 3/7, foot = 2/7, spasticity present and no voluntary movement), with a high degree of asymmetry and phase-inappropriate activity in the paretic BF during the control task. Like the other patients with stroke, P3 was not able to pedal at his prescribed spring load (six springs), which was subsequently reduced to four springs. Of the three patients studied, P3 was the only individual to demonstrate increases in both VL muscles during LLE (although activity was still asymmetric). This change was associated with a large increase in Vo2 (46%) compared to the standard SR task. P3 decreased nonparetic VL activity during the FB task, but did not increase paretic activation. There were no differences in Vo2 compared to the control task.

Back to Top | Article Outline

Summary

None of the poststroke participants were able to pedal at 30% of body weight during LLE, and even with the reduction in tonic load applied through the springs, all three participants had difficulty maintaining a continuous, even rhythm for the duration of the four-minute testing session.Both cardiorespiratory and sensorimotor responses to LLE varied both across individuals and in comparison to the average of the healthy participants. In contrast, the EMG FB task was better tolerated. Although cardiorespiratory responses varied between individuals, all three patients were able to decrease VL activity in the nonparetic limb. However, it is noteworthy that this was not necessarily accompanied by an increase in paretic leg VL activation.

Back to Top | Article Outline

DISCUSSION

The purpose of this study was to examine the sensorimotor and cardiorespiratory characteristics of two modified pedaling tasks to evaluate the feasibility of increasing paretic limb activity during adapted aerobic pedaling after stroke. While only a small number of individuals with stroke were evaluated in this initial case series, the three participants are representative of the types of individuals who are likely to benefit from such an adapted program. Specifically, this subset of the stroke population would be unable to participate in a traditional aerobic walking program. The findings from the case series highlight the challenges faced in identifying aerobic training activities that may also target the primary sensorimotor impairments caused by stroke. Responses from a healthy control group were important to highlight expected variations between tasks and inform selection of a potential task for study with individuals after stroke. A number of observations were particularly noteworthy. Healthy participants and participants with stroke did not show the same adaptations to the modified pedaling tasks. The hypotheses were supported in the healthy participants, such that both tasks successfully increased target muscle activation with limited effects on other muscles and accordingly increased cardiorespiratory output and perceived exertion. However, these adaptations did not translate to the participants with stroke.

Participants with stroke were not able to pedal at the same spring load as healthy participants during the LLE task. This is an issue of inappropriate exercise prescription and a reflection of a lack of any published, standardized guidelines for establishing exercise intensity in a new and previously untested paradigm. Accordingly, we adjusted the spring load for participants with stroke as necessary to allow them to complete the task at a lower intensity. Interestingly, even at reduced spring loads, sensorimotor and cardiorespiratory adaptations did not match those of the control group. This could be due to the differences in load between populations and/or due to the dyscontrol experienced by the individuals with stroke. Moreover, on the whole, these patients were unable to increase paretic limb activation during the LLE task. It appears that even at reduced spring loads, the LLE task was excessively challenging for these individuals. This issue may be addressed in future studies by identifying appropriate spring loads. It is also possible that limb-loaded ergometry may be more appropriate for higher functioning patients as an adjunct to walking exercise. While these case studies provide preliminary insights for guiding development of LLE prescription, clearly more study is required in a broader and larger sample to determine which subset of the stroke population may be most appropriate for limb-loaded training.

Similarly, pedaling with EMG FB induced different adaptations in each of the groups. Healthy participants successfully used EMG FB to increase VL activation with minimal compensations in other muscles. In contrast, although the EMG FB task was better tolerated than LLE task in participants with stroke, they were unsuccessful in increasing paretic limb involvement. One possibility for this is that FB may have been provided too infrequently for participants with stroke to sustain adaptations. FB was generated from cumulative EMG activity, updated every 30 seconds. While this time frame was selected to allow participants adequate time to process the information and make conscious adjustments in pedaling strategies and was effective in the healthy group, participants with stroke may have benefited from increased frequency of FB. Interestingly, although the participants with stroke did not increase paretic leg activity with EMG FB as hypothesized, nonparetic muscle activation was reduced. While the functional implications of such a change are not clear, it indicates that participants were attempting to perform the task. Like LLE, use of EMG FB warrants further study before being recommended for practice. That said, the EMG FB was well tolerated, and, in light of moderate successes with EMG FB in other paradigms, such as the improvement in gait quality with FB-based gait training,17,18 it may be more relevant than LLE for future application in stroke rehabilitation. However, such trials need to have sufficient power and standard assessments to appropriately assess the effectiveness of EMG FB.18

Our results highlight the challenges experienced when adapting traditional exercise models for enhanced training. The current emphasis on SR ergometry was due to (1) the availability of such equipment in clinical facilities and (2) the limited ability of many people with stroke to perform walking-based aerobic training. Selection of an appropriate adapted training activity may not necessarily require both limbs to work equally as long as the paretic limb participates. In fact, differential loading would likely be ideal, particularly if the degree of loading could be varied as the patients’ abilities improved. Both of the proposed models allow patient-specific needs to be addressed during training, such as modification limb loading or of the muscles targeted for FB.

We recognize that the small number of participants with stroke limit the conclusions that can be drawn from this study. However, our examination of feasibility of adapted training has revealed the need for further study of enhanced-use training tasks. For example, while the present encouraged-use strategies did not induce the desired changes in motor control, modifications to both adapted paradigms (such as decreased loading and more frequent FB) may yield more positive results. The current work was limited to evaluation over a short-term interval. The insights gained from the present study are warranted before embarking on future long-term training programs to explore the influence on learning and neurological change.

Back to Top | Article Outline

CONCLUSIONS

The development of adapted training modalities to address both sensorimotor and cardiorespiratory deficits after stroke is important for several reasons. First, many existing approaches do not encourage paretic limb use during aerobic training, and there is a lost opportunity to attempt to concurrently improve sensorimotor control. Furthermore, development of such activities could improve efficiency of rehabilitation sessions and may optimize functional recovery. This feasibility case series represents an important first step and highlights the challenges involved in adapting training activities. The tasks selected in the present study, which have the potential to generate the desired effects, require more refinement and study in a broader group of participants before being recommended on a larger scale. Alternative adapted training paradigms (such as recumbent stepping or force-based FB) also warrant attention.

Back to Top | Article Outline

ACKNOWLEDGMENTS

We thank Mark Bayley, MD, Sandra Black, MD, Valerie Closson, and John Esposito for assistance with medical screening, patient referral, recruitment, and data collection.

Back to Top | Article Outline

REFERENCES

1.Pang MY, Dawson AS, Gylfadottir S. The use of aerobic exercise training in improving aerobic capacity in individuals with stroke: a meta-analysis. Clin Rehabil. 2006;20:97–111.
2.Katz-Leurer M, Sender I, Keren O, Dvir Z. The influence of early cycling training on balance in stroke patients at the subacute stage. Results of a preliminary trial. Clin Rehabil. 2006;20:398–405.
3.Tang A, Sibley KM, Thomas SG, McIlroy WE, Brooks D. Maximal exercise test results in subacute stroke. Arch Phys Med Rehabil. 2006;87:1100–1105.
4.Brown DA, Kautz SA. Increased workload enhances force output during pedaling exercise in persons with poststroke hemiplegia. Stroke. 1998;29:598–606.
5.Potempa K, Lopez M, Braun LT, Szidon JP, Fogg L, Tincknell T. Physiological outcomes of aerobic exercise training in hemiparetic stroke patients. Stroke. 1995;26:101–105.
6.Taub E, Miller NE, Novack TA, et al. Technique to improve chronic motor deficit after stroke. Arch Phys Med Rehabil. 1993;74:347–354.
7.Macko RF, Ivey FM, Forrester LW, et al. Treadmill exercise rehabilitation improves ambulatory function and cardiovascular fitness in patients with chronic stroke: a randomized, controlled trial. Stroke. 2005;36:2206–2211.
8.Patterson KK, Parafianowicz I, Danells CJ, et al. Gait asymmetry in community ambulating stroke survivors. Arch Phys Med Rehabil. 2008;89:304–310.
9.Gowland C, Van Hullennaar S, Torresin W, et al. Chedoke-McMaster Stroke Assessment, Development, Validation and Administration Manual. Hamilton: Chedoke-McMaster Hospitals and McMaster University;1995:723–727.
    10.Gowland C, Straford P, Ward M, et al. Measuring physical impairment and disability with the Chedoke-McMaster Stroke Assessment. Stroke. 1993;24:58–63.
    11.American College of Sports Medicine. ACSM’s Guidelines for Exercise Testing and Prescription. 7th ed. Philadelphia: Lippincott Williams & Wilkins; 2006:78–80.
    12.Borg GAV. Psychophysical bases of perceived exertion. Med Sci Sports Exerc. 1982;14:377–381.
    13.Brown DA, Nagpal S, Chi S. Limb-loaded cycling program for locomotor intervention following stroke. Phys Ther. 2005;85:159–168.
    14.Neptune RR, Kautz SA, Hull ML. The effect of pedaling rate on coordination in cycling. J Biomech. 1997;30:1051–1058.
    15.Rainoldi A, Melchiorri G, Caruso I. A method for positioning electrodes during surface EMG recordings in lower limb muscles. J Neurosci Methods. 2004;134:37–43.
    16.Quant S, Adkin AL, Staines RW, Maki BE, McIlroy WE. The effect of a concurrent cognitive task on cortical potentials evoked by unpredictable balance perturbations. BMC Neurosci. 2004;5:18.
    17.Moreland JD, Thomson MA, Fuoco AR. Electromyographic biofeedback to improve lower extremity function after stroke: a meta-analysis. Arch Phys Med Rehabil. 1998;79:134–140.
    18.Woodford H, Price C. EMG biofeedback for the recovery of motor function after stroke. Cochrane Database Syst Rev. 2007(2):CD14004585.
    Keywords:

    paresis; cerebrovascular accident; cardiovascular deconditioning; exercise; rehabilitation

    © 2008 Neurology Section, APTA