Use of Visual Feedback in Retraining Balance Following Acute Stroke.

To function in daily life, an individual must be able to maintain and adopt various postures, react to external disturbances, and use automatic postural responses that precede voluntary movements. Following a stroke, some or all of these tasks generally become more difficult. Increased sway during quiet standing,[1] uneven weight distribution with increased weight bearing on the unaffected limb,[2] decreased weight-shifting ability in stance,[3] and abnormalities in postural responses[4,5] have been documented. A major focus of rehabilitation programs, therefore, is to improve balance and optimize function and mobility

The ability to balance requires that the body's center of gravity (COG) lie over the base of support.[6] If individuals are provided with accurate visual representation of their CoG position, some authors[7] believe that motor behaviors can be improved. The Balance Master(*) is a commercially available computerized balance assessment and training system that provides the user with visual information about the position of the CoG within predefined (theoretical) limits of stability. By shifting the body weight and CoG over the base of support, the user can track the movement of the CoG on the computer screen. The visual feedback is supposed to be used to match and recalibrate proprioceptive sensory information or input that may be impaired due, for example, to stroke.[7] The theory behind such an approach is that improve CoG control should translate into gains in function. People who were better abel to shift their CoG at least 6 months following their strokes also performed well on activity-based measures such as the Berg Balance Scale.[8] Although it cannot be inferred from these findings that improved CoG control led to the gains in Berg Balance Scale scores, it is generally accepted that the ability to balance underlies the performance of most physical activities.[3,9]

Visual feedback related to weight distribution has been shown to be an efficacious method to gain symmetrical stance following stroke.[10,11] Winstein et al[11] reported that stroke survivors who were provided with visual information about their relative weight distribution through paretic and nonparetic limbs had better standing symmetry than those who received conventional physical therapy (exercises and routine standing balance and weight-shifting training). Sackley and Lincoln[10] extended these findings, demonstrating that improved stance symmetry was associated with superior ability to perform functional tasks. Shumway-Cook et al[1] showed that visual feedback of center-of-pressure position reduced asymmetrical standing more effectively than therapies designed to provide tactile and verbal cues regarding pastural symmetry. The total sway area during standing, however, was similar between groups. In combination, the above studies, which were designed to explore the efficacy of visual feedback training following stroke, have provided clear evidence that abilities specific to the training are enhanced.[1,10,11] Whether such training affords additional benefit in terms of function or the ability to perform everyday tasks remains inconclusive.

The purpose of our study was to explore the relative effectiveness of providing visual feedback of the CoG; position and conventional physical therapy, both offered in addition to physical therapy and occupational therapy (regular therapy) provided 2 hours a day to people with stroke admitted to a rehabilitation unit. A group of patients receiving only regular therapy served as a control group. Preliminary findings have been reported previously.[12]

Method

Subjects

Inpatients with hemiparesis secondary to a first stroke who had been admitted to a stroke unit for rehabilitation were screened for possible inclusion in the study. Individuals were referred if they were within 4 months of their stroke, could stand unassisted for 1 minute, and were in need of balance training according to the judgment of the senior physical therapist, who had more than 10 years of experience in the stroke unit. This individual was unaware of the prospective group allocation. Subjects were subsequently excluded for any of the following reasons: they were medically unstable as determined by examination, history, or medical records; they had a disorder prior to their stroke that affected their balance; they could not commit themselves to a minimum participatory involvement of 3 weeks; they had prior experience with the Balance Master (potential confound of learning); they were unable to see the computer screen at a distance of 60 cm; they scored less than 25 out of 36 on the Motor Free Visual Perceptual Test[13]; or they achieved a score below 22 out of 28 on the Folstein Modified Mini-mental State Examination.[14]

Fifty-four individuals were eligible to participate and provided their informed consent. All subjects were scored on the Clinical Outcome Variables Scale (COVS), which evaluates 13 mobility items (4 items involve ambulation, 2 items relate to arm function, and the remainder address transfers and ability to change positions).[15] Each item was scored on a scale of 1 (fully dependent [or unable]) to 7 (normal), and the scores were totaled (maximum total score = 91 points). The scale has demonstrated high interrater reliability (intraclass correlation coefficient [ICC] = .97).[16] This score characterized the participants' functional status at the time of admission to the study.

Of the eligible participants, 8 subjects withdrew from the study (4 subjects developed medical complications during the study period, 2 subjects failed to complete all of the testing, and 2 subjects did not want to continue in the study).

Balance Training (Phase 1)

Subjects were assigned to either the visual feedback group or the conventional therapy group using block randomization in order to ensure equivalency in group size. In view of the inclusion and exclusion criteria, which eliminated those individuals with more severe deficits, no effort was made to stratify the groups on the basis of any characteristic or measured variable. Sixteen subjects per group were sought based on sample size calculations for an effect size of 6 points on the Berg Balance Scale and assuming a variance of 6.2,[8] a significance level of .05, and power at 0.80.[12] In addition to their regular therapy program, based on a neurodevelopmental approach[17,18] and incorporating everyday activities,[19] these subjects received an additional 30 minutes of balance training. The additional training was provided 5 days a week for 3 to 8 weeks (depending on the length of the inpatient stay) by an experienced physical therapist (practicing for 17 years) who was not part of the regular rehabilitation staff. This individual scheduled and delivered interventions to both treatment groups, taking care to ensure that all sessions were 30 minutes in length.

Visual feedback training involved the use of the Balance Master and accompanying software (version 3.4). The Balance Master consists of 2 forceplates positioned side by side (each measuring 23 x 46 cm) with transducers mounted along the anterior-posterior center line of each plate. The output is digitized, and the software provides the user with feedback about the CoG location (adjusting for subject height [ie, 0.55 x height[20]]) in the form of a cursor displayed on a monitor. For the purpose of training stance symmetry, the forceplates served as weigh scales, and bars reflecting the weight transferred through each leg were displayed on the monitor. Symmetrical weight distribution was presumed when the bars on the computer screen were the same height. Tactile and verbal cues were provided as necessary to ensure proper alignment and stability of the hips, knees, and trunk (erect posture with no observable leaning to one side). The task was progressed through the addition of an upper-extremity activity or introducing trunk rotation. To increase weight bearing on the affected limb, subjects were instructed to shift their weight until the bars on the computer corresponded to a preset target.

To encourage weight shifting, the visual feedback group moved their CoG and observed the corresponding cursor movement (representing CoG position) on the computer screen. Targets positioned on the screen were used to encourage weight shifting as subjects attempted to move the cursor in a desired direction toward the targets. Increasing the distance between the targets, decreasing the time required to move between the targets, adding an upper-extremity activity, or altering the foot position increased the task difficulty. The positioning of the targets was set relative to the theoretical limits of stability (LOS), which is based on the assumption that individuals could shift their CoG 6.25 degrees anteriorly, 4.45 degrees posteriorly, and 8 degrees to each side from a resting position.[20] Initially, the targets were set at positions approximating 30% of the LOS; however, the targets were moved closer toward the LOS as individuals consistently achieved the training goal. In this manner, the task remained challenging. Additionally, rhythmic weight shifting was encouraged by having subjects shift their weight forward and backward or from side to side while keeping pace with a moving target. Software provided with the Balance Master was used for the training protocols.

The conventional therapy training protocol was an extension of the regular rehabilitation program. Symmetrical weight distribution was encouraged through verbal and tactile cues and was made more difficult by the addition of arm activities or actions requiring trunk rotation. Stools of various heights were used to support the nonparetic lower limb and to increase weight bearing on the affected side. In an effort to improve rhythmic weight-shifting ability, subjects practiced shifting their weight in forward and backward directions and side to side while performing reaching tasks such as dropping beanbags through a hoop.

In all cases, programs were set up on an individual basis. The amount of time spent on items varied according to an individual's ability and tolerance as judged by the physical therapist. Competence in a certain skill was not required prior to moving on to the next item.

Outcome Measures

Postural sway measurements were obtained using the Balance Master as a force platform. We did not check the measurement characteristics of the device and based our use of the device on the manufacturer's claims. Subjects stepped onto the platform, their feet were positioned in accordance with the manufacturer's guidelines,[20] and they were instructed to stand as still as possible with arms at their sides looking straight ahead. Postural sway was measured with eyes open and then with eyes closed over a 20-second period. The average sway area was expressed as a percentage of the theoretical limits of stability, as established by the manufacturer.[20,21] A total of 3 trials for each condition (alternating between eyes open and eyes closed) were performed, and average values were calculated. These measures have moderate reliability as demonstrated in subjects with chronic hemiparesis (6 months or more following a stroke).[8]

Measurements were also obtained with 3 activity-based measures of balance--the Berg Balance Scale,[9] the Timed "Up & Go" Test,[22] and gait speed. The Berg Balance Scale is a 14-item task-oriented test that has been used to identify and evaluate balance impairment in people with hemiplegia[23] and that has been reported to be responsive to clinically meaningful changes.[24] When a subject was unable to independently complete a test item, he or she was given 3 attempts and the score on the best attempt was recorded. A total score for all items was determined for each subject (maximum score = 56 points), as this measure has been shown to have excellent intrarater reliability (ICC = .99).[9]

For the Timed "Up & Go" Test, subjects were seated in a chair with armrests and then instructed to stand (using the armrests, if desired) and walk as quickly and as safely as possible for a distance of 3 m. Subjects then turned around, returned to the chair, and sat down. The time from the point at which their spine left the back of the chair until they returned to that same position was recorded using a stopwatch. A practice trial was provided and followed by 3 test trials. The average time of the test trials was calculated. High intrarater (ICC = .99) and interrater (ICC = .99) reliability have been demonstrated using this measure.[22]

Gait speed was determined by having subjects walk as quickly and as safely as possible along a 15-m walkway. The time to traverse the middle 10 m was measured in order to exclude acceleration and deceleration. Following a practice trial, subjects completed 3 trials and the mean speed was determined. Excellent test-retest reliability (ICC = .96) has been reported in people having had strokes beyond 6 months prior to testing.[8]

All outcome measurements were collected from all subjects at baseline (entry into the study), after completing balance training (at discharge from the stroke unit or after 8 weeks, whichever came first), and 1 month later (follow-up). The therapist who administered the out-come tests was aware of individuals' group allocation, which was considered acceptable, given the nature of the outcome measures or the standardized method of scoring.

Phase 2

The second phase of the study began after completion of recruitment for phase 1 (16 subjects per group) when all new admissions to the stroke unit meeting the inclusion and exclusion criteria were assigned to the control group. This group received a regular therapy program, as did subjects in phase 1, which included 2-hour daily sessions, 5 days per week, with physical therapists and occupational therapists. The therapists were the same as those who participated in phase 1 of the study, and they were unaware of the results of the phase 1 testing. The primary aim of the treatment was to maximize independence and improve function. Subjects in the control group did not receive additional balance training. All outcome measurements were obtained from all control subjects in the same manner as described above.

Data Analysis

Data were pooled across subjects according to group (control, visual feedback, and conventional therapy). An analysis of variance for repeated measures with one between-subject factor (3 groups) and one within-subject factor (3 testing sessions) was performed for each of the outcome measures. If an effect of time (3 testing sessions) was observed, post hoc multiple paired-sample t tests were used to determine between which testing sessions the differences lay. For all analyses, a significance level with an alpha less than .05 was adopted.

Results

Forty-six subjects with hemiparesis who had a mean age ([+ or -] 1 standard deviation [SD]) of 64.5 [+ or -] 12.2 years completed the study. The 3 groups were similar in terms of age, side of stroke (with 36%-44% of each group having right hemiparesis), and COVS score, as the average score for each group fell within 2 points of each other (Tab. 1). The average number of days between the stroke and admission to the stroke unit was comparable for the visual feedback, conventional therapy, and control groups (about 18, 20, and 21 days, respectively). The time elapsed between the stroke and the start of the study was also consistent across groups (F = 0.599, df = 2, P = .554) and ranged from 8 to 80 days. There were no differences in the time interval from the initial baseline measurements and the second testing session, which approximated 37 to 39 days for all groups (F = 0.250, df = 2, P = .780). These data and the characteristics associated with each group are summarized in Table 1.

Table 1. Group Characteristics

                               Control Group
                                 (n = 14)

Characteristic           [bar]X    SD    Range

Age (y)                   65.8     9.9   47-85
Female                     5
Right hemiparesis          5
Clinical Outcome
  Variables Scale[15]
  score                   59.7    11.7   42-77
Stroke to stroke unit
  admission (d)           21.1     9.3   11-42
Stroke to study
  admission (d)           42.0    13.4   26-60
Time between testing
  sessions 1 and 2 (d)    36.6    14.2   19-63

                          Visual Feedback Group
                                 (n = 16)

Characteristic           [bar]X    SD    Range

Age (y)                   65.4    13.8   33-83
Female                     4
Right hemiparesis          7
Clinical Outcome
  Variables Scale[15]
  score                   60.7     9.5   39-80
Stroke to stroke unit
  admission (d)           17.7     9.8    8-39
Stroke to study
  admission (d)           40.9    19.5   15-80
Time between testing
  sessions 1 and 2 (d)    39.3    12.0   16-58

                          Conventional Therapy
                             Group (n = 16)

Characteristic           [bar]X    SD    Range   P(a)

Age (y)                   62.4    13.3   30-77   .723
Female                     8
Right hemiparesis          7
Clinical Outcome
  Variables Scale[15]
  score                   61.4    10.2   47-82   .904
Stroke to stroke unit
  admission (d)           19.6    14.6    3-58   .717
Stroke to study
  admission (d)           35.1    22.3    8-79   .554
Time between testing
  sessions 1 and 2 (d)    36.7    10.7   15-58   .780

(a) Probability value associated with the significance of the between-groups F ratio (df=2) associated with a one-way analysis of variance.

The baseline measurements associated with pastural and activity-based balance indicated a similarity in ability across all 3 groups (F = 0.254-0.960; df = 2, P [is greater than] .391). Groups were comparable in terms of the amount of sway they exhibited, with that associated with eyes closed (1.12%, 1.25%, and 0.84% of LOS for the control, visual feedback, and conventional therapy groups, respectively) being almost twice that of eyes open. Mean Berg Balance Scale scores approximated 36 for all groups, and walking speed was slow, averaging between 0.32 and 0.48 m/s for all groups (Tab. 2).

Pooling data across groups revealed marked reductions in eyes-open and eyes-closed pastural sway (F = 17.64, df = 42, P [is less than] .001; F = 17.37, df = 42, P [is less than] .001), almost a doubling in walking speed (F = 29.0, df = 42, P [is less than] .001) paired with a reduction in Timed "Up & Go" Test scores of about 50% (F = 28.9, df = 42, P [is less than] .001), and approximately a 12-point increase in Berg Balance Scale scores (F = 17.4, df = 42, P [is less than] .001). There were, however, no group differences for any outcome measure over time, indicating equivalency in balance performance despite differences in intervention. The mean number of training sessions received by the visual feedback group and the conventional therapy group were 22.0 [+ or -] 6.0 and 23.2 [+ or -] 6.8 sessions, respectively (t = .515, df = 30, P = .610). The data relating to all outcome measures are presented in Table 2.

Table 2. Mean Outcome Measures ([+ or -] 1 SD) and Range as a Function of Group and Time(a)

                                       Baseline

Outcome               Control       VF          CT
Variable               Group       Group       Group        All

Postural sway-eyes
open (% limits
of stability)
  [bar]X             0.57        0.43        0.49        0.49
  SD                 0.46        0.35        0.44        0.41
  Range              0.07-1.35   0.09-1.38   0.08-1.76   0.07-1.76

Postural sway-eyes
closed (% limits
of stability)
  [bar]X             1.12        1.25        0.84        1.07
  SD                 0.77        0.92        0.74        0.81
  Range              0.22-2.67   0.23-3.34   0.09-2.89   0.09-3.34

Berg Balance Scale
score (points)
  [bar]X             36.1        35.9        36.9        36.3
  SD                 13.4         7.5        11.0        10.5
  Range                18-53       22-46       18-55       18-55

Gait speed (m/s)
  [bar]X             0.48        0.32        0.47        0.42
  SD                 0.38        0.28        0.42        0.37
  Range              0.08-1.05   0.02-1.05   0.12-1.54   0.02-1.54

Timed "Up & Go"
Test (s)
  [bar]X             52.8        54.2        45.8        50.9
  SD                 40.1        34.4        26.8        33.7
  Range                12-118      16-152      16-96       12-152

                                       Discharge

Outcome               Control       VF          CT
Variable               Group       Group       Group        All

Postural sway-eyes
open (% limits
of stability)
  [bar]X             0.32        0.21        0.25        0.26
  SD                 0.24        0.11        0.27        0.22
  Range              0.08-0.81   0.08-0.45   0.04-1.10   0.04-1.10

Postural sway-eyes
closed (% limits
of stability)
  [bar]X             0.78        0.57        0.47        0.60
  SD                 0.53        0.42        0.49        0.48
  Range              0.13-2.10   0.21-1.79   0.08-1.65   0.08-2.10

Berg Balance Scale
score (points)
  [bar]X             46.5        46.6        48.8        47.3
  SD                  8.9         6.0         6.2         7.0
  Range                29-56       32-55       36-55       29-56

Gait speed (m/s)
  [bar]X             0.74        0.57        0.89        0.73
  SD                 0.53        0.34        0.65        0.53
  Range              0.14-1.85   0.14-1.18   0.20-2.67   0.14-2.67

Timed "Up & Go"
Test (s)
  [bar]X             29.3        33.4        21.3        28.0
  SD                 21.6        20.3        12.8        18.8
  Range                 8-80       12-73        6-49        6-80

                                       Follow-up

Outcome               Control       VF          CT
Variable               Group       Group       Group        All

Postural sway-eyes
open (% limits
of stability)
  [bar]X             0.23        0.15        0.19        0.19
  SD                 0.14        0.06        0.15        0.12
  Range              0.06-0.48   0.04-0.26   0.05-0.56   0.04-0.56

Postural sway-eyes
closed (% limits
of stability)
  [bar]X             0.61        0.37        0.42        0.46
  SD                 0.36        0.31        0.47        0.39
  Range              0.08-1.05   0.09-1.33   0.06-1.81   0.06-1.81

Berg Balance Scale
score (points)
  [bar]X             47.1        47.9        50.9        48.7
  SD                  9.3         5.1         4.1         6.5
  Range                27-56       42-55       46-56       27-56

Gait speed (m/s)
  [bar]X             0.82        0.63        0.93        0.79
  SD                 0.60        0.36        0.58        0.53
  Range              0.14-1.78   0.18-1.39   0.28-2.27   0.14-2.27

Timed "Up & Go"
Test (s)
  [bar]X             28.8        28.2        17.8        24.9
  SD                 25.2        20.2         9.8        19.6
  Range                 8-92       10-88        6-37        6-92

Outcome
Variable             P(b)

Postural sway-eyes
open (% limits
of stability)
  [bar]X             .435
  SD
  Range

Postural sway-eyes
closed (% limits
of stability)
  [bar]X             .387
  SD
  Range

Berg Balance Scale
score (points)
  [bar]X             .639
  SD
  Range

Gait speed (m/s)
  [bar]X             .268
  SD
  Range

Timed "Up & Go"
Test (s)
  [bar]X             .411
  SD
  Range

(a) VF = visual feedback, CT = conventional therapy.

(b) Probability value associated with significance of the between-groups F ratio (df=2) associated with a 2-way analysis of variance (group and time).

Due to the variance in test scores across subjects and across time, the data were also examined in terms of change scores. The Figure illustrates the mean change scores for the respective outcome measures from baseline to discharge and from discharge to follow-up. The relative improvements in both postural and activity-based measures of balance were greatest in the period from baseline to discharge. Again, there were no between-group differences with respect to any outcome measure.

[Figure ILLUSTRATION OMITTED]

Discussion

The people with stroke who participated in this study demonstrated marked improvements in their balance abilities over time. The introduction of additional therapy incorporating either visual feedback of the CoG or conventional physical therapy strategies afforded no additional benefit to the individuals who participated in the study.

Examining the changes in postural balance ability over time revealed marked reductions in postural sway from subjects' initial baseline levels. In comparison with normative sway values of 0.16 and 0.34 (eyes open and closed, respectively) for individuals with no known health problems aged 60 to 75 years,[25] the participants in our study initially averaged about 3 times the amount of sway. It has been postulated that large oscillatory movements at the ankles may compensate for sensory deficiencies by augmenting proprioceptive feedback.[26] If this were so, it would be appealing to conclude that the normalization of sway at follow-up may be indicative of improved sensation or a recalibration of the contributions made by the relevant sensory systems, as suggested by Moore and Woollacott.[7] In the absence of sensory testing, however, the mechanisms underlying the improvements observed in our study cannot be determined.

In terms of activity-based balance measures, all 3 groups improved in gait speed, Timed "Up & Go" Test scores, and Berg Balance Scale scores over time. Cunningham et al[27] reported a natural walking speed of 1.1 m/s in subjects with no known health problems. Goldie et al[2] documented a mean gait speed of 0.45 m/s in a group of 42 individuals with acute strokes walking at their comfortable speed. In our study, the maximal safe gait speed averaged 0.42 m/s at baseline and nearly doubled by follow-up to 0.79 m/s. This change exceeds the 42% gain observed by Goldie et al[2] after 8 weeks of gait training.

The initial slowness in gait speed followed by marked increases may relate to the relative ability of people with stroke who are early in their rehabilitation to vary walking speeds. Turnbull et al[28] investigated the range of available walking speeds in people with chronic hemiparesis. They reported that unlike age- and sex-matched control subjects who demonstrated 5 distinct walking speeds, those with hemiparesis had only 2 walking speeds--natural and fast. In the early stages of recovery, comfortable versus fast walking speeds may not be discernable. Later, however, a clear distinction may be apparent. If this is the case, measuring maximal safe walking speed as done in our study would produce larger gains than those detected by tests recording comfortable gait speed (eg, Goldie et al[2]).

The improved performance on the Timed "Up & Go" Test was not surprising to us in light of the improvements observed in walking speed. The time taken to complete the task at follow-up was one half of the time taken during the initial measurement session. This finding suggests that, in addition to walking faster, subjects were better able to transfer from a chair and change direction while walking. Pairing these findings with the evidence of gains in Berg Balance Scale scores provides support that activity-based balance performance improved over time. The extent of the gains was most dramatic in the period between initial testing and discharge rather than from discharge to follow-up. This finding was not due to relative differences in the time elapsed between consecutive testing dates, as there was approximately 4 weeks between baseline and discharge testing and between discharge and follow-up testing. The early gains may relate to the daily rehabilitation all subjects received during this time and the natural time course of recovery, which is known to slow down over time.[29]

An unexpected finding of our study was that there were no differences in any of the outcome measures despite different interventions. The provision of feedback relating to the CoG position was intended to recalibrate the postural control system. Researchers[30] have suggested that visual information can compensate for sensorimotor loss and, with training, subjects can assimilate the information, thus establishing a central motor program such that the external feedback would no longer be required. Other authors have reported gains in stance symmetry in subjects with hemiparesis who trained with either visual feedback of the position of the center of pressure[1] or weight distribution[10,11] over those who received conventional training. In cases where the feedback training and testing protocols were similar,[1] the ability to distinguish between performance and learning was limited. Winstein et al[11] indicated that improved stance symmetry was not associated with a reduction in asymmetrical movement patterns in gait, suggesting that skill transference to more complex motor activities is limited. Sackley and Lincoln[10] reported that the initial differential benefit of visual feedback training was lost when subjects were followed up after 8 weeks, suggesting that such training failed to enhance learning or skill retention.

There are several explanations as to why an immediate differential treatment effect was not evident in the visual feedback group. In our study, visual feedback was constant and immediate throughout the training period. Winstein and Schmidt[31] reported that limiting the relative frequency of providing knowledge of results to young adults with no known health problems during the practice sessions improved learning, as evidenced by higher scores on retention tests compared with subjects receiving feedback after every trial. Without externally supplied knowledge of results, subjects apparently were obliged to use intrinsic information produced by the movement itself. Delaying feedback and encouraging subjects to estimate their performance level during the delay interval has also been found to enhance learning.[32] Instantaneous visual feedback, however, may be detrimental to learning, as subjects fail to attend to intrinsic information in favor of the more concrete external information, although it may well contribute to improved error reduction during task performance.[32-34]

Practice is believed to be essential for effective learning of complex tasks,[35,36] and the training activities should resemble real-life tasks as much as possible in order to maximize training effects,[19,37] particularly when component skills are highly interdependent.[38] There is a question as to why subjects receiving additional conventional balance training failed to outperform other groups, given that functional, everyday activities were used to promote stability during weight shifting. We contend that, in the early stages of rehabilitation, it may be extremely difficult to detect performance differences attributable to the specifics of supplemental interventions.

Most spontaneous recovery following stroke occurs during the first 3 to 6 months.[29,39] During this period, patients receiving specialized care in, for example, a stroke or rehabilitation unit do better in terms of functional ability than do people with stroke who are treated on general medical wards.[40-42] These findings suggest that, although natural or spontaneous recovery may account for some of the observed improvements, the nature of the treatment also appears to play a role. In our study, all subjects were admitted to the stroke unit and received approximately 120 minutes of daily physical therapy and occupational therapy. All treatment received by the patients was tracked in 5-minute units and recorded in the patients' files, enabling us to be confident that participants in this study received equivalent therapy time.

All subjects made improvements in postural and activity-based balance ability, with no added benefit associated with additional treatment. It is conceivable that the regular therapy sessions alone sufficed to enable patients to maximize their potential. Limited energy levels may have reduced the efficacy of additional interventions, because fatigue is more prevalent following stroke relative to age-matched peers without health problems.[43] Alternatively, it may be that a certain threshold level of performance needs to be achieved before the intricacies of an intervention become important. In sports, expert performers are better able to interpret and utilize skill-related information, including visually displayed information, than novices.[44,45] Recently Winstein et al[46] confirmed that the execution and control of motor tasks are adversely affected following stroke, but not the learning of the skills. It would be of interest to the rehabilitation community to further explore learning strategies following stroke.

Conclusions

During the early rehabilitation following stroke, subjects who were admitted to a stroke unit and who received daily rehabilitation therapy for up to 8 weeks exhibited improvements in postural sway and activity-based measures of balance over time. The gains were greatest during the inpatient period. Additional daily treatment sessions using either visual feedback of the CoG position or conventional balance training afforded no added benefit to the participants. Standard treatment paired with spontaneous recovery may have been sufficient to maximize the patients' potential. Whether specialized intervention strategies such as visual feedback training are differentially effective in later stages of recovery is not known.

References

[1] Shumway-Cook A, Anson D, Hailer S. Postural sway biofeedback: its effect on reestablishing stance stability in hemiplegic patients. Arch Phys Med Rehabil. 1988;69:395-400.

[2] Goldie PA, Matyas TA, Evans OM, et al. Maximum voluntary weight bearing by the affected and unaffected legs in standing following stroke. Clin Biomech. 1996;11:333-342.

[3] Dettmann MA, Linder MT, Sepic SB. Relationships among walking performance, postural stability, and functional assessments of the hemiplegic patient. Am J Phys Med. 1987;66:77-90.

[4] Badke MB, Duncan PW. Patterns of rapid motor responses during postural adjustments when standing in healthy subjects and hemiplegic patients. Phys Ther. 1983;63:13-20.

[5] Horak FB, Esselman P, Anderson ME, Lynch MK. The effects of movement velocity, mass displaced, and task certainty on associated postural adjustments made by normal and hemiplegic individuals. J Neurol Neurosurg Psychiatry. 1984;47:1020-1028.

[6] Nashner LM. Sensory, neuromuscular, and biomechanical contributions to human balance. In: Duncan PW, ed. Balance: Proceedings of the APTA Forum, Nashville, Tennessee, June 13-15, 1989. 1989:5-12.

[7] Moore S, Woollacott MN. The use of biofeedback devices to improve postural stability. Physical Therapy Practice. 1993;2:1-19.

[8] Liston RA, Brouwer BJ. Reliability and validity of measures obtained from stroke patients using the Balance Master. Arch Phys Med Rehabil. 1996;77:425-430.

[9] Berg KO, Wood-Dauphinee SL, Williams JI, Gayton D. Measuring balance in the elderly: preliminary development of an instrument. Physiotherapy Canada. 1989;41:304-311.

[10] Sackley CM, Lincoln NB. Single blind randomized controlled trial of visual feedback after stroke: effects on stance symmetry and function. Disabil Rehabil. 1997;19:536-546.

[11] Winstein CJ, Gardner ER, McNeal DR, et al. Standing balance training: effect on balance and locomotion in hemiparetic adults. Arch Phys Med Rehabil. 1989;70:755-762.

[12] Grant T, Brouwer BJ, Culham EG, Vandervoort A. Balance retraining following acute stroke: a comparison of two methods. Can J Rehabil. 1997;11:69-73.

[13] Colarusso RP, Hammill DD. Motor Free Visual Perceptual Test. Novato, Calif: Academic Therapy Publications; 1972.

[14] Folstein MF, Folstein SE, McHugh PR. "Mini-mental state": a practical method for grading the cognitive state of patients for the clinician. J Psychiatr Res. 1975;12:189-198.

[15] Seaby L, Torrance G. Reliability of a physiotherapy functional assessment used in a rehabilitation setting. Physiotherapy Canada. 1989;41:264-270.

[16] Seaby L, Torrance G. Measurement properties of the physiotherapy Clinical Outcome Variables (COVs): a mobility functional assessment. Physiotherapy Canada. 1994;46:JS76.

[17] Bobath B. Adult Hemiplegia: Evaluation and Treatment. 2nd ed. London, England: William Heinemann Medical Books Ltd; 1978.

[18] Davies PM. Steps to Follow: A Guide to the Treatment of Adult Hemiplegia. Berlin, Germany: Springer-Verlag; 1985.

[19] Carr JH, Shepherd RB. A Motor Relearning Programme for Stroke. 2nd ed. Rockville, Md: Aspen; 1987.

[20] Balance Master Operator's Manual. Clackamas, Ore: NeuroCom International; 1993.

[21] Brouwer BJ, Culham EG, Liston RA, Grant T. Normal variability of postural measures: implications for the reliability of relative balance performance outcomes. Scand J Rehabil Med. 1998;30:131-137.

[22] Podsiadlo D, Richardson S. The timed "Up and Go": a test of basic functional mobility for frail elderly persons. J Am Geriatr Soc. 1991;39:142-148.

[23] Stevenson TJ, Garland SJ. Standing balance during internally produced perturbations in subjects with hemiplegia: validation of the balance scale. Arch Phys Med Rehabil. 1996;77:656-662.

[24] Wood-Dauphinee S, Berg KO, Bravo G, Williams J. The balance scale: responsiveness to clinically meaningful changes. Can J Rehabil. 1997;10:35-50.

[25] Hammon R, Longridge NS, Mekjavic I, Dickinson J. Effect of age and training schedules on balance improvement exercises using visual biofeedback. J Otolaryngol. 1995;24:221-229.

[26] Patla AE, Frank JS, Winter DA. Assessment of balance control in the elderly: major issues. Physiotherapy Canada. 1990;42:89-97.

[27] Cunningham DA, Rechnitzer PA, Donner AP. Exercise training and the speed of self selected walking pace in men at retirement. Can J Aging. 1986;5:19-25.

[28] Turnbull GI, Charteris J, Wall JC. A comparison of the range of walking speeds between normal and hemiplegic subjects. Scand J Rehabil Med. 1995;27:175-182.

[29] Ashburn A. Physical recovery following stroke. Physiotherapy. 1997;83:480-490.

[30] Mulder T, Hulstyn W. Sensory feedback therapy and theoretical knowledge of motor control and learning. Am J Phys Med. 1984;63:226-244.

[31] Winstein CJ, Schmidt RA. Reduced frequency of knowledge of results enhances motor skill learning. J Exp Psychol Learn Mem Cogn. 1990;16:677-691.

[32] Swinnen SP, Schmidt RA, Nicholson DE, Shapiro DC. Information feedback for skill acquisition: instantaneous knowledge of results degrades learning. J Exp Psychol Learn Mem Cogn. 1990;16:706-716.

[33] Gentile AM. Skill acquisition: action and neuromotor processes. In: Carr JH, Shepherd RB, Gordon J, et al, eds. Movement Science: Foundations for Physical Therapy Rehabilitation. Rockville, Md: Aspen; 1987:93-154.

[34] Schmidt RA, Bjork RA. New conceptualizations of practice: common principles in three paradigms suggest new concepts for training. Psychological Science. 1992;3:207-217.

[35] Schmidt RA. Motor Control and Learning. 2nd ed. Champaign, Ill: Human Kinetics Inc; 1988.

[36] Swanson LR, Sandford JA. Motor learning concepts applied to rehabilitation. In: Pickles B, Compton A, Cott C, et al, eds. Physiotherapy With Older People. London, England: WB Saunders Co Ltd; 1995:107-116.

[37] Ma HI, Trombly CA, Robinson-Podolski C. The effect of context on skill acquisition and transfer. Am J Occup Ther. 1999;53:138-144.

[38] Naylor JC, Briggs GE. Effects of task complexity and task organization on the relative efficiency of part and whole training methods. J Exp Psychol. 1963;65:217-244.

[39] Lehmann JF, DeLateur BJ, Fowler RS Jr, et al. Stroke: does rehabilitation affect outcome? Arch Phys Med Rehabil. 1975;56:375-382.

[40] Indredavik B, Bakke F, Solberg R, et al. Benefit of a stroke unit: a randomized controlled trial. Stroke. 1991;22:1026-1031.

[41] Kalra L. The influence of stroke unit rehabilitation on functional recovery from stroke. Stroke. 1994;25:821-825.

[42] Sivenius J, Pyorala K, Heinonen OP, et al. The significance of intensity of rehabilitation of stroke: a controlled trial. Stroke. 1985;16:928-931.

[43] Ingles JL, Eskes GA, Philips SJ. Fatigue after stroke. Arch Phys Med Rehabil. 1999;80:173-178.

[44] Abernathy B, Russel DG. The relationship between expertise and visual search strategy in a raquet sport. Human Movement Science. 1987;6:283-319.

[45] Borgeaud P, Abernathy B. Skilled perception in volleyball defense. J Sport Psychol. 1987;9:400-406.

[46] Winstein CJ, Merians AS, Sullivan KJ. Motor learning after unilateral brain damage. Neuropsychologia. 1999;37:975-987.

(*) NeuroCom International, 9570 SE Lawnfield Rd, Clackamas, OR 97015.

C Walker, MSc, is Lecturer, School of Rehabilitation Therapy, Queen's University, Kingston, Ontario, Canada.

BJ Brouwer, PhD, is Associate Professor and Chair, Graduate Program, School of Rehabilitation Therapy, Queen's University, LD Acton Bldg, 13 George St, Kingston, Ontario, Canada K7L 3N6 ([e-mail]). Address all correspondence to Dr Brouwer.

EG Culham, PhD, is Associate Professor and Chair, Physical Therapy Program, School of Rehabilitation Therapy, Queen's University.

All authors provided writing and data analysis. Dr Brouwer and Dr Culham provided concept/research design and fund procurement. Dr Brouwer and Ms Walker provided project management, and Ms Walker provided data collection. The authors acknowledge the support of Cally Martin (provision of subjects), Pat Cross (provision of facilities/equipment and institutional liaisons), and Mary Jo Demers (provision of subjects) in the Physiotherapy Department, St Mary's of the Lake Hospital.

The ethics review boards of the Faculty of Health Sciences, Queen's University, and St Mary's of the Lake Hospital approved the study protocol.

This study was supported by grant NA2819 from the Heart and Stroke Foundation of Ontario.

This article was submitted August 26, 1999, and was accepted May 30, 2000.