Volume 87, Issue 5, Pages 656-660 (May 2006)

11 of 25

ABSTRACT

FULL TEXT

FULL-TEXT PDF (95 KB)

Psychometric Properties of a Modified Wolf Motor Function Test for People With Mild and Moderate Upper-Extremity Hemiparesis

Abstract

Whitall J, Savin DN, Harris-Love M, McCombe Waller S. Psychometric properties of a modified Wolf Motor Function Test for people with mild and moderate upper-extremity hemiparesis.

Objective

To test the necessity of videotaping, test-retest reliability, and item stability and validity of a modified Wolf Motor Function Test (WMFT) for people with mild and moderate chronic upper-extremity (UE) hemiparesis caused by stroke.

Design

Raters of videotape versus direct observation; test-retest reliability over 3 observations, item stability, and criterion validity with upper-extremity Fugl-Meyer Assessment (FMA) in the mildly and moderately impaired groups.

Setting

Academic research center.

Participants

Sixty-six subjects with chronic UE hemiparesis who participated in a large intervention study. Subjects were classified into mild and moderate groups for additional analyses.

Interventions

Not applicable.

Main Outcome Measures

Mean and median times of task completion, functional ability, and strength (weight to box) measures of the WMFT. FMA scores for validity assessment.

Results

In a subgroup of 10 subjects, the intraclass correlation coefficient (ICC) for videotape versus direct observation ranged from .96 to .99. For the whole group, test-retest reliability using ICC2,1 ranged from .97 to .99; stability of the test showed that administration 1 differed from administrations 2 and 3 but administrations 2 and 3 did not differ; item analysis showed that 4 of 17 items changed across time, and validity, using a correlation with UE FMA, ranged from .86 to .89. Separate mild- and moderate-group analyses were similar to whole-group results.

Conclusions

Videotaping the modified WMFT was not necessary for accurate scoring. The modified WMFT is reliable and valid as an outcome measure for people with chronic moderate and mild UE hemiparesis and is stable, but 1 repeat testing is recommended when practical.

Key Words: Cerebrovascular accident , Hemiparesis , Rehabilitation , Reliability and validity

Article Outline

• Abstract

• Methods

• Participants

• Test Administration

• Procedures

• Step 1: videotaping versus direct observation

• Step 2: test-retest reliability, item stability, and validity of the WMFT in subjects with chronic mild and moderate hemiparesis

• Results

• Step 1: Videotape Versus Observation

• Step 2: Test-Retest Reliability, Item Stability, and Criterion Validity

• Comparison of subgroups

• Discussion

• Conclusions

• Acknowledgment

• Appendix 1. Functional ability scale of the modified WMFT

• References

• Copyright

STROKE IS THE LEADING CAUSE of long-term disability in the United States, with more than 1.1 million adults reporting difficulty with activities of daily living and functional limitations as a result of stroke.1 In addition, an estimated 95% of stroke survivors have some resultant upper-extremity (UE) dysfunction.2 The importance of investigating novel rehabilitation techniques, particularly for UE dysfunction, is apparent. Equally important is the development of reliable, valid, and clinically useful tests of UE dysfunction. The recognition that improvement of function is more important than improvement of impairment measures alone, such as strength and range of motion, has resulted in the development of several tests designed to measure functional abilities.3, 4, 5, 6, 7 The Wolf Motor Function Test (WMFT),7 previously called the Emory Motor Test, is 1 such test that was initially developed to assess the impact of forced use on return of UE function in chronic stroke subjects. Subsequently it was modified and renamed,8 and it has been used as an outcome measure in several UE rehabilitation studies, most notably studies of constraint-induced movement therapy.9, 10, 11, 12, 13, 14, 15, 16, 17 The WMFT contains 3 parts: (1) timing, the speed at which functional tasks can be completed; (2) functional ability, the movement quality when completing the tasks; and (3) strength, the ability to lift against gravity. The purpose of the present study was to build on previous assessments7, 8, 18 of the WMFT’s reliability and validity, in which only subjects with mild hemiparesis were assessed.

Morris et al8 examined the reliability of the WMFT in assessing UE motor function in a group of 24 subjects with mild chronic hemiparesis. Interrater reliability for median performance times and mean functional ability both were greater than .93, and the agreement of the raters was also high, at .88 and .97. Test-retest reliability of 2 administrations separated by 2 weeks was high, at .90 (time) and .95 (function). However, some individual items had low correlations across administrations and scales (<.50).8 Wolf et al18 examined the WMFT using a group of 19 subjects with mild chronic UE hemiparesis who were age- and sex-matched with 19 subjects without impairment. Interrater reliability again was high, ranging from .97 to .99. In addition, construct validity, shown by the WMFT’s ability “to differentiate the more affected extremity and the less affected extremity from either extremity of subjects without impairment”18(p1637) and criterion validity, shown by a significant relation between UE Fugl-Meyer Assessment (FMA) and WMFT scores for the affected extremity, were supported.18 Test-retest reliability was not reported and only the timed—not the functional ability—portion of the WMFT was used for analysis.

These studies provide evidence that the WMFT has high interrater reliability and validity when used with subjects with mild impairments and when scored from videotape. Nevertheless, there remain some questions regarding test-retest reliability, the stability of the scales, and individual items including the strength tests. For example, using an earlier version of the test in a training study, Wolf et al7 described a training effect before scores become stable on the timed portion of the test. Moreover, further questions remain regarding the usefulness of WMFT as a research and clinical tool in the larger population of stroke survivors. In particular, we were interested in testing the WMFT for subjects with moderate-severity hemiparesis. This study addresses some of these issues. First, as a preliminary step, we investigated whether it was necessary to videotape test sessions for later review, because this is a burden on the laboratory researcher and is almost impossible within the time constraints of the clinical environment. If trained raters can make immediate timing and functional ability decisions, the clinical usefulness and efficiency of this tool will be greatly increased. Second, we examined test-retest reliability and stability over 3 trials, the stability of individual items, and the concurrent validity of the WMFT with the UE FMA on subjects with chronic mild- and moderate-severity hemiparesis.

Methods

Participants

All participants were part of a larger training intervention study, and reliability measurements were performed in the context of pretraining baseline testing. Participants were recruited through local hospitals and advertisements. Informed consent, approved by the joint Veterans Affairs–University of Maryland institutional review board, was obtained from all participants before inclusion in the study. Initial evaluations included a medical history, the Folstein Mini-Mental State Examination, and the National Institutes of Health Stroke Scale (NIHSS). All subjects met the following inclusion criteria: unilateral ischemic stroke, age 25 years or older, at least 6 months poststroke, completion of all conventional inpatient and outpatient therapy, and residual hemiparetic motor dysfunction—specifically, subjects had to demonstrate minimal antigravity shoulder movement (with the distal end moving at least 7.62cm), but active wrist and finger movement was not required. Exclusion criteria included the following: unstable angina, recent (<3mo) myocardial infarction, congestive heart failure (New York Heart Association class II), hemodynamically significant valvular dysfunction, recent hospitalization for severe disease or surgery, poorly controlled hypertension, significant orthopedic or chronic pain conditions limiting UE exercise, active neoplastic disease, pulmonary or renal disease, untreated major poststroke depression, dementia, severe receptive or global aphasia, and a nonstroke neuromuscular disorder restricting exercise. Inclusion and exclusion criteria were assessed first by telephone and subsequently by interview, with tests administered by a physician and a nurse or research administrator.

Test Administration

The WMFT was administered to subjects as part of a larger testing protocol. A dedicated table with marks indicating where objects were to be placed was used for test administration. The WMFT in its original form consisted of 21 tasks.7 We followed a more recent version, which consists of 17 tasks, 2 strength (1 grip, 1 lifting the maximum possible weight onto a box) and 15 functional, arranged from simple to complex.8 Examples of a simple task are subjects bringing the forearm to the table from the lap and extending the elbow. Intermediate tasks include reaching and retrieving a weight and lifting a pencil from the table. More complex tasks include turning a key in a lock and folding a towel. The 15 functional tasks were timed, with a maximum of 120 seconds allowed for each. Each task was described and demonstrated to subjects before they performed the task, and subjects were not allowed to practice the task during its description. The rater ensured that subjects understood the instructions for the task and told them to attempt the task as quickly as they could. The preparatory instruction given to each subject was, “Ready, set, go,” with timing commencing at the word “go.” If a subject was unable to complete a task in the allotted time, a performance time of 120 seconds was recorded.

For the functional ability scoring, we modified the original test instructions, which used a 6-point scoring scale, by telling subjects that they were not allowed to use the nonparetic arm to assist the paretic arm with any of the unilateral tasks. (One exception was that subjects were allowed to use the uninvolved arm during the bilateral task of folding a towel.) This resulted in a rescaled 5-point scoring scale (appendix 1). This modification of the original test was introduced after early pilot work showed that unless we made this stipulation, our moderately affected subjects, in particular, used the nonparetic arm to assist or to complete the task when they did not need to do so. The consequence of this adjustment meant that subjects were not assigned a functional ability score of 2. Therefore, we removed the original score of 2 and rescaled to produce the modified version. Although there are other criteria for a score of 2, taking out this score completely gave us a cleaner method of scoring. Another consequence of the modified instruction is that timing scores may be longer than those in other studies where subjects were permitted to use the nonparetic arm to assist.8 The consequence of our modified instruction is that direct comparison of absolute scores across studies is not possible for functional ability because of the rescaling. Comparison of the absolute performance times also requires caution but, to this point, few other studies have used the WMFT for subjects with moderate to severe impairment. Our modified instruction has ensured a measurement of paretic arm function alone without compensatory nonparetic assistance. In the sense that subjects with mild impairments are also likely to be using only the paretic arm, this study and those with mildly impaired subjects are comparable.

Procedures

Step 1: videotaping versus direct observation

To determine whether videotaping the test sessions is necessary, we assessed 10 subjects (6 women) with a mean age of 64±12 years (range, 39–81y), mean time poststroke of 78±64 months (range, 7–204mo), and median NIHSS score for each subject’s paretic UE of 1 (range, 0–3). Subjects were videotaped as they performed the WMFT. A trained tester assessed the subjects while they performed the test, used a stopwatch to obtain the time score, and dictated the time and functional ability score to an assistant. The same tester reviewed the videotapes along with 2 other trained testers approximately 4 to 6 weeks after the original session. The testers used a stopwatch to assess the time required to complete each task and were permitted to rewind the videotape as needed. Testers were originally trained to 100% agreement of functional criteria and within 0.5 seconds for timing.

We calculated both mean and median task times, because both have been used in the literature,16 and median functional ability scores. The intraclass correlation coefficient model 2,1 (ICC2,1)19 that assesses agreement between the 3 raters and direct observation was calculated.

Step 2: test-retest reliability, item stability, and validity of the WMFT in subjects with chronic mild and moderate hemiparesis

Two trained raters administered the WMFT 3 times over 14 days to study test-retest reliability and item stability. Each subject had the same rater administering the test at all 3 time points. Subjects were not given any intervention during the repeated baseline tests and were instructed not to practice any of the tasks from the WMFT. The UE portion of the FMA20 was administered during the first testing session, after the WMFT, and it provided a method of determining criterion validity because it is a well-validated and reliable test21, 22, 23 that can be considered a criterion standard for measuring impairment. Although the same tester administered the FMA and WMFT, the tester was unaware of the total WMFT scores, group allocation, or intent of this study.

Subjects were 33 women and 33 men (N=66) with a mean age of 58±14 years (range, 32−84y), a mean time poststroke of 59.6±74.6 months (range, 8−443mo), and a median NIHSS score of 0 (range, 0–3). Subjects were considered minimally affected if they met the inclusion criteria of at least 10° of active wrist and finger extension in the paretic UE, which are similar cut-off criteria for the previously mentioned studies.8, 18 This subgroup consisted of 6 women and 11 men (n=17) with a mean age of 61±16 years (range, 34−82y), a mean time poststroke of 46±75.0 months (range, 11−328mo), and a median NIHSS score of 0 (range, 0–3). The average FMA score was 52.2±8.8 (range, 36−63). The remainder of the subjects consisted of 27 women and 22 men (n=49) with a mean age of 57±12 years (range, 32−81y), a mean time poststroke of 64±74.6 months (range, 8−443mo), and a median NIHSS score of 1 (range, 0–3). These subjects can be considered as moderately impaired and had an average FMA score of 28.5±11.2 (range, 10−56). Functionally, this moderate group ranged from those who did not use the paretic arm, even in a supportive role, to those who can use the affected arm on its own without any adaptive equipment or other supports for tasks such as bathing or dressing.

Mean and median task times and median functional ability scores were calculated and correlated across time using ICCs to assess test-retest reliability. To assess the stability of the test items for performance time (interval data), a 1-way analysis of variance (ANOVA) with repeated measures was run on the mean time for each item of the test individually. We made adjustments for inflating α using Tukey-Kramer procedures. To assess the stability of the functional ability scale (ordinal data), we used a Wilcoxon signed-rank test for paired scores. For validation, we used the Spearman correlation coefficient to correlate subjects’ times, functional ability scores, and weight-to-box scores with those from the UE FMA.

Results

Step 1: Videotape Versus Observation

Scores for the mean and median performance times and median functional ability scales are presented in table 1. Because our sample included patients with a range of disability severity, it is noteworthy that using the median time results in a floor effect for subjects 2 and 9. The ICCs for mean and median time were .99 and .96, respectively, and .99 for functional ability.

Table 1.

Individual Subject Scores of Mean Time, Median Time, and Functional Ability From Each Rater and the Direct Observation


Subject	Variable	Rater 1	Rater 2	Rater 3	DO
1	Mean time (s) Median time (s) Functional ability	19.5 5.7 2	19.3 5.5 2	19.4 5.9 2	19.5 5.8 2
2	Mean time (s) Median time (s) Functional ability	115.3 120.0 1	115.2 120.0 1	115.2 120.0 1	115.2 120.0 1
3	Mean time (s) Median time (s) Functional ability	74.2 112.8 2	74.5 112.7 2	74.2 112.6 2	74.4 113.1 2
4	Mean time (s) Median time (s) Functional ability	11.5 3.2 3	11.5 2.9 2	11.5 3.4 3	11.5 3.4 3
5	Mean time (s) Median time (s) Functional ability	33.8 7.4 2	33.7 7.4 2	33.7 7.8 2	33.7 7.9 2
6	Mean time (s) Median time (s) Functional ability	36.2 12.7 2	36.1 12.6 2	36.1 12.8 2	36.1 12.7 2
7	Mean time (s) Median time (s) Functional ability	71.0 120.0 1	64.0 64.2 2	71.9 120.0 1	64.9 64.2 2
8	Mean time (s) Median time (s) Functional ability	29.4 4.6 2	29.5 4.7 2	29.4 4.5 2	29.5 5.1 2
9	Mean time (s) Median time (s) Functional ability	104.7 120.0 1	104.6 120.0 1	104.6 120.0 1	104.6 120.0 1
10	Mean time (s) Median time (s) Functional ability	60.7 37.6 2	60.8 38.2 2	60.8 37.8 2	60.8 37.8 2

Abbreviation: DO, direct observation.

Step 2: Test-Retest Reliability, Item Stability, and Criterion Validity

The test-retest ICC2,1 values of the total scores were high, ranging from .92 to .99, whether or not the subjects were considered as a whole or were segregated into mild and moderate groups (table 2). However, the 1-way ANOVA on the combined subject group showed that both mean time and functional ability scores differed significantly between the first and second test administrations (both P<.02) and between the first and third administrations (both P<.001). There were no differences between the second and third administrations, indicating that the second administration can be considered a stable measure of the total test score. The median time analysis showed no differences between scores at any of the 3 times.

Table 2.

Test-Retest ICC2,1 of WMFT Across 3 Administrations for All Subjects and Mild and Moderate Subgroups


Measure	All	Mild	Moderate
Mean time	.99	.98	.99
Median time	.99	.97	.99
Functional ability	.97	.92	.96

The individual item analyses showed few changes. In the combined subject pool for performance time, only item 12 (stack checkers) differed between the first and second test administrations (P=.05) but not between the first and third or the second and third. For the mildly affected subgroup, only item 1 (forearm to table) differed between the first and third administrations (P<.02). In the moderately affected group, item 6 (hand to box) differed between tests 1 and 2 and tests 1 and 3 (P<.01), item 12 differed between tests 1 and 3 (P<.003), and item 13 (flip cards) between times 2 and 3 (P<.04). There is no apparent similarity among these items that are less stable across time.

Correlations between time and the UE FMA were relatively high, ranging from –.69 to –.89 for time and from .66 to .88 for functional ability (all P<.001). Table 3 presents Spearman correlation coefficient results for all subjects and those classified as mildly and moderately affected for time, functional ability, and weight to box.

Table 3.

Criterion Validity of WMFT Against UE FMA Using Spearman Correlation Coefficients for All Subjects and Mild and Moderate Subgroups


Measure	All	Mild	Moderate
Mean time	−.88	−.75	−.79
Median time	−.89(−.82)	−.69	−.80(−.68)
Functional ability	.88	.66	.79
Weight to box	.86	.85	.79

NOTE. Values in parentheses indicate correlations with 120 scores removed. P<.001 for all.

Comparison of subgroups

The t test (mean) and Wilcoxon (median, UE FMA) analyses on the mild and moderate subgroups confirmed that the groups differed significantly from each other on the FMA (52.2 vs 28.5, P<.001) and all 3 WMFT mean (eg, first administration mean, 12.1 vs 65.3; P<.001) and median (eg, first administration median, 4.2 vs 69.1; P<.001) scores.

Discussion

In this study, we determined the psychometric properties of a modified WMFT for subjects with mild- and moderate-severity hemiparesis. As a preliminary step, in a small sample, we compared results from videotaping with direct observation and determined that the latter provided similar results with a higher correlation as did 3 raters using the videotape. In a larger sample we showed good test-retest reliability and stability of the timed and functional ability scores and individual items across 3 administrations of the test. These results were similar for the total sample and for the subgroups of mildly and moderately affected subjects. In addition, there was a relatively high correlation of the WMFT’s time and functional ability scores with the UE FMA scores.

In general, results are similar to previous reliability and validity studies in which mildly affected subjects were tested. However, our results cannot be directly compared with other studies, in terms of the absolute scores, because we modified the test instructions and rescaled the scoring for functional ability. In fact, the WMFT (and the preceding Emory Test) has been modified over time by others as well, all with the purpose of making it a more accurate and useful test. For example, it is now standard practice to let subjects practice each task first with the nonparetic arm, but our original guidelines did not include this suggestion. Therefore, we ensured that subjects understood the task with a demonstration but did not let them practice with the nonparetic arm. Whether this type of practice influences the performance of the paretic arm has not been reported to our knowledge, although intuitively it would seem a useful addition for future studies. The usefulness of the WMFT, as a tool for assessing functional performance, is enhanced by the addition of the present modification, because it expands the test’s application to the moderately impaired population and this modification has good reliability and validity.

Regarding the necessity of videotaping the WMFT, the good agreement between the individual raters and between direct observation scores and those generated from videotape suggest that videotaping is not required to generate accurate scores. This supports the WMFT’s usefulness as an outcome measure in the clinic (modified or not), where videotaping is often not possible. It also enhances the WFMT’s usefulness as a research tool, because the time commitment of videotaping and rating the tapes can be eliminated.

Because mean and median times correlated equally well over the 3 trials and with the UE portion of the FMA, either method of calculating the timed score can be used. However, the results of the first analysis show that the median time can show a floor effect for moderately affected subjects. For example, in a patient who cannot complete more than half the items, being able to complete 1 additional item, posttraining, might represent a clinically significant functional change but no change in median score. Therefore, a reason to select mean rather than median time would be that as a continuous variable it will not result in a floor effect in those with large movement impairments. In addition, our data suggest that 2 administrations of the test increase the measurement stability, particularly if using mean times. Because it is not clinically practical to recall a patient on a subsequent day, it might be plausible to provide a practice session on the same day. It has not been established, however, that 2 administrations on the same day would provide a stable measure. At the least, it appears advisable to practice the test items with the nonparetic arm.

For criterion validity, our results extend those of Wolf et al18 by including subjects with more severe impairments. In fact, the correlations between the WMFT functional ability scale and the UE FMA reported here are generally higher than those of Wolf, who reported correlations of –.57 and –.54 at session 1 and –.67 and –.68 at session 2 for raters 1 and 2, respectively. Higher correlations in our study may have been caused by our modification of the instructions and scoring, the larger sample size in our study, or bias from the tester administering both tests.

When comparing the subgroups, we found that the FMA correlated more strongly with the WMFT functional ability score in the moderately affected group than in the minimally affected group of stroke subjects. This could be a result of the larger sample size for the moderately affected group or the scoring modification. That is, those who are moderately affected tend to have functional ability scores of 1 or 2, which are easier to discriminate between than scores of 3 to 4 (see appendix 1). This argument implies that the modified WMFT is a more discriminating outcome measure for those who have been moderately versus mildly affected as the result of a stroke. This logic ties in with our rationale for modifying the scoring in the first place.

Conclusions

The WMFT (modified for more impaired subjects) can be used without videotape and has good interrater reliability and high criterion validity with the UE FMA in subjects with both mild- and moderate-severity hemiparesis. The test-retest reliability results are high, but the stability assessment shows that it may be prudent to administer the WMFT a second time to obtain stable measures. Overall, the clinical and scientific applicability of the WMFT is broadened.

Acknowledgments

We thank Richard Macko, MD, for screening subjects; Jennifer Sulin-Stair for assistance with data collection; and Steven Wolf, PT, PhD, for helpful comments on an earlier draft of this article.

Appendix 1. Functional ability scale of the modified WMFT

0=Does not attempt with the involved arm.

1=Involved arm does not participate functionally; however, an attempt is made to use the arm.

2=Arm does participate, but movement is influenced to some degree by synergy or is performed slowly and/or with effort.

3=Arm does participate; movement is close to normal but slightly slower; may lack precision, fine coordination, or fluidity.

4=Arm does participate; movement appears to be normal.

For the determination of normal, the uninvolved arm can be used as an available index for comparison, with premorbid limb dominance taken into consideration.

References

1. 1 American Heart Association . Heart disease and stroke statistics—2003 update . Dallas: AHA; 2002; .

2. 2 Gowland C , deBruin H , Basmajian J , Curcea I . Agonist and antagonist activity during voluntary upper limb movement in subjects with stroke . Phys Ther . 1992;72:624–633 . MEDLINE

3. 3 Lyle RC . A performance test for assessment of upper limb function in physical rehabilitation treatment and research . Int J Rehabil Res . 1981;4:483–492 . MEDLINE

4. 4 Desrosiers J , Herbert R , Dutil E , Bravo G . Development and reliability of an upper extremity function test for the elderly (the TEMPA) . Can J Occup Ther . 1993;60:9–16 .

5. 5 McCulloch K , Cook EW , Fleming WC , Novack TA , Taub E . A reliable test of upper extremity ADL function . [abstract] Arch Phys Med Rehabil . 1988;69:755 .

6. 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 . MEDLINE

7. 7 Wolf SL , Lecraw DE , Barton LA , Jann BB . Forced use of hemiplegic upper extremities to reverse the effect of learned nonuse among chronic stroke and head-injured subjects . Exp Neurol . 1989;104:125–132 . MEDLINE | CrossRef

8. 8 Morris DM , Uswatte G , Crago JE , Cook EW , Taub E . The reliability of the Wolf motor function test for assessing upper extremity function after stroke . Arch Phys Med Rehabil . 2001;82:750–755 . Abstract | Full Text | Full-Text PDF (48 KB) | CrossRef

9. 9 Kunkel A , Kopp B , Muller G , et al. Constraint-induced movement therapy for motor recovery in chronic stroke subjects . Arch Phys Med Rehabil . 1999;80:624–628 . Abstract | Full-Text PDF (1593 KB) | CrossRef

10. 10 Sterr A , Elbert T , Berthold I , Kolbel S , Rockstroth B , Taub E . Longer versus shorter daily constraint-induced movement therapy of chronic hemiparesis (an exploratory study) . Arch Phys Med Rehabil . 2002;83:1374–1377 . Abstract | Full Text | Full-Text PDF (68 KB) | CrossRef

11. 11 Page SJ , Sisto S , Johnson MV , Levine P , Hughes M . Modified constraint-induced therapy in subacute stroke (a case report) . Arch Phys Med Rehabil . 2002;83:286–290 . Abstract | Full Text | Full-Text PDF (44 KB) | CrossRef

12. 12 Pierce SR , Daly K , Gallagher KG , Gershkoff AM , Schaumburg SW . Constraint-induced therapy for a child with hemiplegic cerebral palsy (a case report) . Arch Phys Med Rehabil . 2002;83:1462–1463 . Abstract | Full Text | Full-Text PDF (49 KB) | CrossRef

13. 13 Page SJ , Sisto SA , Levine P , Johnston MV , Hughes M . Modified constraint induced therapy (a randomized feasibility and efficacy study) . J Rehabil Res Dev . 2001;38:583–590 . MEDLINE

14. 14 Sterr A , Freivogel S , Voss A . Exploring a repetitive training regime for upper limb hemiparesis in an in-patient setting (a report of three case studies) . Brain Inj . 2002;16:1093–1107 . MEDLINE | CrossRef

15. 15 Schaechter JD , Kraft E , Hilliard TS , et al. Motor recovery and cortical reorganization after constraint-induced movement therapy in stroke subjects (a preliminary study) . Neurorehabil Neural Repair . 2002;16:326–338 . MEDLINE

16. 16 Whitall J , McCombe Waller S , Silver KH , Macko RF . Repetitive bilateral arm training with rhythmic auditory cueing improves motor function in chronic hemiparetic stroke . Stroke . 2000;31:2390–2395 .

17. 17 Duncan P , Studenski S , Richards L , et al. Randomized clinical trial of therapeutic exercise in subacute stroke . Stroke . 2003;34:2173–2180 . CrossRef

18. 18 Wolf SL , Catlin PA , Ellis M , Archer AL , Morgan B , Piacentino A . Assessing the wolf motor function test as outcome measure for research in subjects after stroke . Stroke . 2001;32:1635–1639 .

19. 19 Domholdt E . In: Physical therapy research principles and applications . Philadelphia: WB Saunders; 1993;p. 288 .

20. 20 Fugl-Meyer AR , Jaasko L , Leyman I , Olsson S , Steglind S . Post-stroke hemiplegic patient (a method for evaluation of physical performance) . Scand J Rehabil Med . 1975;7:13–31 . MEDLINE

21. 21 Berglund K , Fugl-Meyer A . Upper extremity function in hemiparesis . Scand J Rehabil Med . 1986;18:155–157 . MEDLINE

22. 22 Filiatraut J , Arsenault A , Dutil E , Bourbonnais D . Motor function and activities of daily living assessments (a study of three tests for persons with hemiparesis) . Am J Occup Ther . 1991;45:806–810 . MEDLINE

23. 23 Duncan PW , Propst M , Nelson SG . Reliability of the Fugl-Meyer assessment of sensorimotor recovery following cerebrovascular accident . Phys Ther . 1983;63:1606–1610 . MEDLINE

a Department of Physical Therapy and Rehabilitation Science, University of Maryland School of Medicine, Baltimore, MD

b National Institutes of Health, National Institute of Neurological Disorders and Stroke, Human Cortical Physiology Section, Bethesda, MD

Reprint requests to Jill Whitall, PhD, Dept of Physical Therapy and Rehabilitation Science, University of Maryland School of Medicine, 100 Penn St, Baltimore, MD 21201

Supported by the National Institute of Disability and Rehabilitation Research (grant no. H133G010111) and the Claude Pepper Older Americans Independence Center, National Institute on Aging (grant no. P6012583).

No commercial party having a direct financial interest in the results of the research supporting this article has or will confer a benefit upon the author(s) or upon any organization with which the author(s) is/are associated.

PII: S0003-9993(06)00108-0

doi:10.1016/j.apmr.2006.02.004

11 of 25