Ten age-matched, right-handed healthy control subjects (mean age 59 ± 13 years; 5 females, 5 males) were also studied. These subjects were free of neurologic or orthopedic conditions that might affect their upper extremities. The Institutional Review Board of Washington University School of Medicine approved the protocol for this study. Informed consent was obtained from all subjects prior to testing.
Measurement Of Upper Extremity Use
Uni-axial accelerometers (model 7164-2.4 Activity Monitors, MTI Health Services, Fort Walton Beach, FL) were placed on each upper extremity following a published protocol.6,7 For the upper extremity and for the duration of time measured, uni-axial accelerometers provide the same information as multi-axial accelerometers about whether or not the limb is moving.11 This may be because upper extremity movements nearly always involve joint rotations in multiple planes of movement, and not in a single plane. Briefly, accelerometers were placed on the distal arm just above the wrist with the axis parallel to the length of the arm. Data was collected for 24 hours, stored in two-second epochs, and downloaded onto a personal computer for subsequent analyses. Subjects were instructed to wear the accelerometer at all times during the 24 hour data collection period except for times when the devices would be exposed to water (eg personal hygiene). The 24 hour period was chosen because: 1) that is the length of one day, enabling us to capture activity that might occur outside of the typical workday or therapy hours, and 2) people with hemiparesis and controls were less interested in participating or less-compliant with wearing them if they had to be worn for longer periods.
We note that accelerometer measures of upper extremity use have some limitations (see Discussion) since they provide information about whether or not the upper extremity was moving, but not information about what it was moving to do (eg functional task versus arm swing during gait). Thus, measures of upper extremity use obtained via accelerometry may be considered small overestimations of the time spent using the upper extremity for functional activities.
Measurement Of Upper Extremity Impairments
We measured sensorimotor impairments that are commonly assessed by physical and occupational therapists in stroke rehabilitation settings. Light touch sensation was measured at 4 locations on the arm using the 5 monofilament minikit (North Coast Medical, Morgan Hill, CA). The 4 locations were: the dorsal web space between the thumb and index finger, the dorsal aspect of the midforearm, slightly superior to the elbow on the lateral aspect of the arm, and the anterior aspect of the shoulder. The smallest monofilament sensed at each location was recorded and given an ordinal score.12 Ordinal scores were assigned as follows: perception of 2.83 monofilament = 0, perception of 3.61 monofilament = 1, perception of 4.31 monofilament = 2, perception of 4.56 monofilament = 3, perception of 6.65 filament = 4, and unable to perceive the 6.65 monofilament = 5. Because scores across sites were well-correlated (r values ranged from 0.75–0.92, p values < 0.01), a composite light touch sensation score was calculated by averaging the ordinal scores from the 4 test sites. Joint position sense was measured at the 1st metacarpal-phalangeal joint following standard clinical techniques. Joint position sense was scored as normal (correct response ≥ 3 of 5 trials) or impaired (correct response < 3 of 5 trials). Elbow joint spasticity was measured using the Modified Ashworth Scale.13 Shoulder pain was measured immediately prior to testing using a 10 point rating scale. Shoulder flexion, elbow flexion, and wrist extension active ranges of motion (AROM) against gravity were measured while the subject was seated using motion analyses techniques.1 Motion analyses techniques were used because it was part of a larger testing protocol for the VECTORS trial, where subjects underwent kinematic analyses of reaching and grasping movements. AROM in degrees was calculated as the total excursion of the joint against gravity. AROM values from the motion analyses techniques are not different than those that can be obtained with a goniometer. Shoulder flexion values were limited to 90° for all subjects because of the testing protocol. Strength of the shoulder, elbow, and wrist flexor and extensors were measured using a hand-held dynamometer (MICROFET2, Hogan Health Industries, Draper, UT) following a standard protocol14 except that subjects were seated during testing. Maximal voluntary isometric strength values were recorded in lbs. for each muscle group tested. Subjects unable to produce force against the dynamometer were given a score of 0 lbs. for that particular muscle group. The strength of each muscle group was expressed as the ratio of affected side to unaffected side maximal isometric force. Maximum grip strength was measured as part of the Wolf Motor Function Test in kilograms (see below).
Measurement of Upper Extremity Activity
We measured activity using 3 clinical scales: the Action Research Arm Test (ARAT), the Wolf Motor Function Test (WMFT), and the Functional Independence Measure (FIM). All evaluations were performed by trained study personnel.
The ARAT assesses activity limitations of the upper extremity. It includes 19 items divided into four subscales: grasp, grip, pinch, and gross movement. Reliability (interrater 0.99, test-retest 0.98) and validity of the ARAT have been well established.15–20 Performance on the ARAT is strongly correlated to performance on the upper extremity motor portion of the Fugl-Meyer scale and to performance on the Box and Block test.16,21 Item scores on the ARAT are summed to create subtest and full-scale scores with a maximum score of 57, indicating normal performance.
The WMFT is a 17-item measure used to quantify activity limitations of the upper extremity. It is comprised of 2 strength items and 15 timed task performance items. The task performance items begin with the measurement of simple proximal movements and progress to more complex distal and whole limb movements. The WMFT yields two scores: 1) a functional ability score quantifying quality of performance, and 2) a timed score quantifying speed of performance in seconds. The test has published reliability and validity.22–25 In the VECTORS study, the key use task was not collected, and the results reported do not include this item.
The FIM is an 18 item global activity measure incorporating concepts and items of functional performance including activities of daily living, bowel and bladder function, social cognition, functional communication, and functional mobility.26 The FIM is scored on a 7 point ordinal scale, where a score of 1 indicates dependence and a score of 7 indicates independence. Reliability and validity of the FIM have been well established.27–29 The FIM was included in this report because it is frequently used in many inpatient rehabilitation centers and by other rehabilitation professionals. The 15 items requiring motor function were summed to create a FIM motor score, while a subset of 5 items relating to upper extremity use were summed to create a FIM upper extremity score.10,20 Because of the way the items are worded, upper extremity items could have been completed with the paretic or the nonparetic limb. For this study, the FIM was collected during a patient interview and performance of activities was not observed. Thus, scores reflect the self-perceived ability to complete the items.
The raw data obtained from the accelerometer were in “activity counts”, a unitless number representing the summation of all accelerations recorded during an epoch (here, an epoch = 2 seconds). Custom software was therefore written in the C programming language to transform the raw accelerometer data using an established methodology that provides a valid and reliable measure of the duration of upper extremity use.6,7 The transformation involved dichotomizing the raw value for each two-second epoch around a threshold of 1. This is a low threshold, such that even very slow or very small movements in a paretic upper extremity would be counted as movement. If the raw count for the epoch was ≥ 1, then the upper extremity was considered to have been moving during that 2 second time period. If the raw count for the epoch was < 1, then the upper extremity was considered to have been still for that 2 second time period. The sum of the epochs when the upper extremity moved therefore represented the duration of upper extremity movement over the 24 hour data collection period. For ease of communication, this summed variable was converted from seconds to hours.
Statistical analyses were conducted using SPSS for Windows Version 13.0 and the criterion for statistical significance was p < 0.05. Accelerometry data were normally distributed as assessed by Kolmogorov-Smirnov tests. A repeated measures ANOVA with 2 factors (group, side) was used to look for differences in dominant and nondominant upper extremity use in the control group and unaffected and affected upper extremity use in the hemiparetic group. When significant main and interaction effects were found, post hoc t tests with Bonferroni corrections for multiple comparisons were used to determine where the differences existed. Spearman correlation coefficients were used to examine relationships between upper extremity use and impairment and activity level measures. Nonparametric correlational analyses were chosen because, although upper extremity use was normally distributed, other measures were not normally distributed and/or were measured on ordinal rating scales. Based on our sample size, correlation coefficients greater than 0.33 were significant at p < 0.05 and coefficients greater than 0.42 were significant at the p < 0.01 level. The following criteria are provided to assist in the interpretation of the magnitude of the correlation coefficients. Coefficients of 0.25 or below were considered low, coefficients ranging from 0.26 to 0.50 were considered fair, coefficients from 0.51 to 0.75 were considered good, and those greater than 0.75 were considered excellent.30
Amount of Use
Using accelerometers, we measured upper extremity use during a 24 hr period in a group of healthy controls and a group of people with acute hemiparesis post stroke (Fig. 1). The control group used their dominant upper extremity 8.7 ± 1.3 hrs (range 6.5–10.5) and used their nondominant upper extremity 8.4 ± 1.2 hrs (range 6.2–10.2). The hemiparetic group used both arms for less time than the control group (main effect of group, p < 0.0001). Affected upper extremity use in the hemiparetic group was 3.3 ± 1.8 hrs (range 0.8–8.1) and unaffected upper extremity use was 6.0 ± 1.6 hrs (range 3.2–9.4). Affected upper extremity use was not different between those with the dominant side affected and those with the nondominant side affected by the stroke (p = 0.40). There was a differential effect of upper extremity use between the two groups (group × side interaction, p = 0.02). Post hoc testing indicated that that dominant and nondominant use were similar in the control group (p = 0.55) and were greater than either affected or unaffected use (p values < 0.01). Affected upper extremity use was less than unaffected upper extremity use in the hemiparetic group (p < 0.0001). Unaffected side use in the hemiparetic group was positively correlated to affected side use (r = 0.55, p < 0.01), such that hemiparetic subjects who used their affected upper extremity for shorter durations often used their unaffected upper extremity for shorter durations also. Age was negatively correlated with affected side use (r = −0.60, P < 0.01) but was not correlated with unaffected side use (r = −0.26, p > 0.05).
Relationships Between Use and Impairment and Activity Measures
Correlations between affected upper extremity use in the hemiparetic group and impairment and activity level measures are shown in Table 3. For the impairment level measures, AROM, strength, and shoulder pain were positively correlated with affected upper extremity use, while spasticity and somatosensory measures were not significantly correlated. Among the impairment level measures, wrist extension AROM had the highest correlation coefficient with affected upper extremity use, such that increased wrist extension AROM was related to increased use. (Correlations between impairment level measures themselves have been previously reported.1) Clinical tests of upper extremity activity and global activity were correlated with affected upper extremity use (Table 3). Better upper extremity function as measured by the ARAT and WMFT was related to more upper extremity use. Upper extremity use was significantly related to both FIM motor score and FIM upper extremity score. To further illustrate the nature of these relationships and the variability within the data, scatter plots of affected upper extremity use versus wrist extension AROM (Fig. 2A), Wolf Motor Function Test function scores (Fig. 2B), and FIM Motor scores (Fig. 2C) are provided.
Healthy, neurologically-intact adults use their dominant and nondominant upper extremities 8–9 hours per day. After mild-to-moderate stroke, people with hemiparesis in the inpatient rehabilitation setting use their affected and unaffected upper extremities substantially less than the healthy adults. Seven of the ten impairment level and each of the activity level measures were related to affected upper extremity use. The impairment measures that were related to upper extremity use were those measures that assessed the ability to activate muscles (i.e. active range of motion and force production) and the measurement of shoulder pain.
Amount of Use in Control and Hemiparetic Samples
Our small sample of community-dwelling, neurologically-intact adults used their dominant and nondominant upper extremities about 1/3 of the time in a 24 hour period. This amount of use is similar to a preliminary report recording dominant hand, thenar muscle activity during 24 hour periods.31 Thenar muscles may be activated more during the day than other upper extremity muscles since thenar muscles are involved in a variety of grasping tasks when other upper extremity muscle may be more quiescent (eg grasping a steering wheel while driving).32 The fact that our data collected with accelerometers matches this preliminary data collected with surface electromyography provides concurrent validity for both measurement techniques. Earlier surface electromyography studies found other individual upper extremity muscles activated for much smaller periods of time (<2 hours),32–34 using higher thresholds to determine the presence or absence of activity. Interestingly, activity in different upper extremity muscles was not correlated,34 suggesting that longer estimations of amount of overall arm use (which may be closer to the values obtained here) might have been possible if activity of multiple upper extremity muscles had been summed. Our finding that the hours of dominant and nondominant upper extremity use are not different supports observations that the upper extremities are most often used together to perform bimanual tasks during daily activity.35 It may seem unfair to compare upper extremity use in community dwelling, neurologically-intact individuals with upper extremity use in hospitalized individuals with hemiparesis post stroke. In the inpatient setting, many regular daily activities are limited. We consider it likely that the inpatient setting reduces upper extremity use by itself, regardless of whether or not an individual has hemiparesis. This could be tested directly in future studies by including an inpatient control group, such as people with general medical or orthopedic conditions. We chose a community-dwelling control group for the current study because we wanted to see just how close or how far acute hemiparetic upper extremity use was from where it may need to be. Our control group averages can now be used as a goal to which people with hemiparesis post stroke can strive to attain.
As mentioned above, our data on the average duration of hemiparetic upper extremity use was obtained in the inpatient rehabilitation setting. Given that the expectation for therapy services during inpatient rehabilitation in the United States is 3 hours of therapy per day, it is noteworthy that the average upper extremity use for the affected side was 3.3 hours per day. These data from a 24 hour period are consistent with the finding that patients in stroke units are inactive for the majority of their day, as observed from 8 am to 5 pm.36 Thus, even in what is often considered the most intensive rehabilitation setting for people post stroke, the affected upper extremity is being used for only a small portion of the day. The important clinical question that follows from our study is: how can we now increase affected upper extremity use in this population? The answer to this question requires further study. Moreover, a clinically useful answer to this question needs to be both economically feasible and realistic to implement.
We were surprised that upper extremity use on the unaffected side was also lower in the hemiparetic group compared to the control group. Additionally, affected and unaffected upper extremity use in the hemiparetic group were positively correlated, indicating that when the affected hand was used only a small amount, the unaffected hand was generally used only a small amount also. We had wrongly expected that unaffected upper extremity use would be high to compensate for the limited ability to use the affected side. It is possible that the low use on the unaffected side is a due to the fact that subjects were studied while they were inpatients, where the lower accelerometer values in the unaffected side of people with hemiparesis reflect a general decrease in activity. While rehabilitation inpatients are routinely encouraged to do as much as possible for themselves, the hospital environment may not require the daily activities the patient would have to accomplish at home (eg meal preparation, laundry). It is possible that the positive relationship between affected and unaffected upper extremity use would be reversed (a negative correlation) in people with hemiparesis discharged to the home environment.
Relationships Between Use and Impairment and Activity Measures
Affected upper extremity use was correlated to impairment and activity level measures. Because accelerometer use reflects what the subject actually does during the 24 hour period (versus what they may be capable of doing in the clinic or laboratory), we had anticipated that the strongest correlations would be with similar measurements, ie activity measurements of paretic limb function. We found good correlations between upper extremity use and the FIM motor and FIM upper extremity scores, indicating that individuals who reported greater independence with mobility and self-care tended to use their upper extremity more. This was interesting because, as used here, the FIM scores provide information about the self-perceived global capability and amount of assistance needed for daily activities, but are not specific to paretic limb capabilities. We speculate that these relationships are good because it is early after stroke, and that in other environments (eg home) and at later time points post stroke, these relationships could become much weaker. Equally strong relationships were found between upper extremity use and the WMFT scores, measures that are specific to upper extremity capabilities and were originally designed to be used in a laboratory for research studies.22,23 This strong relationship means that individuals with better hand function used their upper extremity more. The good correlations with the WMFT scores provide evidence for the construct validity of the WMFT. Interestingly, equally strong relationships were also found between upper extremity use and wrist extension active range of motion where those individuals who had greater wrist extension active range of motion used their upper extremity more. Our data lend support to the validity of selecting wrist extension active range of motion as a meaningful criterion for implementing constraint-induced movement therapy in the subacute phase after stroke.37
From a clinical perspective, the good correlation with wrist extension active range of motion implies that this is an important impairment to measure during evaluations of people with hemiparesis post stroke. When making clinical decisions based on examination findings, our data suggest that wrist extension active range of motion measures should be weighted more heavily compared to other impairment measures. The low and nonsignificant correlations between upper extremity use and spasticity and somatosensory measures implies that results from these impairment measures should be weighted less heavily in the clinical decision making process. The positive correlation of 0.41 between upper extremity use and shoulder pain is also highly clinically relevant. This finding indicates that those who used their upper extremity more had greater pain. Overall, average pain ratings were low (1/10) and this positive correlation between use and pain highlights the importance of managing hemiparetic shoulder pain earlier on, before more intense pain may discourage upper extremity use. Lastly, caution should be used when interpreting our correlations with respect to individual patients because there is considerable variability within the sample as can be seen in the scatter plots of Figure 2.
Our study population was recruited for a clinical trial based on the presence of hemiparesis post stroke. Based on the inclusion and exclusion criteria for the clinical trial, our subjects represent a subgroup of people with hemiparesis seen in inpatient rehabilitation. This subgroup had relative pure motor hemiparesis, where muscle strength on the affected side ranged from 0–100% of the strength on the unaffected side (Table 2). They generally had minimal deficits in somatosensation and were largely free of cognitive, attentional, and language deficits due to the exclusion criteria. Thus, the amount of upper extremity use found here cannot be generalized to all patients with hemiparesis post stroke or to other time points post stroke. We speculate that the subpopulation studied here are the individuals who might have the greatest amount of upper extremity use shortly after stroke because they are relatively free of deficits in other domains. If this is the case, then people with stroke-induced deficits in multiple domains might be expected to use their upper extremities for even smaller amounts of time than the values provided here.
We used accelerometry to measure the amount of upper extremity use for a 24 hour period. While wrist accelerometers have been shown to produce valid and reliable measures of upper extremity use in people with hemiparesis,6,7 they are not perfect measurement instruments. The term “upper extremity use” is not synonymous with the term “upper extremity purposeful use”. When obtaining the number of hours with accelerometry, movements such as arm swing during gait are counted equally with movements that are more purposeful, such as feeding oneself. Movements such as riding passively in a wheelchair however do not register on the accelerometers. Thus, the values provided here potentially overestimate actual purposeful upper extremity use in our subjects. It is also possible that collecting data for only 24 hours biased our results in one direction or the other. This could have happened if the accelerometers were worn on a day when a particular subject was unusually active or unusually inactive. We believe however, that additional random error generated with only a 24 hour data collection period (vs a 48 or 72 hour period) would be minimal with our hemiparetic sample size (n = 34).
We found that people with hemiparesis post stroke use their affected upper extremity only a small amount while in the inpatient rehabilitation setting. Our data provide information to clinicians about how much the upper extremity is used during acute inpatient rehabilitation and how upper extremity use is related to other clinical measurements they may take.
In the future, accelerometry measurements of upper extremity use could become part of routine clinical practice. Commercially-available accelerometers are slightly larger than a big watch and relatively easy to apply. They require a few minutes of set-up prior to placing them on a patient and a few minutes after removing them to obtain the data. Based on their time and monetary costs, it may be feasible to use them as a clinical tool to measure individual patient’s progress over the course of therapy, especially in the inpatient rehabilitation setting. If the ultimate goal of upper extremity rehabilitation is to promote independence and a return to productive activities in the real world, not just in the hospital, clinic, or laboratory, then objective information about actual upper extremity use provided by wrist accelerometers would be of great value to clinicians as they select treatments and evaluate progress.
This work was supported by NIH NS41261, HD047669, James S. McDonnell Foundation 21002032, and the Foundation for Physical Therapy Promotion of Doctoral Studies Scholarship. We thank Lily Hu for her assistance with data collection and processing and the therapists who assisted with recruitment and scheduling during this project.
1. Wagner JM, Lang CE, Sahrmann SA, et al. Relationships between sensorimotor impairments and reaching deficits in acute hemiparesis. Neurorehabil Neural Repair.
2. Caro JJ, Huybrechts KF, Duchesne I. Management patterns and costs of acute ischemic stroke: An international study. For the stroke economic analysis group. Stroke.
3. Spieler JF, Lanoe JL, Amarenco P. Costs of stroke care according to handicap levels and stroke subtypes. Cerebrovasc Dis.
4. Bernmark E, Wiktorin C. A triaxial accelerometer for measuring arm movements. Appl Ergon.
5. Chen KY, Acra SA, Majchrzak K, et al. Predicting energy expenditure of physical activity using hip- and wrist-worn accelerometers. Diabetes Technol Ther.
6. Uswatte G, Miltner WH, Foo B, et al. Objective measurement of functional upper-extremity movement using accelerometer recordings transformed with a threshold filter. Stroke.
7. Uswatte G, Foo WL, Olmstead H, et al. Ambulatory monitoring of arm movement using accelerometry: An objective measure of upper-extremity rehabilitation in persons with chronic stroke. Arch Phys Med Rehabil.
8. Lang CE, Wagner JM, Edwards DF, Sahrmann SA, Dromerick AW. Recovery of grasp vs. Reach in people with hemiparesis post stroke. Neurorehabil Neural Repair.
9. Lang CE, Wagner JM, Bastian AJ, et al. Deficits in grasp versus reach during acute hemiparesis. Exp Brain Res.
10. Dromerick AW, Lang CE, Birkenmeier R, et al. Relationships between upper-limb functional limitation and self-reported disability three months after stroke. J Rehabil Res Dev.
11. Redmond DP, Hegge FW. Observations on the design and specification of a wrist-worn human activity monitoring system. Behav Res Methods.
12. Zackowski KM, Dromerick AW, Sahrmann SA, Thach WT, Bastian AJ. How do strength, sensation, spasticity and joint individuation relate to the reaching deficits of people with chronic hemiparesis? Brain.
13. Bohannon RW, Smith MB. Interrater reliability of a modified ashworth scale of muscle spasticity. Phys Ther.
14. Andrews AW, Thomas MW, Bohannon RW. Normative values for isometric muscle force measurements obtained with hand-held dynamometers. Phys Ther.
15. Lyle RC. A performance test for assessment of upper limb function in physical rehabilitation treatment and research. Int J Rehabil Res.
16. De Weerdt W, Harrison MA. Measuring recovery of arm-hand funciton in stroke patients: A comparison of the brunnstrom-fugl-meyer test and the action research arm test. Physiother Can.
17. Hsieh CL, Hsueh IP, Chiang FM, Lin PH. Inter-rater reliability and validity of the action research arm test in stroke patients. Age Ageing.
18. van der Lee JH, De Groot V, Beckerman H, et al. The intra- and interrater reliability of the action research arm test: A practical test of upper extremity function in patients with stroke. Arch Phys Med Rehabil.
19. van der Lee JH, Beckerman H, Lankhorst GJ, Bouter LM. The responsiveness of the action research arm test and the fugl-meyer assessment scale in chronic stroke patients. J Rehabil Med.
20. Lang CE, Wagner JM, Dromerick AW, Edwards DF. Measurement of upper extremity function early after stroke: Properties of the action research arm test. Arch Phys Med Rehabil.
21. Platz T, Pinkowski C, van Wijck F, et al. Reliability and validity of arm function assessment with standardized guidelines for the Fugl-Meyer test, action research arm test and box and block test: A multicentre study. Clin Rehabil.
22. Wolf SL, Catlin PA, Ellis M, et al. Assessing wolf motor function test as outcome measure for research in patients after stroke. Stroke.
23. 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 patients. Exp Neurol.
24. Taub E, Miller NE, Novack TA, et al. Technique to improve chronic motor deficit after stroke. Arch Phys Med Rehabil.
25. Morris DM, Uswatte G, Crago JE, Cook EW3rd, Taub E. The reliability of the wolf motor function test for assessing upper extremity function after stroke. Arch Phys Med Rehabil.
26. Granger CV, Hamilton BB. The uniform data system for medical rehabilitation report of first admissions for 1991. Am J Phys Med Rehabil.
27. Hamilton BB, Laughlin JA, Fiedler RC, Granger CV. Interrater reliability of the 7-level functional independence measure (FIM). Scand J Rehabil Med.
28. Linacre JM, Heinemann AW, Wright BD, Granger CV, Hamilton BB. The structure and stability of the functional independence measure. Arch Phys Med Rehabil.
29. Ottenbacher KJ, Hsu Y, Granger CV, Fiedler RC. The reliability of the functional independence measure: A quantitative review. Arch Phys Med Rehabil.
30. Portney LG, Watkins MP. Foundations of clinical research: Applications to clinical practice.
Norwalk: Appleton & Lange; 1993:439–455.
31. Thomas CK, Peterson LB, Klein CS, et al. Daily use of human thenar muscles. Society for Neuroscience Abstract 2005; Program No. 397.
32. Monster AW, Chan H, O’Connor D. Activity patterns of human skeletal muscles: Relation to muscle fiber type composition. Science.
33. Fuglevand AJ, Bilodeau M, Enoka RM. Short-term immobilization has a minimal effect on the strength and fatigability of a human hand muscle. J Appl Physiol.
34. Kern DS, Semmler JG, Enoka RM. Long-term activity in upper- and lower-limb muscles of humans. J Appl Physiol.
35. Kilbreath SL, Heard RC. Frequency of hand use in healthy older persons. Aust J Physiother.
36. Bernhardt J, Dewey H, Thrift A, Donnan G. Inactive and alone: Physical activity within the first 14 days of acute stroke unit care. Stroke.
© 2007 Neurology Section, APTA
37. Winstein CJ, Miller JP, Blanton S, et al. Methods for a multisite randomized trial to investigate the effect of constraint-induced movement therapy in improving upper extremity function among adults recovering from a cerebrovascular stroke. Neurorehabil Neural Repair.