Sage Journals: Discover world-class research

Abstract

Objective. To evaluate for any clinical effects of home-based mirror therapy and subsequent cortical reorganization in patients with chronic stroke with moderate upper extremity paresis. Methods. A total of 40 chronic stroke patients (mean time post .onset, 3.9 years) were randomly assigned to the mirror group (n = 20) or the control group (n = 20) and then joined a 6-week training program. Both groups trained once a week under supervision of a physiotherapist at the rehabilitation center and practiced at home 1 hour daily, 5 times a week. The primary outcome measure was the Fugl-Meyer motor assessment (FMA). The grip force, spasticity, pain, dexterity, hand-use in daily life, and quality of life at baseline—posttreatment and at 6 months—were all measured by a blinded assessor. Changes in neural activation patterns were assessed with functional magnetic resonance imaging (fMRI) at baseline and posttreatment in an available subgroup (mirror, 12; control, 9). Results. Posttreatment, the FMA improved more in the mirror than in the control group (3.6 ± 1.5, P < .05), but this improvement did not persist at follow-up. No changes were found on the other outcome measures (all Ps >.05). fMRI results showed a shift in activation balance within the primary motor cortex toward the affected hemisphere in the mirror group only (weighted laterality index difference 0.40 ± 0.39, P < .05). Conclusion. This phase II trial showed some effectiveness for mirror therapy in chronic stroke patients and is the first to associate mirror therapy with cortical reorganization. Future research has to determine the optimum practice intensity and duration for improvements to persist and generalize to other functional domains.

Keywords

stroke rehabilitation upper extremity randomized controlled trial functional magnetic resonance imaging (fMRI)

Introduction

From 55% to 75% of stroke survivors have a paretic arm¹ that may improve primarily within 6 months.² Further intensive training can lead to improved motor function and associated cortical reorganization.³ Yet these training programs often make use of expensive apparatus⁴ or require an intensive one-on-one interaction with a therapist,⁵ which hinders implementation on a large scale. Mirror therapy may be a suitable alternative. Designed by Ramachandran et al,⁶ mirror therapy was originally developed to diminish phantom limb pain in amputees. The reflection of the unimpaired arm in a mirror gave patients the sensation of having 2 moving arms, which led to a reduction in pain. In 1999, Altschuler et al⁷ introduced mirror therapy for recovery of hemiparesis following stroke. In a crossover design, they showed that motor performance of chronic stroke patients improved. Although several additional studies were small and often not well controlled,^8-10 2 recent, high-quality, randomized controlled trials have also reported mirror therapy to improve motor function in patients with subacute¹¹ and acute¹² stroke.

Despite the encouraging clinical results, little is known about the underlying mechanisms of mirror therapy. Ramachandran et al⁶ referred to a “learned paralysis” in the brain, which could possibly be “unlearned” as a result of the mirror illusion.¹³ Other studies have attributed the positive effects of mirror therapy in stroke to motor imagery⁸ or the mirror neuron system.¹¹ The prevailing idea is that observing mirrored movements causes additional neural activity in motor areas located in the affected hemisphere, which should eventually result in cortical reorganization and improved function. Research in healthy subjects has provided some evidence of such mechanisms, with the use of either transcranial magnetic stimulation^14,15 or functional magnetic resonance imaging (fMRI).¹⁶ However, results of these studies are not conclusive, do not give insight into long-term neuronal changes, and do not necessarily apply to stroke patients.

The aim of this phase II study was 2-fold. First, we evaluated the effect of mirror therapy on upper extremity function in a group of chronic stroke patients. As the intervention had to be both effective and efficient, an unsupervised training program to be performed at home was developed and carried out with supervised weekly sessions. Our primary focus was on improvements in motor function, but to get a detailed insight in any effects of mirror therapy, outcomes were also measured at the other International Classification of Functioning, Disability and Health (ICF)¹⁷ domains. Second, we used fMRI to examine whether mirror therapy could induce cortical reorganization.

Materials and Methods

Participants

After contacting 182 outpatients (hospitalized between January 1998 and September 2007) from the Rijndam Rehabilitation Centre in Rotterdam, the Netherlands, we enrolled 40 patients. Inclusion criteria were knowledge of the Dutch language, a Brunnstrom score for the upper extremity between III and V¹⁸ (the 6 stages of the Brunnstrom score range from [I] flaccidity toward [VI] full-range voluntary extension and individual finger movements present but less accurate than on the opposite side), home dwelling status, and at least 1 year poststroke. Patients with neglect, comorbidities that influenced upper extremity usage, or a history of multiple strokes were excluded. For patients to participate in the fMRI experiment the following additional inclusion criteria applied: ability to perform a hand squeezing movement, no metal implants, no claustrophobia, and no severe obesity. The study was approved by the Medical Ethics Committee of the Erasmus Medical Centre, Rotterdam, and all patients gave written, informed consent before participating.

Sample Size

To calculate the necessary number of patients, we performed a power analysis on data from the literature. Assuming a clinical relevance of 10% on the Fugl-Meyer motor assessment¹⁹ (FMA) score, a standard deviation of 3.2,²⁰ and a loss of patients at follow-up of 10%, we calculated that 20 patients in each group would be sufficient to have an 80% chance of detecting a statistically significant difference in improvements between the 2 groups.

Study Design

All participants were randomly assigned to either the experimental group receiving mirror therapy (mirror group) or the control group. One of the authors otherwise not involved in the intervention used a computer random number generator to create the randomization sequence and constructed sealed envelopes containing the assignments. Patients received their group allocation assignments after baseline measurements were performed, just before the first training session.

Measurements of upper extremity function were performed before intervention (baseline), right after intervention (posttreatment), and 6 months after intervention (follow-up). All assessments were made by the same investigator, who was blinded to group allocation. Blinding of the patients or the physiotherapist was not possible because of the nature of the therapy. Examinations with fMRI were done in scanning sessions at baseline and posttreatment. Posttreatment clinical measurements and fMRI examinations were performed within a week after the last treatment session.

Intervention

All patients participated in a 6-week training program. Both mirror and control groups performed bimanual exercises, with the difficulty of the exercises depending on the patients’ individual levels of functioning. Exercises were not only based on the Brunnstrom phases of motor recovery but also consisted of functional exercises such as moving objects. The control group had a direct view of both hands, whereas the mirror group practiced with the affected hand positioned behind the mirror while they looked at the reflection of the unaffected hand in the mirror. To ensure that patients focused at the mirror reflection of their unaffected hand instead of their moving unaffected hand itself, a cover was placed over their unaffected hand (Figure 1). Patients practiced at the rehabilitation center once a week under the supervision of a physiotherapist and were instructed to practice 5 times a week, 1 hour per day, at home. Home practice materials consisted of an instruction booklet with photographs and a digital video disk with film fragments of the exercises to be performed. Regular telephone calls were made by the physiotherapist to assure that patients complied with their exercise regimens. Furthermore, patients were instructed to keep detailed accounts of their practice schedules and experiences. These diaries were inspected by the physiotherapist during each training session in the rehabilitation center.

Figure 1.

Setup for mirror therapy

Outcome Measures

As mirror therapy consists of exercises mainly on the level of body function, we chose as primary endpoint the difference in improvement between both treatment groups in motor function as measured with the upper extremity part of the FMA (including arm, wrist, and hand function measurements). Additionally, in the body function ICF domain, we measured grip force with a Jamar handheld dynamometer (Sammons Preston, Bolingbrook, IL), spasticity with the Tardieu scale,²¹ and pain with a visual analog scale ranging from 0 to 100 mm. In the activity domain, we measured motor capacity with the Action Research Arm Test,²² self-perceived performance with the ABILHAND questionnaire,²³ and actual performance in daily life during 24 hours with the Stroke-ULAM.²⁴ The Stroke-ULAM consists of accelerometers placed on both arms and has been described in detail and tested for its validity in an earlier study. Because of the inconvenience of wearing the ULAM, measurements were not performed at 6-month follow-up. The ratio between the amount of use of the unaffected and the affected arms was used as outcome. In the participation domain, we measured quality of life with the EQ-5D.²⁵ To evaluate the effects of mirror therapy at cortical levels, we calculated the difference between groups in the change in activation balance between affected and unaffected hemispheres as measured with fMRI (see the following).

fMRI Experiment

In all, 12 patients from the mirror group and 9 patients from the control group were eligible to participate in this part of the study. During the scanning sessions, patients lay on their backs in the scanner with their upper arms comfortably resting on the scanner table alongside their torsos, with their elbows flexed so that their hands were 20 cm apart above their waists. By means of 2 mirrors attached to the head coil above each patient’s head, patients were able to look in the direction of their feet and view both hands.

The experimental task consisted of 10 alternating 30-second periods of 5 rest and 5 active conditions (block design). In the active periods, patients had to open and close the affected hand; in the rest periods patients had to hold the hand still. Patients were instructed to pace the opening of their hand to a metronome with a rhythm of 0.5 Hz. The onsets of the rest and active conditions were indicated verbally by using simple words (start, rest) generated by a computer program (Matlab version 7.1; Mathworks, Sherborn, MA). Auditory stimuli were presented to the patients through MRI-compatible headphones. The hand movement was practiced before the scan session started.

Imaging was performed on a 3T MR system (HD platform, GE Healthcare, Milwaukee, WI). For anatomical reference, a high-resolution, 3-dimensional, inversion recovery, fast spoiled gradient echo, T1-weighted image was acquired (TR/TE/TI 10.7/2.2/300 ms, 18° flip angle, matrix 416 × 256, and field of view 250 × 175 mm²). For functional imaging, a single-shot, T2*-weighted, gradient echo echo-planar imaging (EPI) sequence was used (TR/TE 3000/30 ms, 75° flip angle, matrix 64 × 96, field of view 220 × 220 mm²). The imaging volume covered the entire brain, including the cerebellum.

Statistical Analyses

Clinical outcome

To test the study hypothesis, we used a generalized estimating equations approach. Under the assumption that missing data are random and not due to group allocation or treatment effect, this model estimates missing data values, thereby allowing the use of data from all participants, irrespective of whether they were measured at all time points. Each outcome measure was used as a separate response variable, and group (mirror vs control) and time (baseline vs posttreatment vs follow-up) were inserted in the model as predictors. The interaction of group × time was used to determine the efficacy of the intervention. Significance was set at .05.

fMRI data

The imaging data were analyzed using statistical parametric mapping software (SPM5; Wellcome Department of Cognitive Neurology, University College London, UK), implemented in Matlab version 7.1 (MathWorks, Natick, MA).

All functional images for each participant were realigned to the first scan of each session and then coregistered to the T1-weighted anatomical scan. Subsequently, images were transformed to standard Montreal Neurologic Institute space. To prevent warping around the lesions, we used a segmentation-based normalization approach.²⁶ Finally, normalized images were spatially smoothed by using a Gaussian filter of 8-mm full width at half maximum.

Preliminary analyses showed that the realignment parameters estimated during spatial preprocessing were sometimes correlated with task design. Therefore, we decided not to model the realignment parameters in the design matrix as regressors of no interest, as this would have resulted in canceling out task-related activation. Instead, we used the ArtRepair Toolbox,^27,28 which evaluates all volumes and detects the ones most affected by movement. Those volumes are repaired by interpolation to avoid side effects in the high-pass filter and then deweighted in the general linear model estimation to maintain unbiased estimates. The experimental block design was convolved with the canonical hemodynamic response function, and the resulting model was estimated using a high-pass filter at 128 seconds to remove low-frequency artifacts.

In the first-level analysis, contrast maps were calculated for the active periods versus rest for each patient and each session separately. In addition, contrasts were calculated for the pretreatment versus posttreatment sessions for each patient. Contrast images from patients with left-sided lesions were flipped about the mid-sagittal plane, so that the affected hemisphere corresponded to the right side of the brain for all patients.

Second-level analyses

We merged both groups and performed a random effect analysis on the baseline contrasts of task versus rest to show the typical activation patterns. To assess the differences between groups following therapy, we concentrated on effects within certain regions of interest (ROIs). Using the Anatomy Toolbox²⁹ and the Anatomical Automatic Labeling Atlas (AAL),³⁰ we defined ROIs for each hemisphere separately for the following areas: primary motor cortex (M1), dorsal premotor cortex (PMd), primary sensory cortex (S1), supplementary motor area (SMA), and cerebellum. M1 was constructed of Brodmann areas 4a and 4p; PMd of Brodmann area 6 excluding SMA and the ventral premotor cortex (area below z = +51); S1 of Brodmann areas 1, 2, and 3. SMA and the cerebellum were integrally derived from the AAL atlas.

These ROIs were used to calculate weighted laterality indexes (wLI), which measure the relative amount of activation between the 2 hemispheres.^31,32 We chose this measure because previous research showed that it correlates with residual clinical deficit and is sensitive to detecting subtle therapy-associated changes in cortical activation.³³ Using this method, we first normalized for intersubject variations in global fMRI signal to reduce the chances of floor or ceiling values. For each condition, a threshold was defined as half the average value of the 5% highest t values within a certain ROI. Subsequently, within that ROI, the sum of the t values of all voxels above this threshold (sum-t) was computed. The wLI was then calculated as

WLI = (\sum t_{ipsi} - \sum t_{CONTRA}) / (\sum t_{IPSI} + \sum t_{contra}),

where the subscripts contra and ipsi, respectively, refer to contralateral and ipsilateral with respect to the side of the lesion. wLI values can range between −1 and 1, with higher values indicating a larger contribution of the affected hemisphere.

Changes in wLI within each ROI for the 2 groups separately were assessed using the Wilcoxon rank sum test. Differences in the amount of change in the wLI between groups before and after treatment were assessed with the Mann–Whitney U test. We chose to use nonparametric statistics because of the risk of ceiling or floor effects (in spite of our use of the wLI instead of LI). Significance was set at .05.

Finally, a correlation analysis (Spearman’s correlation coefficient) was performed to assess the relationship between changes in wLI and changes in FMA scores.

Results

Table 1 presents demographic and clinical characteristics of the 2 treatment groups, showing no differences. Figure 2 shows the flow chart of the study. None of the reported dropouts were because of group allocation or treatment effect. Figures 3 and 4 show the structural brain images of the participants of the fMRI subgroup.

Table 1.

Patient Characteristics

	All Patients		Patients With fMRI
Characteristic	Mirror Group	Control Group	Mirror Group	Control Group
No. of patients	20	20	9	7
Age (years)	55.3 ± 12.0	58.7 ± 13.5	51.9 ± 9.3	59.0 ± 10.4
Time since stroke (years)	4.7 ± 3.6	4.5 ± 2.6	6.2 ± 4.4	5.2 ± 2.5
Gender (male/female)	7/13	13/7	2/7	5/2
Affected side (dom/ndom)	6/14	6/14	2/6	3/4
Type of stroke (infarct/hemorrhage)	14/6	14/6	7/2	6/1
FMA score baseline	39.7 ± 14.1	36.4 ± 14.7	43.7 ± 12.5	36.9 ± 9.1
FMA score posttreatment	43.5 ± 14.0	36.6 ± 14.2	48.2 ± 11.2	37.6 ± 8.5
Location of stroke lesion
Cortical (with or without subcortical)	8	5	6	2
Subcortical	7	5	3	3
Brainstem^a	1	3	—	2
Unknown	4	7	—	—
Bamford classification
LACS	8	8	3	1
PACS	8	7	5	3
TACS	3	2	1	1
POCS	1	3	—	2

Abbreviations: fMRI, functional magnetic resonance imaging; dom, dominant; ndom, nondominant; LACS, lacunar stroke; PACS, partial anterior circulation stroke; TACS, total anterior circulation stroke; POCS, posterior circulation stroke.

Brainstem lesions are located above the corticospinal tract.

Figure 2.

Flow chart

Figure 3.

T1-weighted magnetic resonance imaging scans at the level of maximum infarct volume for each patient of the mirror group

Figure 4.

T1-weighted magnetic resonance imaging scans at the level of maximum infarct volume for each patient of the control group

Clinical Measures

All patients attended all 6 training sessions in the rehabilitation center, and the home-kept diaries showed no differences in total home-based practice time between the groups. All patients kept to their regimens, which resulted in an average home training time of 30 hours.

Analyses followed the intention-to-treat principle. Table 2 shows the estimated marginal means and standard deviations for all measurement tools at baseline, posttreatment, and follow-up as well as the between-group comparisons of the change scores from baseline to posttreatment and from baseline to follow-up. Posttreatment, the mirror group had improved significantly more on the FMA than the control group (P = .04), but this difference was not present at follow-up (P = .53). No other significant effects were found between or within groups.

Table 2.

Mean Scores on the Clinical Outcome Measures and Their Changes Over Time

		Baseline^a	Posttreatment^a	Follow-up^a	Δ Posttreatment (95% CI)^b,d	P Value	Δ Follow-up (95% CI)^c,d	P Value
FMA	Mirror	39.7 ± 14.1	43.5 ± 14.0	41.1 ± 14.9	3.6 ± 1.8	.04	1.5 ± 2.5	.53
	Control	36.4 ± 14.7	36.6 ± 14.2	36.3 ± 16.2	(0.2 to 7.1)		(−3.4 to 6.5)
Grip force (kg)	Mirror	11.2 ± 7.8	12.3 ± 9.9	11.6 ± 5.7	1.2 ± 1.6	.47	0.4 ± 2.4	.86
	Control	15.4 ± 9.7	15.2 ± 8.5	15.3 ± 8.4	(−2.0 to 4.4)		(−4.2 to 5.1)
Tardieu elbow	Mirror	61.0 ± 41.3	70.2 ± 34.8	72.3 ± 37.3	13.1 ± 9.5	.17	12.4 ± 12.0	.30
	Control	78.7 ± 14.3	74.7 ± 30.3	77.9 ± 31.3	(−5.5 to 31.7)		(−11.1 to 35.9)
Tardieu wrist	MirrorControl	20.0 ± 27.0	23.6 ± 25.3	21.4 ± 21.5	−1.8 ± 6.3	.79	4.5 ± 7.8	.57
	Control	30.1 ± 33.0	35.5 ± 28.2	26.9 ± 25.9	(−10.9 to 19.8)		(−10.9 to 19.9)
Pain (VAS [mm])	Mirror	9.3 ± 19.0	8.8 ± 10.8	8.0 ± 12.8	2.6 ± 6.8	.68	−4.0 ± 8.4	.63
	Control	12.3 ± 21.6	9.2 ± 14.1	14.9 ± 25.3	(−9.9 to 15.1)		(−20.4 to 12.5)
ARAT	MirrorControl	23.8 ± 15.8	25.5 ± 17.4	24.6 ± 18.7	1.1 ± 2.0	.58	0.4 ± 2.8	.89
	Control	20.6 ± 17.0	21.1 ± 16.8	20.9 ± 17.6	(−2.9 to 5.1)		(−5.0 to 5.8)
ABILHAND	Mirror	0.97 ± 0.90	1.35 ± 1.23	1.17 ± 1.03	0.35 ± 0.15	.15	−0.04 ± 0.28	.89
	Control	1.10 ± 1.32	1.13 ± 1.42	1.34 ± 1.69	(−0.13 to 0.83)		(−0.59 to 0.51)
Accelerometric proportion^e	Mirror	0.30 ± 0.14	0.29 ± 0.09		−0.01 ± 0.02	.59
	Control	0.28 ± 0.09	0.28 ± 0.13		(−0.05 to 0.03)
EQ-5D	MirrorControl	0.75 ± 0.11	0.76 ± 0.13	0.76 ± 0.20	−0.05 ± 0.05	.38	−0.03 ± 0.06	.63
	Control	0.75 ± 0.18	0.81 ± 0.12	0.79 ± 0.14	(−0.15 to 0.06)		(−0.16 to 0.09)

Abbreviations: Δ indicates change; CI, confidence interval; FMA, Fugl-Meyer assessment; VAS, visual analog scale; ARAT, action research arm test; EQ-5D, EuroQol.

Values are mean ± standard deviation.

Between groups difference of mean change at posttreatment from baseline.

Between groups difference of mean change at follow-up from baseline.

Values are mean ± standard error of the mean.

Accelerometric proportion was measured only at baseline and posttreatment.

Cortical Reorganization

In all, 3 patients who were scanned at baseline did not complete the training program and were not scanned posttreatment. Two other patients were discarded from further analysis because their data sets contained scanner artifacts. The remaining analyses were therefore conducted on data from 9 patients in the mirror group and 7 patients in the control group.

Figure 5 shows the activation maps as calculated in the random effects analysis of the task condition versus the rest condition. In general, in all sessions, activation patterns were in accordance with the expected activation for a hand motor task. Activity was observed bilaterally in the precentral and postcentral gyri (M1 and S1), the medial superior frontal gyrus (SMA), at the junction of the superior frontal sulcus and the precentral sulcus (PMC), and in the cerebellum.

Figure 5.

Activation map of the 2 experimental groups combined at baseline of the task condition versus rest (P < .0001)

Table 3 presents the wLI for each of the ROIs. Within M1, a difference in wLI change between the mirror group and control group was found. The activation in the mirror group shifted toward the affected hemisphere, whereas a small shift in activation toward the unaffected hemisphere was observed in the control group. In the other regions no differences in the wLI were observed between the 2 groups. No significant correlations were observed between pretreatment and posttreatment changes in wLI and FMA scores.

Table 3.

Voxel Count in Affected and Unaffected Hemisphere and Weighted Laterality Indices (wLI) for the Baseline and the Posttreatment fMRI Examinations and the Corresponding Changes for each of the 5 Regions of Interest^a,b

		Baseline		Posttreatment		Unaffected			Affected
Area		Unaffected	Affected	Unaffected	Affected	Change (95% CI)	P _within	P _between	Change (95% CI)	P _within	P _between
M1	Mirror	6268 ± 4052	8010 ± 4250	3743 ± 4321	5353 ± 2810	2525 ± 2849 (335 to 3715)	.04	.05	2658 ± 3554 (−74 to 5390)	.07	.03
	Control	4551 ± 4237	6408 ± 3952	5342 ± 3227	5419 ± 2066	−791 ± 3258 (−3804 to 2222)	.61		990 ± 3473 (−2223 to 4302)	.50
PMd	Mirror	8182 ± 4157	12 081 ± 4132	4061 ± 3859	7178 ± 4105	4121 ± 4433 (713 to 7529)	.05	.15	4903 ± 5354 (788 to 9019)	.04	.43
	Control	6339 ± 5240	9070 ± 6994	5653 ± 3265	7830 ± 6702	686 ± 5231 (−4152 to 5523)	.87		1240 ± 5528 (−3872 to 6353)	.50
SMA	Mirror	7009 ± 4694	10 534 ± 4083	3757 ± 2861	6780 ± 4097	3252 ± 3606 (480-6024)	.03	.19	3754 ± 5382 (−383 to 7891)	.11	.96
	Control	7183 ± 3676	8835 ± 6473	6585 ± 3413	7938 ± 5085	598 ± 4039 (−3183 to 4334)	.87		897 ± 7159 (−5723 to 7518)	.74
S1	Mirror	10895 ± 8048	16 688 ± 9845	6022 ± 5680	9620 ± 5712	4874 ± 6854 (−394 to 10 139)	.11	.26	7068 ± 7890 (1003-13 133)	.04	.49
	Control	8121 ± 5949	13 339 ± 8334	7988 ± 6118	11623 ± 5593	133 ± 8167 (−7421 to 7687)	.74		1716 ± 6113 (−3937 to 7369)	.40
CB	Mirror	5904 ± 4848	6045 ± 7228	8065 ± 10457	4905 ± 5458	−2161 ± 7862 (−8204 to 3882)	.86	.87	1140 ± 6734 (−4036 to 6316)	.52	.96
	Control	8331 ± 7215	10 473 ± 8905	8828 ± 7513	6965 ± 7159	−497 ± 5497 (−5581 to 4587)	.90		5307 ± 9090 (−4899 to 11 915)	.31
Area		Baseline	Posttreatment	Change (95% CI)	P _within	P _between
M1	Mirror	0.11 ± 0.21	0.44 ± 0.50	0.33 ± 0.39 (.03 to .63)	.05	.03
	Control	0.17 ± 0.31	0.10 ± 0.38	−0.07 ± 0.28 (−0.33 to 0.19)	.40
PMd	Mirror	0.21 ± 0.17	0.41 ± 0.38	0.20 ± 0.30 (−0.03 to 0.43)	.11	.43
	Control	0.05 ± 0.50	0.07 ± 0.44	0.02 ± 0.30 (−0.26 to 0.30)	.50
SMA	Mirror	0.24 ± 0.29	0.33 ± 0.30	0.09 ± 0.21 (−0.07 to 0.25)	.17	.96
	Control	0.00 ± 0.30	0.11 ± 0.19	0.11 ± 0.23 (−0.10 to 0.32)	.31
S1	Mirror	0.12 ± 0.47	0.34 ± 0.51	0.22 ± 0.34 (−0.04 to 0.48)	.07	.49
	Control	0.18 ± 0.40	0.19 ± 0.49	0.01 ± 0.33 (−0.30 to 0.32	.99
CB	Mirror	0.16 ± 0.59	0.33 ± 0.53	0.17 ± 0.49 (−0.21 to 0.54)	.52	.96
	Control	-0.05 ± 0.48	0.13 ± 0.64	0.18 ± 0.53 (−0.31 to 0.66)	.31

Abbreviations: M1, primary motor cortex; PMd, dorsal premotor cortex; SMA, supplementary motor area; S1, primary sensory cortex; CB, cerebellum.

P values are estimated using the Wilcoxon (P _within) and the Mann–Whitney U (P _between) Tests.

All values are mean ± standard deviation.

Discussion

Our study has 2 important findings. We showed that in patients with chronic stroke, practicing with a mirror resulted in modest, but statistically significant, improvements in upper extremity motor function. This effect disappeared in the follow-up measurements and did not transfer to the ICF domains activity and participation. In addition, we demons-trated that mirror therapy caused a shift in activation balance M1 toward the lesioned hemisphere, suggesting neural reorganization. As we did not measure cortical activation at the 6-month follow-up, we do not know whether this change persisted.

The results on our clinical outcome measures are in agreement with previous studies on patients with subacute¹¹ and acute¹² stroke. The size of the FMA change in the present study is not very different from that in a study of patients with more acute stroke. Mirror therapy may thus be beneficial at all stages after stroke. The fact that the current study was based on unsupervised, home-based treatment is in this respect promising.

The effects we observed on motor function did not persist at the 6-month follow-up and did not transfer to any other ICF domains. This is in disagreement with the study of Yavuzer et al,¹¹ who reported improved motor function as well as improved activities of daily living also at 6-months’ follow-up. This discrepancy might be because of the more sensitive activities of daily living (ADL) measurement tool used by Yavuzer et al, as well as the difference in time post-onset. Whereas the subacute stroke patients in the study by Yavuzer et al were still in the rehabilitation center, the chronic stroke patients in our study were home based with daily living routines that were already well adjusted to their disabilities. Improvements in motor function may therefore cause less change in these routines. This lack of transfer to the activity and participation domain may also explain the lack of persistence of improvements in motor function: as patients do not involve the affected arm more in a home setting, therapeutic improvements will deteriorate more rapidly, following the “use it or lose it” principle.³⁴

The improvements we found in motor function did not reach the generally accepted clinically relevant level of 10%.³⁵ However, other training studies in chronic stroke patients have reported similarly small improvements, often using more complicated therapeutic strategies.^4,36 Although one could argue that the improvements we found are small and would not have survived a correction for the multiple comparisons (eg, a Bonferroni correction), from a clinical point of view, any improvements in this group are promising, especially considering that this was a short, easily implemented and nonintensive home training program.

A novel aspect of this study was that we evaluated the effect of mirror therapy on cortical organization. Our results suggest that after a period of mirror therapy the hemispheric activation balance shifts toward the affected hemisphere. This is in agreement with previous research showing a similar shift during the recovery following stroke³⁷ and after arm-tracking training.³⁸ At least several adaptation mechanisms occur following stroke, amongst them increased recruitment of the undamaged hemisphere.³⁹ Whether this increased recruitment is beneficial or detrimental remains controversial.⁴⁰ However, several studies have shown that increased activation in contralesional motor areas is associated with worse motor function,^31,41 whereas a more normal distribution of hemispheric activation is usually found in better recovered patients.⁴² In our study, the mirror illusion seems to bring the disturbed activation balance within the hemispheres more toward normality. Our data show that this shift in hemispheric balance was mainly caused by a decrease in contralesional M1 activation, which is also in line with previous research.³¹

We found no correlation between functional gains and changes in hemispheric balance. This may be because of our small sample size and the small changes we found on both measures. Although several studies have reported correlations between recovery of motor function and changes in brain activation patterns,^31,37,42 such relationships are not always found,^43,44 illustrating that the relationship between measures of brain reorganization and behavioral improvement remains complex.^45,46

The question that remains is which mechanism is responsible for the functional improvements and neural changes following mirror therapy. As mirror therapy resulted in larger improvements than the control intervention, the mirror does seem to have an additional effect beyond repetitive task-orientated training.⁴⁷ It is possible that by providing an image of a normal moving hand, the mirror illusion enriches the training environment and increases somatosensory input, thereby inducing excitability of the motor cortex.⁴⁰ Research in healthy subjects using transcranial magnetic stimulation has provided some proof of such a mechanism, although these results are not conclusive.^14,48 Alternatively, the mirror illusion might increase attention to the motor task, which is known to increase cortical activity⁴⁹ and is of importance during motor recovery following stroke.⁴² In literature, mirror therapy is often linked with the mirror neuron system or motor imagery. One imaging study in healthy participants provided some proof for the activation of the mirror neuron system by the mirror illusion, but again these results were not conclusive.¹⁶ Definite evidence for the contribution of specific neural areas to the effects of mirror therapy mirror therapy has to come from patient studies investigating neural activation patterns directly associated with mirror viewing.

The results of the present study should be interpreted with some caution. First, our treatment groups were of only moderate size, and the findings on neural activation changes are based on only a small number of participants. Because of the latter, we were unable to perform a random effects fMRI analysis to make population based interferences. Furthermore, our sample included patients with large lesions, including partial M1 damage. Previous fMRI studies have shown, however, that partial infarction of M1 still allows for functional recovery, accompanied by adaptive functional reorganization within spared M1.⁴⁴ In addition, some of the differences in activation patterns following treatment could be because of adaptation to the scanner or task. Although the overall decrease in activation observed in both hemispheres (Table 3) indicates that some adaptation occurred, it is not likely that changes in wLI are caused by such adaptation.

The generalizability of our results has limitations. First, our sample consisted mainly of nondominant hemisphere stroke. Second, although lesion size and motor deficits of our participants were considerable compared with patients in most fMRI research, we did exclude patients with complete hemiplegia. Whereas small lesions and modest motor deficits are preferable for making strong fMRI inferences, therapies such as mirror therapy may be especially relevant for patients with large motor deficits for whom virtually no treatment is available.⁵⁰

In conclusion, our phase II trial shows that mirror therapy improves motor function in chronic stroke patients more than a control intervention and leads to changes in cortical organization. However, the clinical improvements we found were small, did not persist 6 months after therapy, and were not reflected in ICF domains of activity and participation. A key question in future research is how to augment the effect of mirror therapy. The focus herein should be on identifying the optimal treatment regimens and subpopulations for the effects of mirror therapy to persist and translate to improvements in performance of daily activities and participation.

Footnotes

The author(s) declared no potential conflicts of interest with respect to the authorship and/or publication of this article.

The author(s) disclosed receipt of the following financial support for the research and/or authorship of this article: Fonds NutsOhra [SNO-T-0602-23]; Innovatiefonds Zorgverzekeraars [06-262]; Wetenschappelijk College Fysiotherapie [WU/2007/07] and Hersenstichting Nederland [15F07.54].

References

Jorgensen

Nakayama

Raaschou

Vive-Larsen

Stoier

Olsen

. Outcome and time course of recovery in stroke. Part II: time course of recovery. The Copenhagen Stroke Study. Arch Phys Med Rehabil. 1995;76:406-412.

Duncan

Goldstein

Matchar

Divine

Feussner

. Measurement of motor recovery after stroke. Outcome assessment and sample size requirements. Stroke. 1992;23:1084-1089.

Taub

Miller

Novack

. Technique to improve chronic motor deficit after stroke. Arch Phys Med Rehabil. 1993;74:347-354.

Mehrholz

Platz

Kugler

Pohl

. Electromechanical and robot-assisted arm training for improving arm function and activities of daily living after stroke. Cochrane Database Syst Rev. 2008;(4):CD006876. doi:10.1002/14651858.CD006876.pub2.

Fasoli

Krebs

Hogan

. Robotic technology and stroke rehabilitation: Translating research into practice. Top Stroke Rehabil. 2004;11:11-19.

Ramachandran

Rogers-Ramachandran

Cobb

. Touching the phantom limb. Nature. 1995;377:489-490.

Altschuler

Wisdom

Stone

. Rehabilitation of hemiparesis after stroke with a mirror. Lancet. 1999;353:2035-2036.

Stevens

Stoykov

. Using motor imagery in the rehabilitation of hemiparesis. Arch Phys Med Rehabil. 2003;84:1090-1092.

Rothgangel

Morton

Hout

JWE

van den Beurskens

AJHM

. Phantoms in the brain. Spiegeltherapie bij chronische CVA-patienten: een pilot-study. Ned Tijdschr Fysiotherapie. 2004;114:36-40.

10.

Sathian

Greenspan

Wolf

. Doing it with mirrors: a case study of a novel approach to neurorehabilitation. Neurorehabil Neural Repair. 2000;14:73-76.

11.

Yavuzer

Selles

Sezer

. Mirror therapy improves hand function in subacute stroke: A randomized controlled trial. Arch Phys Med Rehabil. 2008;89:393-398.

12.

Dohle

Pullen

Nakaten

Kust

Rietz

Karbe

. Mirror therapy promotes recovery from severe hemiparesis: a randomized controlled trial. Neurorehabil Neural Repair. 2009;23:209-217.

13.

Ramachandran

Altschuler

. The use of visual feedback, in particular mirror visual feedback, in restoring brain function. Brain. 2009;132:1693-1710.

14.

Garry

Loftus

Summers

. Mirror, mirror on the wall: viewing a mirror reflection of unilateral hand movements facilitates ipsilateral M1 excitability. Exp Brain Res. 2005;163:118-122.

15.

Fukumura

Sugawara

Tanabe

Ushiba

Tomita

. Influence of mirror therapy on human motor cortex. Int J Neurosci. 2007;117:1039-1048.

16.

Matthys

Smits

Van der Geest

. Mirror-induced visual illusion of hand movements: a functional magnetic resonance imaging study. Arch Phys Med Rehabil. 2009;90:675-681.

17.

World Health Organization. International Classification of Functioning, Disability and Health (ICF). Geneva, Switzerland: World Health Organization; 2001.

18.

Sawner

LaVigne

. Brunnstrom’s Movement Therapy in Hemiplegia: A Neurophysiological Approach. Philadelphia, PA: J. B. Lippincott; 1992.

19.

Fugl-Meyer

Jaasko

Leyman

Olsson

Steglind

. The post-stroke hemiplegic patient. 1. A method for evaluation of physical performance. Scand J Rehabil Med. 1975;7:13-31.

20.

van der Lee

Beckerman

Lankhorst

Bouter

. The responsiveness of the action research arm test and the Fugl-Meyer assessment scale in chronic stroke patients. J Rehabil Med. 2001;33:110-113.

21.

Tardieu

Shentoub

Delarue

. Research on a technique for measurement of spasticity [in French]. Rev Neurol (Paris). 1954;91:143-144.

22.

Hsueh

Hsieh

. Responsiveness of two upper extremity function instruments for stroke inpatients receiving rehabilitation. Clin Rehabil. 2002;16:617-624.

23.

Penta

Tesio

Arnould

Zancan

Thonnard

. The ABILHAND questionnaire as a measure of manual ability in chronic stroke patients: Rasch-based validation and relationship to upper limb impairment. Stroke. 2001;32:1627-1634.

24.

de Niet

Bussmann

Ribbers

Stam

. The stroke upper-limb activity monitor: its sensitivity to measure hemiplegic upper-limb activity during daily life. Arch Phys Med Rehabil. 2007;88:1121-1126.

25.

Lamers

McDonnell

Stalmeier

Krabbe

Busschbach

. The Dutch tariff: results and arguments for an effective design for national EQ-5D valuation studies. Health Econ. 2006;15:1121-1132.

26.

Crinion

Ashburner

Leff

Brett

Price

Friston

. Spatial normalization of lesioned brains: performance evaluation and impact on fMRI analyses. Neuroimage. 2007;37:866-875.

27.

Mazaika

Whitfield-Gabrieli

Reiss

. Artifact repair for fMRI data from high motion clinical subjects. Paper presented at: Human Brain Mapping Conference; June 10-14, 2007; Chicago, IL.

28.

Artifact repair software for fMRI data from clinical subjects. http://cibsr.stanford.edu/tools/ArtRepair/ArtRepairTopPage.htm . Accessed September 13, 2010.

29.

Eickhoff

Stephan

Mohlberg

. A new SPM toolbox for combining probabilistic cytoarchitectonic maps and functional imaging data. Neuroimage. 2005;25:1325-1335.

30.

Tzourio-Mazoyer

Landeau

Papathanassiou

. Automated anatomical labeling of activations in SPM using a macroscopic anatomical parcellation of the MNI MRI single-subject brain. Neuroimage. 2002;15:273-289.

31.

Calautti

Naccarato

Jones

. The relationship between motor deficit and hemisphere activation balance after stroke: A 3T fMRI study. Neuroimage. 2007;34:322-331.

32.

Sharma

Simmons

Jones

. Motor imagery after subcortical stroke: a functional magnetic resonance imaging study. Stroke. 2009;40:1315-1324.

33.

Calautti

Baron

. Functional neuroimaging studies of motor recovery after stroke in adults: a review. Stroke. 2003;34:1553-1566.

34.

Sawaki

Butler

Leng

. Constraint-induced movement therapy results in increased motor map area in subjects 3 to 9 months after stroke. Neurorehabil Neural Repair. 2008;22:505-513.

35.

van der Lee

Wagenaar

Lankhorst

Vogelaar

Devillé

Bouter

. Forced use of the upper extremity in chronic stroke patients: results from a single-blind randomized clinical trial. Stroke. 1999;30:2369-2375.

36.

Takahashi

Der-Yeghiaian

Motiwala

Cramer

. Robot-based hand motor therapy after stroke. Brain. 2008;131:425-437.

37.

Askim

Indredavik

Vangberg

Haberg

. Motor network changes associated with successful motor skill relearning after acute ischemic stroke: a longitudinal functional magnetic resonance imaging study. Neurorehabil Neural Repair. 2009;23:295-304.

38.

Carey

Kimberley

Lewis

. Analysis of fMRI and finger tracking training in subjects with chronic stroke. Brain. 2002;125:773-788.

39.

Marshall

Perera

Lazar

Krakauer

Constantine

DeLaPaz

. Evolution of cortical activation during recovery from corticospinal tract infarction. Stroke. 2000;31:656-661.

40.

Ward

Cohen

. Mechanisms underlying recovery of motor function after stroke. Arch Neurol. 2004;61:1844-1848.

41.

Murase

Duque

Mazzocchio

Cohen

. Influence of interhemispheric interactions on motor function in chronic stroke. Ann Neurol. 2004;55:400-409.

42.

Ward

Brown

Thompson

Frackowiak

. Neural correlates of outcome after stroke: a cross-sectional fMRI study. Brain. 2003;126:1430-1448.

43.

Calautti

Leroy

Guincestre

Marie

Baron

. Sequential activation brain mapping after subcortical stroke: changes in hemispheric balance and recovery. Neuroreport. 2001;12:3883-3886.

44.

Dong

Winstein

Albistegui-Dubois

Dobkin

. Evolution of fMRI activation in perilesional primary motor cortex and cerebellum with rehabilitation training-related motor gains after stroke: a pilot study. Neurorehabil Neural Repair. 2007;21:412-428.

45.

Levin

Kleim

Wolf

. What do motor “recovery” and “compensation” mean in patients following stroke? Neurorehabil Neural Repair. 2009;23:313-319.

46.

Buma

Lindeman

Ramsey

Kwakkel

. Functional neuroimaging studies of early upper limb recovery after stroke: a systematic review of the literature. Neurorehabil Neural Repair. 2010;24:589-608.

47.

Kwakkel

Kollen

Wagenaar

. Long term effects of intensity of upper and lower limb training after stroke: a randomised trial. J Neurol Neurosurg Psychiatry. 2002;72:473-479.

48.

Funase

Tabira

Higashi

Liang

Kasai

. Increased corticospinal excitability during direct observation of self-movement and indirect observation with a mirror box. Neurosci Lett. 2007;419:108-112.

49.

Binkofski

Fink

Geyer

. Neural activity in human primary motor cortex areas 4a and 4p is modulated differentially by attention to action. J Neurophysiol. 2002;88:514-519.

50.

Hodics

Cohen

Cramer

. Functional imaging of intervention effects in stroke motor rehabilitation. Arch Phys Med Rehabil. 2006;87(Suppl 2):S36-S42.

Motor Recovery and Cortical Reorganization After Mirror Therapy in Chronic Stroke Patients

Abstract

Keywords

Introduction

Materials and Methods

Participants

Sample Size

Study Design

Intervention

Outcome Measures

fMRI Experiment

Statistical Analyses

Clinical outcome

fMRI data

Second-level analyses

Results

Clinical Measures

Cortical Reorganization

Discussion

Footnotes

References