Neural interface of mirror therapy in chronic stroke patients: A functional magnetic resonance imaging study
Correspondence Address: Source of Support: The research was funded by Department of
Science and Technology (DST), Government of India, Conflict of Interest: None DOI: 10.4103/0028-3886.105188
Source of Support: The research was funded by Department of Science and Technology (DST), Government of India, Conflict of Interest: None
Background: Recovery in stroke is mediated by neural plasticity. Neuro-restorative therapies improve recovery after stroke by promoting repair and function. Mirror neuron system (MNS) has been studied widely in humans in stroke and phantom sensations. Materials and Methods: Study subjects included 20 patients with chronic stroke and 10 healthy controls. Patients had clinical disease-severity scores, functional magnetic resonance imaging (fMRI) and diffuse tensor imaging (DTI) at baseline, 8 and at 24 weeks. Block design with alternate baseline and activation cycles was used with a total of 90 whole brain echo planar imaging (EPI) measurements (timed repetition (TR) = 4520 ms, timed echo (TE) = 44 ms, slices = 31, slice thickness = 4 mm, EPI factor 127, matrix = 128 × 128, FOV = 230 mm). Whole brain T1-weighted images were acquired using 3D sequence (MPRage) with 120 contiguous slices of 1.0 mm thickness. The mirror therapy was aimed via laptop system integrated with web camera, mirroring the movement of the unaffected hand. This therapy was administered for 5 days in a week for 60-90 min for 8 weeks. Results: All the patients showed statistical significant improvement in Fugl Meyer and modified Barthel Index (P < 0.05) whereas the change in Medical Research Council (MRC) power grade was not significant post-therapy (8 weeks). There was an increase in the laterality index (LI) of ipsilesional BA 4 and BA 6 at 8 weeks exhibiting recruitment and focusing principles of neural plasticity. Conclusions: Mirror therapy simulated the "action-observation" hypothesis exhibiting recovery in patients with chronic stroke. Therapy induced cortical reorganization was also observed from our study.
Keywords: Functional neuroimaging, motor learning, neural plasticity, visuomotor system
Stroke is second leading cause of mortality in developing countries. The estimated prevalence of stroke in Asian countries is about 250-300 per 100,000 with a death rate of 1.2%. , Various therapies, aim to improve patient outcomes by promoting repair and restoration of function in the sub-acute or chronic phase, eg. cell and gene therapy, physiotherapy techniques like electromagnetic stimulation, and task-oriented training have been evaluated. , Longitudinal studies suggest that approximately 50% of patients with significant arm paresis recover useful arm function within the first 3 years of stroke. ,
Mirror therapy was first introduced by Ramachandran and co-workers to alleviate phantom limb pain in amputees.  Imagination and observation of a movement forms a source of information that could be useful for motor rehabilitation after stroke with the rationale that brain areas that are normally involved in movement planning and execution are also active during the imagination of a movement.  Mirror therapy is based on the concept of mirror neuron system (MNS). It was first described in monkey ventral pre-motor cortex (Brodmann area 5) and it has been observed that MNS discharges when an animal performs a goal-directed hand action and also when it observes someone else performing a similar action. Their activation leads to recruitment of functionally interconnected cortical structures coupling action execution and observation. , Motor imagery has been proven to be effective in stroke patients with mirror induced-visual feedback facilitating recovery.  It may thus be possible for a system capable of appropriately stimulating the action observation system to encourage plasticity and repair. Since task-oriented rehabilitation is known to be beneficial, it thus suggests that mirror therapy based motor imagery may induce cortical plasticity and promote recovery using goal-directed arm and/or hand movements.  The aim of the study is to evaluate the effectiveness of mirror therapy by a computer-assisted (laptop) webcam system in rehabilitating stroke patients. It also studies cortical reorganization when the patients were subjected to physiotherapy regime.
Twenty chronic stroke (first ever) patients with the following inclusion criteria: Duration 3 months to 2 years; Medical Research Council (MRC) motor power grade at least >2/5 of wrist and hand flexor muscles; Brunnstorm stage between II and IV, National Institute of Health Stroke scale (NIHSS) score between 4 and 15, conscious and able to comprehend, were recruited for the study from the Neurology clinics. The exclusion criteria were: Progressive neurological disorders, unilateral neglect, comorbidities which influence upper extremity usage, contractures, pregnancy, and contraindications to magnetic resonance imaging (MRI). The study was approved by the Institutional Ethics Committee and informed consent was taken from all the subjects. All of them underwent clinical and radiological examination at baseline, 8 and 24 weeks. Ten healthy control subjects also had functional imaging for comparison.
The subjects were examined and assessed by a neurologist and a neurophysiotherapist. Motor function was assessed by NIHSS score, Fugl-Meyer (FM) scale and modified Barthel Index (mBI) at baseline (0 week) and post-physiotherapy (8 weeks). The Edinburgh Handedness Inventory was used to assess the dominance in hand activities. ,
Functional magnetic resonance imaging studies
Blood oxygenation level dependent (BOLD) data were acquired using the echo planar imaging (EPI) sequence using 1.5T MR scanner (Avanto; Siemens Medical Solutions, Erlangen, Germany) with a standard head coil. Block design with alternate baseline and activation cycles was used with a total of 90 whole brain EPI measurements (timed repetition (TR) = 4520 ms, timed echo (TE) = 44 ms, slices = 31, slice thickness = 4 mm). ,
The subjects were asked to perform the motor task with paretic/affected hand, with self-paced (minimum 0.5 Hz) fist clenching/extension of the wrist joint depending upon the extent of motor damage. For healthy controls, the dominant and non-dominant hand motion was used. To control the consistency of rate of motion, and to reduce movement artefacts, the movement was visually monitored using a remote digital camera.
All the patients received physiotherapy by the same therapist for the paretic upper and lower limbs. The treatment regime was administered for 5 days in a week for 8 weeks for 60-90 min. , The treatment incorporated bilateral hand exercises in such a way that the patient observed his unaffected hand on the laptop screen, imagining it to be the affected hand. The movement of the unaffected hand was captured using a web cam and the mirror image of the same motion was projected on the laptop screen to the subject [Figure 1]. This resulted in the facilitation and movement of the paretic hand.  The difficulty of the exercise was dependent on the patient's individual level of functioning. Two similar objects were used for the treatment. For example, two yellow colored soft balls for squeezing, two identical pens to have a prehensile grasp wood blocks to hold in hands.
For clinical data, the results were analysed using Statistical Package for the Social Sciences (SPSS Inc. SPSS for Windows, Version 11.5. Chicago, USA). Parametric and non-parametric t-tests were used for analysis between baseline, 8 and 24 weeks. fMRI data were analysed using SPM2 software (Wellcome Department of Cognitive Neurology, London, UK) running under the MATLAB environment (Mathworks). The functional images were realigned, normalized and then smoothed by a 6-mm Gaussian filter before statistical analysis. Montreal Neurological Institute (MNI) co-ordinates were co-related with Talairach's Atlas More Details for the gray and white matter areas of the brain. The volume of the lesion was analyzed by IMAGEJ (Version 1.42q, Wayne Rasband, National Institute of Health, USA) software.
Of the 20 stroke patients (mean age 45.45 ± 6.6 years, M: F: 18:2) included with the pre-defined inclusion criteria, 13 patients had cortical lesions and 7 had subcortical lesions. The mean time of stroke onset was 8.5 months. The clinical scores at baseline (0 week) and at 8 weeks are given in [Table 1]. The mean FM scale score at baseline was 18.90 ± 7.60 and at 8 weeks 29.45 ± 9.07 (t = −14.36, P = 0.0001). The mean FM scale score at 24 weeks was 35.65 ± 8.5 with statistical significant improvement between 8 and 24 weeks ( t = -8.929, P = 0.0001) and between baseline and 24 weeks ( t = −16.37, P = 0.0001). The mean mBI at baseline, 8 and 24 weeks was 46.95 ± 10.04 and 58 ± 9.3, respectively ( P < 0.05) [Table 1]. The mean mBI score at 24 weeks was 68.4 ± 9.2 showing statistical significant improvement between baseline and 24 weeks' scores. No significant difference was observed in MRC grade scale for power and Brunnstrom stage of stroke recovery. Repeated measures of ANOVA were applied to calculate the difference between 0 (baseline), 8 and 24 weeks which was found to be statistically significant.
We found a significant improvement in the laterality index (LI) of ipsilesional BA 4 and BA 6 in all patients with P value < 0.05 on Wilcoxon sign rank test. [Table 2] shows the number of cluster activation in the ipsilesional and the contra-lesional hemisphere. There was a consistent increase in the cluster activation of the motor and pre-motor Brodmann areas post-therapy (P < 0.05). Ipsilateral or contra-lateral cerebellum was also found to be active in our patients. There was a shift of LI from negative (patient id 4, 9, and 12) to positive at 8 weeks. The inter-rater reliability for all the clinical outcome measures was 0.89 and for the functional imaging parameters was 0.87.
The group analysis of BOLD activation in the right hemispheric stroke at 24 weeks showed right BA 6 activation with cluster counts of 91 voxels, inferior parietal lobule with 90 voxels and at 8 weeks showed a cluster activation of 155 voxels in right BA 4 [Table 3]. During BOLD activation of the left hemispheric stroke at 24 weeks, it was found that right and left cerebellum were active with 430 and 85 voxels, respectively, right BA 6 with cluster counts of 118 voxels and left BA 6 with cluster counts of 180 voxels. At 8 weeks, it was observed that right BA 6 had cluster counts of 90 voxels, left BA 6 with 120 voxels.
BOLD activation results in healthy controls
The mean age of healthy controls (M: F: 7:3) was 44.6 ± 7.8 years. All of them were right handed. A larger activation of the pre-central sulcus or primary motor areas (BA 4) was observed when the dominant hand was active.
Comparison between stroke subjects and healthy controls
One way ANOVA was done to compare controls and patients. [Table 4] shows the number of cluster activated in the right hemispheric in stroke patients and healthy controls. We observed a cluster activation of 26 voxels in right BA 6, 18 voxels in right BA 40 and 17 voxels in left BA 24 in comparison between the two groups. Similar results were observed in the left hemispheric stroke patients and healthy controls [Figure 2].
The first objective in this study was to assess the efficacy of 8 weeks physiotherapy regime in chronic stroke patients. There was significant change in the clinical, FM, and mBI parameters at 8 weeks, suggesting improvement in the motor recovery. The impairment in the Ashworth tone grade and motor power by MRC grading for wrist and hand did not show statistically significant improvement when measured at 8 and 24 weeks. The percentage change in the mean FM score and mBI from baseline to 8 weeks were 55.08% and 23%, respectively. This suggests that a structured exercise regime leads to considerable improvement in the clinical and functional recovery.
Clinical and functional recovery is attributed to reorganization processes in the damaged brain. Within-system reorganization may be possible when damage to a system is partial. However, when a functional system is completely damaged, recovery is achieved largely by a process of substitution, that is, other brain areas are recruited to take over the functions of the damaged areas. , The observations in this study suggest that there was an increase in the activation of primary motor area BA 4 post-therapy explaining the "restitution" principle of neural plasticity.  A shift in the position of the BA 4, 6 was also observed post-therapy, suggesting that physiotherapy in the form of mental imagery promotes a focused activation of the injured brain, augmenting recovery [Figure 3]. ,
The role of cerebellum in the recovery and movement generation following stroke is still being investigated and many varied views have been put forth. All patients in the present study showed an increased cerebellar activation (ipsilateral or contra-lateral to the paretic hand movement) during pre-therapy consistent with the hypothesis that blood flow and the cerebello-cortico network enhance at the beginning of the movement or in performing a new task in diseased subjects whereas a decrease in the post-therapy results suggest that patients had achieved skilled motor performance due to the intense physiotherapy regime. ,
It has been observed that there is an increase in the signal intensity as well an increased change in the LI as measured post-therapy.  These results suggest a focused activation of the perilesional cortex, the plasticity. These observations support the hypothesis that a structured physiotherapy regime improves the outcome measures. ,
The treatment regimens in this study were based on the principles of virtual reality (VR) and motor imagery. , Patients observed the moving hand on the laptop screen as the affected hand and imagined it to be the affected hand (mirror image of the unaffected hand captured by a web cam) and tried to imitate the movement in real environment. The brain areas involved are pre-motor cortex, dorsolateral pre-frontal cortex, and the primary hand motor area as evident by the results in this study. In this therapy, the subject uses two different strategies: (a) Producing a visual representation of the moving limb, in which case the subject is a (third-person) spectator of the movement (visual imagery, [VI]) or (b) mentally simulating the movement, associated with a kinesthetic feeling of the movement, in which case the person is a (first-person) performer of the movement (kinetic imagery). ,
In healthy subjects, ipsilateral cortical activation was more pronounced during left hand motor tasks as compared with the right hand similar to the earlier observations. During the dominant (right) hand movement, pre-central sulcus or the primary motor areas (BA 4) was highly active in healthy subjects. When the BOLD activation pattern were compared with stroke and healthy controls, the ipsilesional hemisphere (left) was active when compared with healthy controls, left BA 6 active with cluster counts of 26 voxels, left BA 40 with 18 voxels. Similar results were observed with the right hemispheric stroke with primary motor cortex active along with the processing areas.
Chemical and anatomic plasticity of the cerebral cortex have been demonstrated in adult animals, suggesting that the neuronal cortical connections can be remodeled by experience, training, and sensory inputs.  Animals reared in complex environments with access to various toys and activities developed more dendritic branching and have higher gene expression for trophic factors than those in standard cages. ,
Both the time of onset of symptoms and the site of cortical lesion play major role in the degree of motor impairment. In the present study, the volume of the lesion varied from a small lacuna (5.8 ml) to a large cortical lesion (45.5 ml). The site and size of the infarct had no correlation with improvement. Patients with <10 ml of lesion showed good recovery as compared with large volume lesion more than 30 ml. Nevertheless a patient with a volume of 21.2 ml had similar clinical scores as compared with a patient with a large volume, 45 ml lesion. It was also observed that lesion at posterior limb of internal capsule (PLIC) had less probability of recovery than the cortical stroke with nearly same volume of lesion.  In this study patients were recruited according to their functional potential i.e, MRC motor power grade of at least 2, NIHSS between 4 and 15, which were contrary to the factors like site and side of lesion, stroke topography, premorbid status, clinical status, the time of onset of symptom, frequency of attack, and acute stroke interventions.
Post-stroke, intensive repetition of exercises leads to learning and improves the motor potential of hemiplegic limb. , Motor imagery activates the primary and pre-motor areas for a task execution. During observation of a movement, areas in the pre-motor cortex become active when the (same) movement is executed. Our treatment involved the basic principles of learning: Practice and active participation. During bilateral hand movements, when the patient observes the mirror image of the unaffected hand on the laptop screen, thinking it to be the affected hand, the patient gets reinforced to move the affected hand in the real world. This kind of cognition and mental activity stimulates the motor processing areas (BA4 and BA6) and showed improvement in the affected hand clinically and functionally. ,
A recent phase II randomized trial evaluated clinical effects and cortical reorganization of home-based mirror therapy in chronic stroke patients. It was observed that post-treatment FM scores improved more in the mirror than in the control group.  Our results are also similar to this study showing greater improvement in the clinical outcome measures at 8 weeks than at 24 weeks when the therapy was withdrawn. The functional and behavioral recovery still needs to be understood to evolve a definite pattern to help the medical and health care professionals to deal with load of stroke patients in India. Also, the post-stroke rehabilitation services are non-structured and have not yet been standardized.
The first author is the motor rehabilitation scientist who administered physiotherapy with motor imagery to all patients. SSK helped in the BOLD activation results, MVP helped in recruitments of patients. RB helped in analysis of results.
[Figure 1], [Figure 2], [Figure 3]
[Table 1], [Table 2], [Table 3], [Table 4]