Effect and mechanism of mirror therapy on lower limb rehabilitation after ischemic stroke: A fMRI study

Abstract

BACKGROUND:

Mirror therapy has been gradually adopted for lower limb rehabilitation, but its efficacy and neural mechanism are not well understood.

OBJECTIVE:

This study aims to investigate the effect and neural mechanism of mirror therapy on lower limb rehabilitation after ischemic stroke by using resting state functional magnetic resonance imaging (rs-fMRI).

METHODS:

A single-blind and randomized controlled pilot study was conducted. 32 patients with ischemic stroke were included in this study and randomly divided into two groups – the control group (CT, n = 16) and the mirror therapy group (MT, n = 16). Both the CT and MT groups received medication and routine rehabilitation training. In addition, mirror therapy was added to the MT group 5 times a week for 30 minutes each time over a period of 3 weeks. Patients’ motor functions, functional connectivity (FC), regional homogeneity (ReHo), and fractional amplitude of low-frequency fluctuations (fALFF) were analyzed both before and immediately after the treatment.

RESULTS:

Patients’ motor functions showed significant improvement in both groups compared to those before treatment (p < 0.01). Moreover, the MT group showed significantly better improvement than the CT group after the treatment (p < 0.05). FC, ReHo and fALFF indicated enhanced neuronal activities in motor function-related brain regions in the MT group compared to the CT group.

CONCLUSION:

Mirror therapy promotes the recovery of lower limb motor functions in patients with ischemic stroke. Through the comparative rs-fMRI analysis, it is found that the mirror therapy promotes the functional reorganization of the injured brain.

Keywords

Ischemic stroke mirror therapy lower limb rehabilitation resting-state functional magnetic resonance imaging (rs-fMRI)

1 Introduction

Ischemic stroke, also known as cerebral infarction, refers to the necrosis or encephalomalacia of brain tissues caused by ischemia or vascular embolism that results in a decrease of blood flow in the brain. According to statistics, about 80% of stroke patients have lower limb motor dysfunction (Lv, Wang, & Shao, 2020). Although many of the patients with hemiplegia post ischemic stroke are able to regain their ability to walk (Langhorne, Bernhardt, & Kwakkel, 2011), they often exhibit gait abnormalities such as strephenopodia. Abnormal gait patterns may cause imbalance and compromise the ability to walk, leading to high energy consumption (MSc, MCSP, & Smith, 2001) and decreased ability to perform activities of daily living, all of which impose a heavy financial burden and mental pressure on patients and their families.

The lower limb rehabilitation of stroke patients with hemiplegia has been primarily relying on traditional functional rehabilitation therapies (Ya-Long, Yan, & Bai-Ya, 2019), such as walking training, kinesiotherapy, functional electrical stimulation, rehabilitation robotics, and ankle foot orthosis (Aliyeh, Mokhtar, & Gholamreza, 2018; Hong et al., 2018; Yu-Rong et al., 2015; Zhang, Yue, Wang, & Kuang, 2017). With advances in rehabilitation theory and technology innovation, rehabilitation professionals are constantly exploring new ways of promoting the recovery of lower limb motor functions after stroke, and mirror therapy has been gradually adopted for this purpose.

Mirror therapy (Altschuler et al., 1999), also known as mirror visual feedback (MVF), is a rehabilitation therapy used to treat unilateral disability, especially the disability of limbs. It is conducted by placing a mirror vertically along the sagittal plane in front of the patient. The reflection of the contralateral limb in the mirror overlaps with the affected limb, thus establishing a visual illusion of two intact limbs to promote the recovery of motor functions of the affected side. In the past, mirror therapy was primarily used for the rehabilitation of upper limbs (Kim, Lee, Kim, Lee, & Kim, 2016; Minjae & Hyun-mo, 2018; Thieme et al., 2019). But more and more attempts have been made to apply mirror therapy to lower limb rehabilitation after stroke, and it has led to good clinical outcomes (Ji & Kim, 2015; Kenji et al., 2015; Lee, Kim, & Lee, 2017; Salem & Huang, 2015; Sutbeyaz, Yavuzer, Sezer, & Koseoglu, 2007; Uthra et al., 2013). Mirror therapy is a low-cost treatment that is easy to implement, which is well-suited for long-term intervention. Furthermore, mirror therapy is beneficial for severe hemiparetic patients whose conditions are too severe to complete tasks of other treatments that require patients to retain certain degree of walking ability, whereas mirror therapy can be conducted while sitting or lying down if patients cannot walk. Therefore, mirror therapy is worthwhile and potentially useful for the rehabilitation of lower limb functions such as motion, balance, transfer, and activity of daily living. However, compared to its wide adoption in the upper limb rehabilitation, the research and clinical adoption of mirror therapy in the lower limb rehabilitation are relatively lacking, and the neural mechanism of mirror therapy in the lower limb rehabilitation is not well understood either.

Functional magnetic resonant imaging (fMRI) is one of the most commonly used tools to study brain functions and neural mechanisms because of its unique non-invasive, atraumatic, and high-resolution imaging capability. There are two main forms of fMRI – task-based fMRI and resting-state fMRI (rs-fMRI). Task-based fMRI examines the brain activities induced by tasks during the imaging session. Compared to task-based fMRI, rs-fMRI focuses more on the intrinsic interrelationship between functional areas in the brain in the task-free resting state. Several studies looked into the effect of mirror therapy by using fMRI (E, Marion, et al., 2011; E, W, et al., 2011; Farsin et al., 2012; Fritzsch et al., 2014; Jing et al., 2013; Koen et al., 2009; Rjosk et al., 2017; Wang et al., 2013), but few of them explained the neural mechanism of mirror therapy based on rs-fMRI (Rjosk et al., 2017).

We hypothesize that the mirror therapy could improve lower limb functions such as motion, balance, transfer and activity of daily living for stroke patients, and such effects are due to enhanced brain functional connectivity and activity. There are two main objectives in current study. The first is to validate the effect of mirror therapy on the lower limb rehabilitation after stroke and compare it with conventional rehabilitation therapies. The second is to reveal the possible neural mechanism of mirror therapy for lower limb rehabilitation by rs-fMRI.

2 Methods

2.1 Participants

32 patients with ischemic stroke who were treated at the Department of Rehabilitation of Sichuan Provincial People’s Hospital from March 2016 to June 2017 were recruited and randomly divided into the control group (CT) and the mirror therapy group (MT) with 16 patients in each group.

Inclusion criteria:

Patients experienced a first-ever ischemic stroke with lesions limited to one hemisphere, and the symptoms met the diagnostic criteria stated in the “Guidelines for the diagnostics and treatment of acute ischemic stroke in China” set by the Neurology Subcommittee of the Chinese Medical Association in 2014. All patients were diagnosed with ischemic stroke by head CT or MRI.

Patients were in stable conditions. When the patients were enrolled in the study, they were within 30 days from the onset of ischemic stroke.

Patients exhibited hemiplegia; Modified Ashworth Scale for lower extremity was not higher than 2; Brunnstrom score for the lower extremity (BRS-LE) was between I and IV.

Patients showed no cognitive impairment that would affect their ability to cooperate with the treatment. Their Mini Mental State Examination (MMSE) score was greater than 23.

Patients could keep static balance in the sitting position.

Patients were right-handed.

Exclusion criteria:

Patients showed unstable vital signs.

Patients had a history of cerebrovascular diseases with sequelae that impaired neural or motor functions.

Patients had a history of epilepsy, dementia, depression or other conditions that may compromise the brain function.

Patients had psychological conditions, cognitive impairment and other medical conditions that would affect the patients’ ability to comply with the study protocol.

Patients had metal implants or other medical conditions that are unsuitable for MRI examination.

Patients had impaired vision.

2.2 Study design

Our study used a single-blind and randomized controlled design. After informed consent were obtained, all patients were randomly assigned to either the MT group or CT group. Both groups received medication and routine rehabilitation training. The CT group received routine rehabilitation, including good limb positioning, maintenance and improvement of joint mobility, control of muscle tension, promotion of active movement, transfer training, balance training, gait training, occupational therapy and traditional Chinese medicine rehabilitation such as acupuncture. In addition to these routine rehabilitation therapies, mirror therapy was added to train the motor functions of lower limbs in the MT group 5 times a week for 30 minutes each time over a period of 3 weeks. A quiet environment was selected during the mirror therapy session. Patients were seated in a stable chair, and a mirror of 85×189 cm×cm was placed in the front of the patients along the sagittal plane. Patients’ legs were located on both sides of the mirror. The non-paretic limb was placed on the reflective side of the mirror to generate a reflected image for visual feedback, and the paretic limb was blocked by the mirror (Fig. 1). During mirror therapy, the patients were given instructions to perform the same action on both lower limbs at the same time. They were instructed to imagine the reflected image of the non-paretic limb in the mirror as their own on the affected side and actively move the paretic limb as much as possible. If the paretic limb was unable to complete the instructed movement, the therapist could assist the movement of the limb behind the mirror. Patients were asked to complete five sets of movement, including the internal and external rotation of hip joint, adduction and abduction of hip joint, flexion and extension of knee joint, dorsiflexion and plantar flexion of ankle joint, and varus and valgus of ankle joint, with each movement reaching the maximum range of the joint motion.

Fig. 1

Setup for mirror therapy of lower limbs. The left side is the unaffected limb, and the unaffected limb move in front of the reflective side of the mirror.

2.3 Evaluation

2.3.1 Motor function assessment

2.3.1.1 Primary outcome

Fugl-Meyer assessment for lower extremity (FMA-LE) was used to quantify the recovery and coordination of lower limb motor functions. This assessment included 17 items with a maximum score of 34, and a higher score indicated better motor functions. FMA-LE has been proved to have intra-rater reliability (r = 0.99) and inter-rater reliability (r = 0.94) (Sanford, Moreland, Swanson, Stratford, & Gowland, 1993).

2.3.1.2 Secondary outcomes

Berg’s balance scale (BBS) was used to evaluate the balance function. This assessment consisted of 14 activities related to balance function with a total score of 56, and a higher score indicated better balance function. This scale has high intra-rater reliability (r = 0.98) and inter-rater reliability (r = 0.97) (Downs, Marquez, & Chiarelli, 2013; K, S, & Ji, 1995). Modified Rivermead mobility index (MRMI) was used to evaluate the mobility change. The MRMI score ranged from 0 to 40, and a higher score indicated a better mobility. The scale has been shown to be highly reliable between raters (ICC = 0.98) and have high internal consistency (Cronbach’s alpha = 0.93) (Lennon & Johnson, 2000).Modified Barthel index (MBI) was used to assess the activities of daily living. The MBI score ranged from 0 to 100, and a higher score indicated better self-care abilities. The MBI shows excellent test-retest reliability with high ICC = 0.90– 0.96 (Yang, Wang, Lee, Chen, & Hsieh, 2020). Brunnstrom score for the lower extremity (BRS-LE) was used to assess the baseline motor function. It consisted of six stages from I to VI, representing the stage from no voluntary movement to near normal movement. BRS has good clinical metrology characteristics, which helps to quantify the degree of motor impairment after stroke and the difference of motor function in patients after stroke (Huang et al., 2016; Safaz, Ylmaz, Yaşar, & Alaca, 2009).

2.3.2 Resting-state Functional Magnetic Resonant Imaging (rs-fMRI)

All the rs-fMRI data in this study were processed on MATLAB 2013a. First, the images were flipped if the lesions were in the right hemisphere, After the flipping, all lesions were in the left hemisphere in the images. Consequently, the left hemisphere was defined as the affected hemisphere, and the right hemisphere as the unaffected hemisphere. The DPARSF software package based on SPM8 was then used to preprocess the fMRI data according to the following standard procedure (Chao-Gan & Yu-Feng, 2010; Mekbib et al., 2020; Yin et al., 2020):

The first 5 volumes were removed from each time series to account for the initial magnetization instability and the time required for the patients to adapt to the scanning. The remaining 155 EPI images were further processed and analyzed.

Slice time correction was applied to the images to make the acquisition time of each layer consistent in a TR time, i.e. all images collected in a TR time started at the same time point.

The subjects whose average movement was less than 1 mm and rotation less than 1° in all directions (translation in X, Y, and Z, and rotation around these three directions) were used.

The corrected images were spatially normalized to the Montreal Neurological Institute (MNI) space and resampled to 3×3×3 mm voxels using a segmentation-based normalization approach estimated during unified segmentation.

In order to reduce the remaining differences between individuals after standardization, improve the signal-to-noise ratio, and make up for the statistical deviation caused by uneven registration, the resulting images were spatially smoothened with a Gaussian kernel of 6×6×6 mm³ (full-width half-maximum, FWHM). Regional homogeneity (ReHo) was not smoothened to avoid errors.

The linear drift of the spatially normalized image was removed to reduce the baseline drift caused by the thermal noise and preamplifier of the tested coil.

The time series for each voxel was band-pass filtered (0.01– 0.08 HZ) to reduce the interference of very low-frequency drift and high-frequency noise. fALFF was not processed by filtering.

The linear regression method was used to remove the effect due to head movements, white matter signals, and cerebrospinal fluid on the low-frequency synchronous oscillation signals.

2.3.3 Statistical analysis

All statistical analyses were performed by using the IBM SPSS software (version 20.0). All data were tested for homogeneity of variance, and the Shapiro-Wilk test was used to assess the normal distribution. If the variance was homogeneous and conformed to the normal distribution, the data were compared by 2×2 ANOVA test. If the variance was uneven or did not conform to the normal distribution, the Wilcoxon rank-sum test or chi-square test was used for comparison between groups, and the Wilcoxon signed-rank test was used for comparison within groups. Based on meeting statistical assumptions, the Wilcoxon rank-sum test was used to assess differences between MT and CT groups after the treatment in lower extremity impairment level as measured by the FMA-LE, and the 2×2 ANOVA test was used to compare the FMA-LE differences from pre- to post-treatment. Quantitative data were presented as means±SD or the median (interquartile range (IQR)). Qualitative data were presented as frequencies (percentages). A p value of < 0.05 was considered statistically significant for all comparisons.

The REST software package was used to compare the functional connectivity (FC) value, regional homogeneity (ReHo) value, and fractional amplitude of low-frequency fluctuations (fALFF) value of the two groups by using unpaired two-sample t-test. The brain region with p < 0.05 was considered the area with significant difference. Alphasim was used for correction. The corrected threshold was p < 0.05, cluster size≥60 voxels, and the single voxel threshold was p < 0.01.

3 Results

A total of 48 patients were recruited at the beginning of the study, but eight did not meet the inclusion criteria or declined to participate in the study. The rest 40 patients were randomly divided into two groups (CT, n = 20; MT, n = 20;). Eight patients dropped out from the study. Out of the eight dropouts, four were discharged from the hospital and quit the study, three refused or unable to complete the intervention, and one had failed rs-fMRI scan. At the end of the study, the results of 32 patients (MT:16 and CT:16) were used for evaluation. The flow chartof this study is shown in Fig. 2.

Fig. 2

Flowchart of the study design.

There was no significant difference in demographic and clinical characteristics, such as age, gender, side of lesion, time since stroke and baseline motor functions, between the two groups (p > 0.05) as shown in Table 1.

Table 1

Patient characteristics

	CT (n = 16)	MT (n = 16)	Between-group comparisons
Age (Year, Mean±SD)	58.5±11.15	61.5±9.93	F = 0.646
			p = 0.428^a
Sex
Female	8 (50%)	7 (43.75%)	χ² = 0.723
			p = 1.0^b
Male	8 (50%)	9 (56.25%)
Side of lesion
Right	7 (43.75%)	5 (31.25%)	χ² = 0.533
			p = 0.716^b
Left	9 (56.25%)	11 (68.75%)
Time since stroke (Day, mean±SD)	20.06±4.42	21.38±5.19	F = 0.593
			p = 0.447^a
Brunnnstrom recovery stage of low extremity	Z = – 0.168
			p = 0.897^c
(BRS-LE)	Stage I:1 (6.25%)	Stage I:2 (12.5%)
	Stage II:4 (25%)	Stage II:4 (25%)
	Stage III:10 (62.5%)	Stage III:8 (50%)
	Stage IV:1 (6.25%)	Stage IV:2 (12.5%)
FMA-LE	11.31±6.37	10.06±6.64	F = 0.295
			p = 0.591^a
MRMI	12.88±5.63	12.19±5.43	F = 0.124
			p = 0.728^a
BBS	7.50 (3.25,14.75)	8.00 (4.00,16.50)	Z = – 0.190
			p = 0.867^c
MBI	22.50 (10.00,27.75)	21.50 (20.00,25.75)	Z = – 0.133
			p = 0.897^c

a, 2×2 ANOVA test; b, Chi-square test; c, Wilcoxon rank-sum test; CT – control group; MT – mirror therapy group; FMA-LE – Fugl-Meyer Assessment for Lower Extremity; BBS – Berg Balance Scale; MRMI – Modified Rivermead Mobility Index; MBI – Modified Barthel Index.

3.1 Assessment of lower limb function

3.1.1 Primary outcome

Results of 2×2 ANOVA test showed no significant difference in FMA-LE scores before treatment between groups (F = 0.295, p = 0.591) (Table 1), but FMA-LE scores in both groups improved significantly from pre- to post-treatment (CT: F = 11.628, p = 0.002; MT: F = 24.845, P < 0.01) (Table 2).In addition, Wilcoxon rank-sum test was used to analyze the mean change of FMA-LE between the two groups after treatment, and the results showed a better effect of treatment for MT group (Z = – 4.526, p < 0.01) (Table 2).

Table 2
Comparison of outcomes within group and between groups

CT (n = 16) Within group MT (n = 16) Within group Mean change Between groups

Pre-treatment Post-treatment Pre-treatment Post-treatment CT (n = 16) MT (n = 16)

FMA-LE 11.31±6.37 17.94±5.74 F = 11.628 10.06±6.64 22.44±6.51 F = 24.845 6.50(5.00,8.00) 12.00(10.00,14.00) Z = – 4.526

p < 0.01^a p < 0.01^a p < 0.01^c

BBS 10.69±8.62 21.75±10.46 F = 10.894 8.00(4.00,16.50) 28.50(22.50,42.75) Z = – 3.520 10.69±8.62 21.75±10.46 F = 36.985

p < 0.01^a p < 0.01^b p < 0.01^d

MRMI 12.88±5.63 25.19±7.26 F = 19.789 12.19±5.43 30.75±5.25 F = 59.177 12.88±5.63 25.19±7.26 F = 27.171

p < 0.01^a p < 0.01^a p < 0.01^d

MBI 22.50(10.00,27.75) 40.50(36.25,50.75) Z = – 3.517 20.50±8.78 53.13±10.76 F = 45.630 20.25±12.22 43.75±14.25 F = 9.830

p < 0.01^b p < 0.01^a p < 0.01^d

	CT (n = 16)	Within group	MT (n = 16)	Within group	Mean change	Between groups
FMA-LE	11.31±6.37	17.94±5.74	F = 11.628	10.06±6.64	22.44±6.51	F = 24.845	6.50(5.00,8.00)	12.00(10.00,14.00)	Z = – 4.526
			p < 0.01^a			p < 0.01^a			p < 0.01^c
BBS	10.69±8.62	21.75±10.46	F = 10.894	8.00(4.00,16.50)	28.50(22.50,42.75)	Z = – 3.520	10.69±8.62	21.75±10.46	F = 36.985
			p < 0.01^a			p < 0.01^b			p < 0.01^d
MRMI	12.88±5.63	25.19±7.26	F = 19.789	12.19±5.43	30.75±5.25	F = 59.177	12.88±5.63	25.19±7.26	F = 27.171
			p < 0.01^a			p < 0.01^a			p < 0.01^d
MBI	22.50(10.00,27.75)	40.50(36.25,50.75)	Z = – 3.517	20.50±8.78	53.13±10.76	F = 45.630	20.25±12.22	43.75±14.25	F = 9.830
			p < 0.01^b			p < 0.01^a			p < 0.01^d

Before treatment, there was no significant difference (p > 0.05) between the CT and MT groups. After the treatment, there were differences between the groups (p < 0.05). The significance level was set at p < 0.05 for differences between the two groups. CT– control group; MT– mirror therapy group; FMA-LE – Fugl-Meyer Assessment-Lower Extremity; BBS– Berg balance scale; MRMI– modified Rivermead mobility index; MBI– modified Barthel index. a, 2×2 ANOVA test; b, Wilcoxon signed-rank test; c, Wilcoxon rank-sum test.

3.1.2 Secondary outcomes

Results of Wilcoxon rank-sum test showed no significant difference in BBS scores before treatment between groups (Z = – 0.190, p = 0.867) (Table 1), but results of 2×2 ANOVA test in CT group (F = 10.894, p = 0.002) and Wilcoxon signed-rank test in MT group (Z = – 3.520, p < 0.01) showed that BBS balance function scores were significantly improved from pre- to post- treatment (Table 2). Moreover, 2×2 ANOVA test was used to analyze the mean change of BBS between the two groups after treatment, and the results showed a better effect of treatment for MT group (F = 36.985, p < 0.01) (Table 2).

Results of 2×2 ANOVA test showed no significant difference in MRMI scores before treatment between groups (F = 0.124, p = 0.728) (Table 1), but MRMI scores in both groups improved significantly from pre- to post-treatment (CT:F = 19.789, p< 0.01; MT:F = 59.177, p < 0.01) (Table 2). And the results of the mean change of MRMI scores showed a better effect of treatment for MT group after the treatment (F = 27.171, p < 0.01) (Table 2).

Results of Wilcoxon rank-sum test showed no significant difference in MBI scores before treatment between groups (Z = – 0.133, p = 0.897) (Table 1), but results of Wilcoxon signed-rank test in CT group (Z = – 3.517, p < 0.01) and 2 × 2 ANOVA test in MT group (F = 45.630, p < 0.01) showed that MBI scores were significantly improved from pre- to post-treatment (Table 2). Moreover, 2 × 2 ANOVA test was used to analyze the mean change of MBI scores between the two groups after treatment, and the results showed a better effect of treatment for MT group (F = 9.830, p = 0.004) (Table 2).

3.2 Functional Connectivity (FC)

The M1 region of the lesioned hemisphere was selected as the region of interest (ROI). A spherical ROI with its center at MNI coordinates (-38, -22, 56) (Xiong, Parsons, Gao, & Fox, 1999) and a radius of 5 mm was used. After the treatment, the global FC values of the two groups were calculated and compared. Compared with the CT group, patients in the MT group showed significantly stronger FC values after the treatment in the left paracentral lobule, superior frontal gyrus, left postcentral gyrus, left parietal gyrus, thalamus, and left parahippocampal gyrus. The middle frontal gyrus exhibited decreased FC value. (Table 3 and Fig. 3A)

Table 3
Difference of FC, ReHo and fALFF between the CT group and the MT group after treatment

Cluster size (voxel) Anatomical region BA MNI coordinates Peak intensity p value

FC

Increased FC

60 Superior Frontal Gyrus BA6 [27 60 12] 5.18 p < 0.05

37 Paracentral_Lobule_L (aal) BA6

50 Postcentral_L (aal) BA3 [– 27 – 51 72] 4.252 p < 0.05

31 Parietal_Sup_L (aal) BA5

38 Halamus / [– 24 30 12] 3.818 p < 0.05

87 Parahippocampa Gyrus BA34

BA27 [– 30 – 15 – 27] 4.376 p < 0.05

74 ParaHippocampal_L (aal) BA36

Decreased FC

49 Middle Frontal Gyrus BA46 [33 – 3 42] – 5.073 p < 0.05

ReHo

Increased ReHo

172 Superior Frontal Gyrus BA4

81 Precentral_L (aal) BA6 [– 39 39 33] 5.941 p < 0.05

46 Precentral_R (aal) BA6 [18 – 21 72] 3.816 p < 0.05

59 Postcentral_L (aal) BA2,3

43 Inferior Parietal Lobule BA40 [– 33 – 51 57] 6.802 p < 0.05

47 Cerebellum / [12 – 30 – 33] 6.571 p < 0.05

40 Cingulate Gyrus BA34 [21 – 39 27] 5.047 p < 0.05

44 Precuneus_L (aal) BA7

35 Occipital_Mid_L (aal) BA18 [– 36 – 84 33] 3.849 p < 0.05

Decreased ReHo

57 Middle Temporal Gyrus BA22 [51 – 60 9] – 3.455 p < 0.05

52 Medial Frontal Gyrus BA10 [– 18 45 9] – 5.22 p < 0.05

76 SupraMarginal_R (aal) BA39 [42 – 30 39] – 4.149 p < 0.05

fALFF

Increased fALFF

120 Superior Frontal Gyrus BA6,8 [– 15 18 57] 4.6906 p < 0.05

93 Frontal_Mid_L (aal) BA46

77 Precentral_L (aal) BA4 [27 – 9 60] 6.703 p < 0.05

59 Postcentral Gyrus BA3 [33 – 51 57] 6.802 p < 0.05

79 Cerebellum Anterior Lobe /

78 Culmen / [0 – 48 – 30] 5.9514 p < 0.05

68 Supp_Motor_Area_L (aal) BA6

66 Paracentral_Lobule_L (aal) BA6 [3 – 6 48] 5.9612 p < 0.05

40 Precuneus BA5 [– 36 – 81 39] 5.1986 p < 0.05

Decreased fALFF

45 Inferior Frontal Gyrus BA44 [45 30 6] – 4.468 p < 0.05

55 Cerebellum Posterior Lobe BA37 [– 33– 48 – 42] – 6.164 p < 0.05

76 SupraMarginal_R (aal) BA40 [42 – 30 39] – 6.164 p < 0.05

41 Temporal_Inf_R (aal) BA19 [27 – 42 39] – 5.795 p < 0.05

Cluster size (voxel)	Anatomical region	BA	MNI coordinates	Peak intensity	p value
FC
Increased FC
60	Superior Frontal Gyrus	BA6	[27 60 12]	5.18	p < 0.05
37	Paracentral_Lobule_L (aal)	BA6
50	Postcentral_L (aal)	BA3	[– 27 – 51 72]	4.252	p < 0.05
31	Parietal_Sup_L (aal)	BA5
38	Halamus	/	[– 24 30 12]	3.818	p < 0.05
87	Parahippocampa Gyrus	BA34
BA27	[– 30 – 15 – 27]	4.376	p < 0.05
74	ParaHippocampal_L (aal)	BA36
Decreased FC
49	Middle Frontal Gyrus	BA46	[33 – 3 42]	– 5.073	p < 0.05
ReHo
Increased ReHo
172	Superior Frontal Gyrus	BA4
81	Precentral_L (aal)	BA6	[– 39 39 33]	5.941	p < 0.05
46	Precentral_R (aal)	BA6	[18 – 21 72]	3.816	p < 0.05
59	Postcentral_L (aal)	BA2,3
43	Inferior Parietal Lobule	BA40	[– 33 – 51 57]	6.802	p < 0.05
47	Cerebellum	/	[12 – 30 – 33]	6.571	p < 0.05
40	Cingulate Gyrus	BA34	[21 – 39 27]	5.047	p < 0.05
44	Precuneus_L (aal)	BA7
35	Occipital_Mid_L (aal)	BA18	[– 36 – 84 33]	3.849	p < 0.05
Decreased ReHo
57	Middle Temporal Gyrus	BA22	[51 – 60 9]	– 3.455	p < 0.05
52	Medial Frontal Gyrus	BA10	[– 18 45 9]	– 5.22	p < 0.05
76	SupraMarginal_R (aal)	BA39	[42 – 30 39]	– 4.149	p < 0.05
fALFF
Increased fALFF
120	Superior Frontal Gyrus	BA6,8	[– 15 18 57]	4.6906	p < 0.05
93	Frontal_Mid_L (aal)	BA46
77	Precentral_L (aal)	BA4	[27 – 9 60]	6.703	p < 0.05
59	Postcentral Gyrus	BA3	[33 – 51 57]	6.802	p < 0.05
79	Cerebellum Anterior Lobe	/
78	Culmen	/	[0 – 48 – 30]	5.9514	p < 0.05
68	Supp_Motor_Area_L (aal)	BA6
66	Paracentral_Lobule_L (aal)	BA6	[3 – 6 48]	5.9612	p < 0.05
40	Precuneus	BA5	[– 36 – 81 39]	5.1986	p < 0.05
Decreased fALFF
45	Inferior Frontal Gyrus	BA44	[45 30 6]	– 4.468	p < 0.05
55	Cerebellum Posterior Lobe	BA37	[– 33– 48 – 42]	– 6.164	p < 0.05
76	SupraMarginal_R (aal)	BA40	[42 – 30 39]	– 6.164	p < 0.05
41	Temporal_Inf_R (aal)	BA19	[27 – 42 39]	– 5.795	p < 0.05

FC – Functional connectivity; ReHo – Regional Homogeneity; fALFF – Fractional amplitude of low-frequency fluctuations; BA – Brodmann area; MNI – Montreal Neurological Institute.

Fig. 3

fMRI data analysis and comparison between the CT and MT groups after treatment. a) Differences of FC between the CT and MT groups after treatment. (Unpaired two-sample t-test, p < 0.05, AlphaSim correction). The yellow areas are the regions of increased FC value, and the blue areas are the regions of decreased FC value. B) Differences of ReHo between CT and MT groups after treatment. (Unpaired two-sample t-test, p < 0.05, AlphaSim correction). The yellow areas are the regions of increased ReHo value, and the blue areas are the regions of decreased ReHo value. c) Differences of fALFF between CT and MT groups after treatment. (Unpaired two-sample t-test, p < 0.05, AlphaSim correction). The yellow areas are the regions of increased fALFF value, and the blue areas are the regions of decreased fALFF value.

3.3 Regional Homogeneity (ReHo)

Compared to the CT group, the MT group showed significantly increased ReHo values after the treatment in the precentral gyrus, superior frontal gyrus, cerebellum, inferior parietal lobule, left postcentral gyrus, middle occipital gyrus, cingulate gyrus, and left precuneus. The brain regions with decreased ReHo value were middle temporal gyrus, medial frontal gyrus, and superior marginal gyrus (Table 3 and Fig. 3B).

3.4 Fractional Amplitude of Low-Frequency Fluctuations (fALFF)

Compared to the CT group, patients in the MT group exhibited increased fALFF after the treatment in the left supplementary motor area (SMA), left paracentral lobule, left precentral gyrus, superior frontal gyrus, cerebellum anterior lobe, postcentral gyrus, and precuneus. Brain regions with decreased fALFF included the right Inferior frontal gyrus, inferior occipital gyrus, and inferior temporal gyrus (Table 3 and Fig. 3C).

4 Discussion

Mirror therapy creates an illusion that enables visual feedback and virtual reality. When combined with movement observation, imitation learning, motor imagery and other rehabilitation training techniques, mirror therapy leads to desired clinical outcomes (Altschuler et al., 1999). Mirror therapy has been widely applied to the treatment of upper limb motor dysfunction after stroke and has achieved good therapeutic effects. In recently years, more and more studies have been exploring the curative effect of mirror therapy for lower limb rehabilitation. Arya et al. (Arya, Pandian, & Kumar, 2019) found that the FMA score and Rivermead visual gait assessment (RVGA) score of the lower limb in patients with chronic stroke showed more significant changes after mirror therapy training compared to the control group. Ji et al.¹⁰ came to a conclusion that mirror therapy was a feasible intervention to improve the lower limb functions of stroke patients after treating 34 patients with subacute stroke using mirror therapy and virtual reaction. The work by Ghadir and other groups (Ghadir, Rifaii, & Soha, 2016; Sutbeyaz et al., 2007) also confirmed the efficacy of mirror therapy on the recovery of lower limb functions after stroke. Our study shows that the FMA-LE score, BBS score, MBI score, and MRMI score are improved after the treatment in both the MT group and CT group. The MT group shows a more significant improvement than the control group (p < 0.01). Our results show that mirror therapy is an effective rehabilitation treatment of lower limb functions in patients with hemiplegia after ischemic stroke, which agrees with the prior works mentioned above.

Although many studies (Holm et al., 2018) proved that mirror therapy could improve motor functions of limbs in rehabilitation, the underlying neural mechanisms of mirror therapy are seldom studied. Up to now, only a couple of possible hypotheses have been proposed. The first hypothesis is related to the mirror neuron system (MNS) (Gunes et al., 2008; Ros, eacute, ouml, & Lundborg, 2005; Sutbeyaz et al., 2007). MNS stimulates the understanding and imitation of the motivation, judgment, behavioral intention, and actions when an individual observes the same action (Rizzolatti, 2005). MNS is closely related to human observation, imitation and motor learning, and is believed to be one important mechanism of mirror therapy by playing a role in action recognition and imitation (Giovanni, Ana, & L, 2006). The rs-fMRI results in this study reveal that the degree of FC in the superior frontal gyrus, left middle frontal gyrus, left supplementary motor area and other brain areas in the MT group is significantly enhanced compared to the CT group, and the synchronization (ReHo) of activities in the superior frontal gyrus, precentral gyrus, inferior parietal lobule, ipsilateral postcentral gyrus, precuneus and other brain areas is also enhanced in the MT group (p < 0.05). Most of these areas with enhanced FC and synchronization are located in brain regions where mirror neurons are located (Giovanni et al., 2006; Rizzolatti, 2005). It has been reported that observing a reflected image in the mirror may trigger the MNS (Gunes et al., 2008). During the mirror therapy, the mirror neurons are stimulated by the illusion in the mirror and directly participate in the imitation and understanding of actions, which may improve the efficiency of information transmission and processing in related brain regions, increase the synchronization of neuronal activities, promote the reorganization and repair of motor pathways, accelerate the recovery of motor nerves in corticospinal tract, and hence improve the recovery of limb motor functions (Dani, euml, Ezendam, Bongers, & Jannink, 2009).

Another hypothesis attributes mirror therapy to a specific form of visual-guided motor imagery (S. J. A & Phillips, 2003). When patients undergo mirror therapy, motor imagery is enhanced, which simulates the motion internally without obvious movement and activates the neural circuits involved in the movement control. In other words, mirror therapy improves the limb motor function by repairing the neural pathway of motor control. FC analysis is extensively used to evaluate correlations of neural activation among spatially distinct brain regions (Hallett et al., 2017). In this study, it is found that the FC between M1 and motor function-related brain regions, such as superior frontal gyrus, thalamus, paracentral lobule and postcentral gyrus, are significantly enhanced after the mirror therapy compared to the CT group, denoting stronger neuronal activities in the MT group (Ying et al., 2020). It is therefore speculated that mirror therapy improves the motor function of lower limbs by inducing the reorganization of neural networks in the motor cortex and promoting the brain plasticity through visual-guided motor imagery.

Furthermore, cross-limb transfer insinuates another possible mechanism. Cross-limb transfer is a phenomenon in which the functions of untrained limbs are significantly improved after unilateral limb training (S, P, & Paul, 2016). The unilateral exercise gives rise to bilateral enhancement in corticospinal excitability and produces follow-up effect due to the adaptive projection of nerves onto the untrained limb (L & G, 2013). Lara et al. (G. L. A & A, 2018) guided forty participants through a six-week unilateral training program of dominant wrist flexion or dorsiflexion, and found that the strength of the contralateral upper and lower limbs increased. Osumi et al. (Osumi, Sumitani, Otake, & Morioka, 2019) also discovered that bilateral upper limbs had mutual influence on sensory and motor functions. The sensory and motor functions of the affected upper limb were improved by the participation of the healthy upper limb. Based on the cross-limb transfer phenomenon, the authors boldly speculate that mirror therapy has a positive effect on the affected limb even without the stimulation of the mirror image. Nonetheless, the mirror image enhances the spatial attention of the affected limb by giving the patient an illusion of the affected limb in motion (Dohle, 2009). By transferring the patient’s attention to the affected limb, the motor consciousness of the affected limb is further strengthened, which restores the balance of cerebral hemispheres of stroke patients and improves the efficiency of information transmission and processing in motor function-related brain regions. Compared with the control group, the intensity of functional activity in these brain regions (fALFF, p < 0.05) is significantly different in the MT group after the mirror therapy. Mirror therapy not only confirms the theory of cross-limb transfer, but also provides a more effective treatment method in clinical practice based on this theory.

5 Conclusion

In conclusion, as a top-down rehabilitation intervention, mirror therapy improves the lower-limb motor function, balance, transfer capability and activity of daily living of stroke patients by enhancing the functional connectivity and activity of motor function-related regions of the brain. It is a safe and effective rehabilitation treatment method. This work is one of the first studies that investigate the possible underlying mechanism of mirror therapy in promoting the rehabilitation of lower limb function in patients after ischemic stroke by using rs-fMRI.

There are several limitations of our study. Firstly, a major limitation of this study is the lack of control of treatment time between the 2 groups. The MT group was provided with additional 30 mins of mirror therapy per session compared to the CT group, which adds up to an additional 7.5 hours of therapy than the CT group received. In fact, a sham therapy was provided to the CT in our pilot study, wherein the patients were asked to control the hemiplegic lower limb to imitate the movements of the normal limb while a wood plate was placed between the lower limbs. However, many patients showed low compliance and refused to complete the treatment. As a result, the sham therapy was not conducted in the current study. We will explore better sham therapy methods with better patient’s compliance in future studies. Secondly, the sample size is relatively small, and the treatment duration is relatively short. We hope to expand the sample size and prolong the treatment in future studies. Thirdly, depending on whether the affected limb participates in the activity or not, mirror therapy has different clinical treatment paradigms, including unilateral activity and bilateral activity (Narayan, Shanta, Dharmendra, & Vinod, 2015; Vural, Yuzer, Ozcan, Ozbudak, & Ozgirgin, 2016). In this study, we used the bilateral activity mode with the help of the therapist, but there is no control group to compare the unilateral activity mode and bilateral activity mode. More standardized operation scheme of mirror therapy is required in future research.

Footnotes

Acknowledgments

This work is financially supported by Sichuan Province Pharmaceutical Administration (Grant No. 2014B064), the Key R&D Program of Sichuan Province (No.2020YFS0415).

Conflict of interest

The authors declare no conflict of interest.

Ethical approval

Approval from the Ethics Committee of Sichuan Provincial People’s Hospital (IRB number 2014.20) for this study and informed consent of participants were obtained prior to the data collection.

References

A, G. L., & A, G. D. (2018) The cross education of strength and skill following unilateral strength training in the upper and lower limbs, Journal of Neurophysiology 120(2).

A, S. J., & Phillips,

S. M. E.

(2003) Using motor imagery in the rehabilitation of hemiparesis, Archives of Physical Medicine and Rehabilitation 84(7).

Aliyeh,

, Mokhtar,

, & Gholamreza,

(2018) Effect of different designs of ankle-foot orthoses on gait in patients with stroke: A systematic review, Gait & Posture 62.

Altschuler,

E. L.

, Wisdom,

S. B.

, Stone,

, Foster,

, Galasko,

, Llewellyn,

D. M. E.

, & Ramachandran,

(1999) Rehabilitation of hemiparesis after stroke with a mirror, The Lancet 353(9169).

Arya,

K. N.

, Pandian,

, & Kumar,

(2019) Effect of activity-based mirror therapy on lower limb motor-recovery and gait in stroke: A randomised controlled trial, Neuropsychological Rehabilitation 29(8), 1193–1210.

Chao-Gan,

, & Yu-Feng,

(2010) DPARSF: A MATLAB Toolbox for “Pipeline” Data Analysis of Resting-State fMRI, frontiers in Systems Neuroscience 4, 13–13. doi: 10.3389/FNSYS.2010.00013.

Dani, euml, Ezendam,

, Bongers,

R. M.

, & Jannink,

M. J. A.

(2009) Systematic review of the effectiveness of mirror therapy in upper extremity function, Disability & Rehabilitation 31(26).

Dohle,

(2009) Mirror therapy promotes recovery from severe hemiparesis: a randomized controlled trial, Neurorehabilitation and Neural Repair 23(3).

Downs,

, Marquez,

, & Chiarelli,

(2013) The Berg Balance Scale has high intra- and inter-rater reliability but absolute reliability varies across the scale: a systematic review, Journal of Physiotherapy 59(2), 93–99. doi: 10.1016/S1836-9553(13)70161-9.

10.

E,M.M.,Marion, S.,M, R. G., J, S. H., N, v. d. G. J., J, B. J. B., & W, S. R. (2011) The neuronal correlates of mirror therapy: an fMRI study on mirror induced visual illusions in patients with stroke, Journal of Neurology, Neurosurgery, and Psychiatry 82(4).

11.

E, M. M., W, S. R., N, v. d. G. J., Martine, E., Gunes, Y., J, S. H.,... J, B. J. B. (2011) Motor recovery and cortical reorganization after mirror therapy in chronic stroke patients: a phase II randomized controlled trial, Neurorehabilitation and Neural Repair 25(3).

12.

Farsin,

, Henning,

L. C.

, Volkmar,

, Irina,

, Michel,

, & Cornelius,

(2012) Functional plasticity induced by mirror training: the mirror as the element connecting both hands to one hemisphere, Neurorehabilitation and Neural Repair 26(5).

13.

Fritzsch,

, Wang,

, Santos,

L. F. d.

, Mauritz,

K.-H.

, Brunetti,

, & Dohle,

(2014) Different effects of the mirror illusion on motor and somatosensory processing, Restorative Neurology and Neuroscience 32(2).

14.

Ghadir,

, Rifaii,

S. A.

, & Soha,

(2016) Mirror therapy combined with functional electrical stimulation for rehabilitation of stroke survivors’ ankle dorsiflexion, Conference proceedings: IEEE Engineering in Medicine and Biology Society 2016.

15.

Giovanni,

, Ana,

, & L, S. S. (2006) Functions of the mirror neuron system: implications for neurorehabilitation, Cognitive and Behavioral Neurology: Official Journal of the Society for Behavioral and Cognitive Neurology 19(1).

16.

Gunes,

, Ruud,

, Nebahat,

, Serap,

, B, B. J., Füsun,

, ...J, S. H. (2008) Mirror therapy improves hand function in subacute stroke: a randomized controlled trial, Archives of Physical Medicine and Rehabilitation 89(3).

17.

Hallett,

, Iorio,

R. D.

, Rossini,

P. M.

, Park,

J. E.

, Chen,

, Celnik,

, ...

Ugawa, Y.

(2017) Contribution of transcranial magnetic stimulation to assessment of brain connectivity and networks, Clinical Neurophysiology 128(11).

18.

Holm,

, Nadine,

, Jan,

, Marcus,

, Johann,

, Bernhard,

, & Christian,

(2018) Mirror therapy for improving motor function after stroke, The Cochrane database of Systematic Reviews 7.

19.

Hong,

, Sui,

, Zhuang,

, Liu,

, Zheng,

, Cai,

, & Jin,

(2018) Effectiveness of neuromuscular electrical stimulation on lower limbs of patients with hemiplegia after chronic stroke: a systematic review, Archives of Physical Medicine and Rehabilitation 99(5).

20.

Huang,

C.-Y.

, Lin,

G.-H.

, Huang,

Y.-J.

, Song,

C.-Y.

, Lee,

Y.-C.

, How,

M.-J.

, ...

Hsieh, C.-L.

(2016) Improving the utility of the Brunnstrom recovery stages in patients with stroke: Validation and quantification, Medicine 95(31). doi: 10.1097/MD.0000000000004508.

21.

Ji,

S. G.

, & Kim,

M. K.

(2015) The effects of mirror therapy on the gait of subacute stroke patients: a randomized controlled trial, Clinical Rehabilitation 29(4).

22.

Jing,

, Claire,

, Johannes,

, Susanne,

, Karl-Heinz,

, Maddalena,

, & Christian,

(2013) A comparison of neural mechanisms in mirror therapy and movement observation therapy, Journal of Rehabilitation Medicine 45(4).

23.

K, B., S, W.-D., & Ji,

(1995) The Balance Scale: reliability assessment with elderly residents and patients with an acute stroke, Scandinavian Journal of Rehabilitation Medicine 27(1), 27–36.

24.

Kenji,

, Hiroyuki,

, Sayaka,

, Kazuya,

, Yusuke,

, Genichi,

, ...

Shigeru, S.

(2015) Randomized controlled comparative study on effect of training to improve lower limb motor paralysis in convalescent patients with post-stroke hemiplegia, Journal of Physical Therapy Science 27(9).

25.

Kim,

, Lee,

, Kim,

, Lee,

, & Kim,

(2016) Effects of mirror therapy combined with motor tasks on upper extremity function and activities daily living of stroke patients, Journal of Physical Therapy Science 28(2), 483–487. doi: 10.1589/JPTS.28.483.

26.

Koen,

, Marion,

, N, V. d. G. J., Aad, V. d. L., Ruth, S., J, S. H., & W, S. R. (2009) Mirror-induced visual illusion of hand movements: a functional magnetic resonance imaging study, Archives of Physical Medicine and Rehabilitation 90(4).

27.

L, R. K., & G, C. R. (2013) Neural pathways mediating cross education of motor function, Frontiers in Human Neuroscience 7.

28.

Langhorne,

, Bernhardt,

, & Kwakkel,

(2011) Stroke rehabilitation, The Lancet 377(9778).

29.

Lee,

H. J.

, Kim,

Y. M.

, & Lee,

D. K.

(2017) The effects of action observation training and mirror therapy on gait andbalance in stroke patients, The Society of Physical Therapy Science 29(3).

30.

Lennon,

, & Johnson,

(2000) The modified rivermead mobility index: validity and reliability, Disability and Rehabilitation 22(18), 833–839. doi: 10.1080/09638280050207884.

31.

Lv,

, Wang,

, & Shao,

(2020) Rehabilitation training activity analysis method for motor function in lower limb in elderly patients with hemiplegia after stroke. +, Rehabilitation Medicine 30(02), 161. doi: 10.3724.

32.

Mekbib,

D. B.

, Zhao,

, Wang,

, Xu,

, Zhang,

, Cheng,

, ...

Xu, D.

(2020) Proactive Motor Functional Recovery Following Immersive Virtual Reality– Based Limb Mirroring Therapy in Patients with Subacute Stroke, Neurotherapeutics 17(4), 1919–1930. doi: 10.1007/S13311-020-00882-X.

33.

Minjae,

, & Hyun-mo,

(2018) The Effects of Mirror Therapy with Functional Electrical Stimulation Therapy for The Stroke Patients. In (Vol. 6, pp. 123-130).

34.

MSc, G. B., MCSP, & Smith,

(2001) The recovery of walking ability and subclassification of stroke, Physiotherapy Research International 6(3).

35.

Narayan,

A. K.

, Shanta,

, Dharmendra,

, & Vinod,

(2015) Task-based mirror therapy augmenting motor recovery in poststroke hemiparesis: a randomized controlled trial, Journal of Stroke and Cerebrovascular Diseases: The Official Journal of National Stroke Association 24(8).

36.

Osumi,

, Sumitani,

, Otake,

, & Morioka,

(2019) A hypothetical explanatory sensorimotor model of bilateral limb interference, Medical Hypotheses 122.

37.

Rizzolatti,

(2005) The mirror neuron system and its function in humans, Anatomy and Embryology 210(5-6).

38.

Rjosk,

, Lepsien,

, Kaminski,

, Hoff,

, Sehm,

, Steele,

C. J.

, ...

Ragert, P.

(2017) Neural Correlates of Mirror Visual Feedback-Induced Performance Improvements: A Resting-State fMRI Study, Frontiers in Human Neuroscience 11. doi: 10.3389/FNHUM.2017.00054.

39.

Ros,

, eacute, ouml, & Lundborg, r. (2005) Training with a mirror in rehabilitation of the hand, Scandinavian Journal of Plastic and Reconstructive Surgery and Hand Surgery 39(2).

40.

S, B. T., P, P. G. E., & Paul,

Z. E.

(2016) Cross-education of strength and skill: an old idea with applications in the aging nervous system, The Yale Journal of Biology and Medicine 89(1).

41.

Safaz,

İ.

, Ylmaz,

, Yaşar,

, & Alaca,

(2009) Brunnstrom recovery stage and motricity index for the evaluation of upper extremity in stroke: analysis for correlation and responsiveness, International Journal of Rehabilitation Research 32(3), 228–231. doi: 10.1097/MRR.0B013E32832A62AD.

42.

Salem,

H. M. A.

, & Huang,

(2015) The effects of mirror therapy on clinical improvement in hemiplegic lower extremity rehabilitation in subjects with chronic stroke, International Journal of Biomedical and Biological Engineering 9(2), 163–166.

43.

Sanford,

, Moreland,

, Swanson,

L. R.

, Stratford,

P. W.

, & Gowland,

(1993) Reliability of the Fugl-Meyer Assessment for Testing Motor Performance in Patients Following Stroke, Physical Therapy 73(7), 447–454. doi: 10.1093/PTJ/73.7.447.

44.

Sutbeyaz,

, Yavuzer,

, Sezer,

, & Koseoglu,

B. F.

(2007) Mirror therapy enhances lower-extremity motor recovery and motor functioning after stroke: a randomized controlled trial, Archives of Physical Medicine and Rehabilitation 88(5), 555–559. doi: 10.1016/j.apmr.2007.02.034.

45.

Thieme,

, Morkisch,

, Mehrholz,

, Pohl,

, Behrens,

, Borgetto,

, & Dohle,

(2019) Mirror Therapy for Improving Motor Function After Stroke: Update of a Cochrane Review, Stroke 50(2). doi: 10.1161/STROKEAHA.118.023092.

46.

Uthra,

, Karthik,

B. S.

, Vijay,

K. K.

, K. K., V, S. B., K, M. Z., & M, C. (2013) Effectiveness of mirror therapy on lower extremity motor recovery, balance and mobility in patients with acute stroke: A randomized sham-controlled pilot trial, Annals of Indian Academy of Neurology 16(4).

47.

Vural,

S. P.

, Yuzer,

G. F. N.

, Ozcan,

D. S.

, Ozbudak,

S. D.

, & Ozgirgin,

(2016) Effects of mirror therapy in stroke patients with complex regional pain syndrome type 1: a randomized controlled study, Archives of Physical Medicine and Rehabilitation 97(4), 1.

48.

Wang,

, Fritzsch,

, Bernarding,

, Krause,

, Mauritz,

K.-H.

, Brunetti,

, & Dohle,

(2013) Cerebral activation evoked by the mirror illusion of the hand in stroke patients compared to normal subjects, Neuro Rehabilitation 33(4).

49.

Xiong,

, Parsons,

L. M.

, Gao,

J. H.

, & Fox,

P. T.

(1999) Interregional connectivity to primary motor cortex revealed using MRI resting state images, Human Brain Mapping 8(2-3).

50.

Ya-Long,

, Yan,

, & Bai-Ya,

(2019) Effectiveness of neuromuscular electrical stimulation combined with rehabilitation training for treatment of post-stroke limb spasticity, Medicine 98(39).

51.

Yang,

C. M.

, Wang,

Y.-C.

, Lee,

C.-H.

, Chen,

M.-H.

, & Hsieh,

C.-L.

(2020) A comparison of test-retest reliability and random measurement error of the Barthel Index and modified Barthel Index in patients with chronic stroke. Disability and Rehabilitation, 1-5. doi:10.1080/09638288.2020.1814429.

52.

Yin,

, Liu,

, Zhang,

, Zheng,

, Peng,

, Ai,

, ...

Hu, X.

(2020) Effects of rTMS Treatment on Cognitive Impairment and Resting-State Brain Activity in Stroke Patients: A Randomized Clinical Trial, Frontiers in Neural Circuits 14. doi: 10.3389/FNCIR.2020.563777.

53.

Ying,

, Kai,

, Zhenyu,

, Guanmao,

, Yanbin,

, Shuming,

, ...

Jie, T.

(2020) Classification of unmedicated bipolar disorder using whole-brain functional activity and connectivity: a radiomics analysis. Cerebral Cortex (New York, N.Y. : 1991), 30(3).

54.

Yu-Rong,

, Leung,

L. W.

, Qiang,

, Le,

, Xiang,

, Preeti,

, & Dong-Feng,

(2015) The effect of body weight support treadmill training on gait recovery, proximal lower limb motor pattern, and balance in patients with subacute stroke, BioMed Research International 2015.

55.

Zhang,

, Yue,

, Wang,

, & Kuang,

(2017) Robotics in lower-limb rehabilitation after stroke, Behavioural Neurology 2017.